Biowerkzeug Wiki - User contributions [en]

Analysis software

2011-04-21T01:08:19Z

Oliver: /* MD Analysis libraries */

Running simulations is often the easy bit. The hard bit is to extract meaningful information from the Gigabytes of trajectory data. This list can act as a starting point. For most advanced uses, however, one will probably have to write analysis code in [[python]], [[Perl]], [[tcl]], [[C/C++]], [[bash]] ... or any other language that "gets the job done".

== "Native" tools ==
Many of the [[Molecular dynamics software|MD packages]] come with their own analysis tools or scripting language. Sometimes it is possible to [[convert data formats]] between packages and use the other package's analysis tools.

;[[Gromacs]] analysis tools: oone of the strengths of Gromacs is that it comes with a large number of useful analysis tools that make many of the standard analysis tasks simple to perform

;NAMD/[[VMD]]: VMD can be used through its GUI or by scripting it in [[tcl]] to great effect

;[[Charmm]]: Charmm is feature-rich but its scripting language can cause a steep learning curve

;LAMMPS/[[pizza]]: pizza.py is a python library geared towards output from [[LAMMPS]]

;Amber/[[ptraj]]: command-line based analysis

== MD Analysis libraries ==

;[http://mdanalysis.googlecode.com/ MDAnalysis]: a python library to analyze a range of trajectories (e.g. DCD, XTC, TRR, XYZ) and single frames (PDB, GRO, CRD, PQR).

;[http://dirac.cnrs-orleans.fr/MMTK/ MMTK]: Another python-based framework for doing analysis is the ''Molecular Modelling Tool Kit''. However, it does not natively read [[Charmm]] dcd files and hence it can be cumbersome to use.

;[http://loos.sourceforge.net LOOS]: The ''Lightweight Object-Oriented Structure library (LOOS)'' from Alan Grossfield's lab provides a lightweight C++ library for analysis of molecular dynamics simulations. This includes parsing a number of PDB variants, as well as the native system description and trajectory formats for CHARMM, NAMD, and Amber. LOOS is not intended to be an all-encompassing library and it is primarily geared towards reading data in and processing rather than manipulating the files and structures and writing them out.

== Specialized tools ==

;[http://hole.biop.ox.ac.uk/hole HOLE]: Oliver Smart's program to trace out pore surfaces and estimate single channel conductances.

;CAVER: [http://loschmidt.chemi.muni.cz/caver/index.php CAVER] provides rapid, accurate and fully automated calculation of pathways leading from buried cavities to outside solvent in static and dynamic protein structures. Calculated pathways can be visualized by graphic program PyMol dissecting anatomy and dynamics of entrance tunnels. CAVER allows analysis of any molecular structure including proteins, nucleic acids, inorganic materials, etc. CAVER is available as [http://loschmidt.chemi.muni.cz/caver/online.php online version] or [[PyMol]] plugin suitable for calculation of pathways in discrete protein structures and stand alone version enabling analysis of trajectories from the molecular dynamics simulations.

;[http://swift.cmbi.ru.nl/gv/dssp/ dssp]: ''Definition of secondary structure of proteins given a set of 3D coordinates.'' The DSSP program defines secondary structure, geometrical features and solvent exposure of proteins, given atomic coordinates in Protein Data Bank format. The program does NOT PREDICT protein structure. According to the Science Citation Index (July 1995), the program has been cited in the scientific literature more than 1000 times.

; [http://www.compbio.dundee.ac.uk/manuals/stamp.4.2/stamp.html STAMP]: ''Structural Alignment of Multiple Proteins''. STAMP is a package for the alignment of protein sequence based on three-dimensional (3D) structure. It provides not only multiple alignments and the corresponding `best-fit' superimpositions, but also a systematic and reproducible method for assessing the quality of such alignments. It also provides a method for protein 3D structure data base scanning. In addition to structure comparison, the STAMP package provides input for programs to display and analyse protein sequence alignments and tertiary structures. Please note that, although STAMP outputs a sequence alignment, it is a program for 3D structures, and NOT sequences.

;[[swinker]]: finds and calculates helix hinges. It optionally finds the hinge point and calculates kink and swivel angles.

== General purpose mathematical packages ==
;[[Scientific Python]] and [[pylab]]: a [[matlab]]-like [[python]] module that has sophisticated analysis and plotting capabilities
;[[matlab]]:
;[[Mathematica]]:
;[[R]]: R is a language and environment for statistical computing and graphics. R provides a wide variety of statistical (linear and nonlinear modelling, classical statistical tests, time-series analysis, classification, clustering, ...) and graphical techniques, and is highly extensible. One of R's strengths is the ease with which well-designed publication-quality plots can be produced, including mathematical symbols and formulae where needed. Great care has been taken over the defaults for the minor design choices in graphics, but the user retains full control.

[[Category:Analysis]]
[[Category:Software]]

Protein/Membrane system size

2010-05-22T10:28:02Z

Oliver: math -> manual formatting

This page collects notes on how to decide how to initially setup a membrane protein simulation. ''Please do not follow these notes blindly — they are more [http://www.imdb.com/title/tt0325980/quotes guidelines] than actual rules.''

Once you've decided on box dimensions and number of lipids you can use any of the [[Transmembrane protein insertion methods]] methods to set up the system.

== System size ==
* Decide on the bilayer composition. Have you got [[Lipid FFDB|force field parameters for the lipids]]?
* System size:
** rule of thumb 1: have at least 2-3 lipid layers between protein and periodic box boundary in the x-y plane to ensure(?) membrane-like behaviour; more may be necessary: you will need to monitor the hydrophobic mismatch and make the system big enough so that the bilayer at the box edges has its natural thickness.
** rule of thumb 2: have at least 1-1.5nm of water between protein and z-boundary (so that Coulomb forces are effectively screened)
** Especially for coarse-grained self-assembly simulations you also need to take into account the [[BONDINI#A rough guide to System Content and Membrane Structure|lipid-to-water ratio and its affect on membrane structure]].
* Add ions at e.g. 100 mM (+counter ions) for additional Coulomb screening
* Remember that run time scales like ''N'' log ''N'' (or even ''N''2) and ''N'' scales with the volume, i.e. ''L''3! Thus: keep your system small and pretty (i.e. the smallest system that still behaves like a big system... invariably you are trading size-artifacts versus speed and thus sampling)

== Number of lipids ==

Estimate the cross sectional area of the protein, e.g. using the radius of gyration

''A''prot = π ''R''G

and with the approximate area per lipid ''A''l = 64 Å the number of lipids in two leaflets is for a box with dimensions ''L''x and ''L''y

''N'' = 2(''L''x''L''y - ''A''prot) / ''A''l

If you are simulating a mixed bilayer, e.g. POPE:POPG 4:1 then you will need 4''N''/5 POPE lipids and ''N/''5 POPG ones.

== Number of ions ==

Once the system is solvated you can calculate the numbers of ions <math>N</math> needed for a given concentration ''c'' (in addition to any counter ions):

''c'' = ''n''/''V'' = ''N''/(''NAV)

(where ''N''A = 6.02214179 × 1023mol-1 is the [http://physics.nist.gov/cgi-bin/cuu/Value?na Avogadro constant]).

The problem is to get the water volume of the inhomogenous system, consisting of membrane and protein with the water. An approximation is to use the volume of a water molecule at standard conditions, ''v''w = 30 Å3 and calculate ''V'' = ''N''w ''v''w from the number of waters in the system, ''N''w. The equations become even simpler if we use the standard concentration of water, ''c''w = ''n''w/(''N''w ''v''w) = 55.5 mol/l.

Using all this we get

N+ = ''c''+/''V'' = ''c''+/''c''w ''N''w

for the (monovalent) cations, and correspondingly for monovalent anions

''N''- = ''N''+

=== Example ===
For a concentration of 100 mM = 0.1 M and 8000 water molecules we will need to add

''N''- = ''N''+ = 0.1 M/55.5 M × 8000 = 14

ions of each kind (e.g. 14 sodium and 14 chloride and whatever counterions are required to make the system charge neutral).

[[Category:MembraneInsertion]]
[[Category:Lipids]]
[[Category:Tutorial]]

Transmembrane protein insertion methods

2010-05-22T10:15:03Z

Oliver:

Published methods for setting up membrane protein simulations; please add more.

; [http://wwwuser.gwdg.de/~ggroenh/membed.html g_membed]: Slowly growing a shrunken peptide into a bilayer. See ''g_membed: Efficient insertion of a membrane protein into an equilibrated lipid bilayer with minimal perturbation''. Maarten G. Wolf, Martin Hoefling, Camilo Aponte-Santamaría, Helmut Grubmüller, Gerrit Groenhof. J Comp Chem (2010). doi:[http://dx.doi.org/10.1002/jcc.21507 10.1002/jcc.21507]
; Tielemann's "shrinking an exploded bilayer": Scale all COM coordinates of lipids to increase space ("explode"), insert peptide, shrink via MD (see recent Tieleman review). ''Setting up and running molecular dynamics simulations of membrane proteins.'' Kandt C, Ash WL, Tieleman DP. Methods '''41''' (2007), 475-88. doi:[http://dx.doi.org/10.1016/j.ymeth.2006.08.006 10.1016/j.ymeth.2006.08.006]
; TaraGrid: an update on the Faraldo-Goméz/Smith "make hole" approach (?); ''Automated Protein-Insertion into Membranes for Molecular Dynamics Simulation Set-Up Using Taragrid'', René Staritzbichler, Lucy R. Forrest and José Faraldo-Gómez. Biophysics 2010 abstract [http://dx.doi.org/10.1016/j.bpj.2009.12.2105 10.1016/j.bpj.2009.12.2105]
; Griffin: Biophysics 2010 abstract ''Automated and Optimized Embedding of Proteins into Membranes for Molecular Dynamics Simulations using Griffin'', René Staritzbichler, Lucy R. Forresta and José D. Faraldo-Gómez doi:[http://dx.doi.org/10.1016/j.bpj.2009.12.3070 10.1016/j.bpj.2009.12.3070]
; gromacs make_hole: Faraldo-Goméz/Smith: use GRASP surface to gently generate a protein shaped hole in the bilayer; requires a special Gromacs binary [http://www.gromacs.org/@api/deki/files/48/=mdrun_make_hole.tar.gz mdrun_make_hole]. ''Setting up and optimization of membrane protein simulations.'' Faraldo-Gómez JD, Smith GR, Sansom MSP. Eur Biophys J. '''31'' (2002), 217-227. doi:[http://dx.doi.org/10.1007/s00249-002-0207-5 10.1007/s00249-002-0207-5]
; CHARMM style (Woolf & Roux): build bilayer from a library of lipid conformers+bound water that are translated and rotated as rigid units Files in the contributed section of CHARMM and at [http://thallium.bsd.uchicago.edu/RouxLab/membrane.html membrane builder]. ''Molecular dynamics simulation of the gramicidin channel in a phospholipid bilayer.'' Woolf TB, Roux B. Proc Natl Acad Sci U S A. '''91''' (1994), 11631-5 and ''Structure, energetics, and dynamics of lipid-protein interactions: A molecular dynamics study of the gramicidin A channel in a DMPC bilayer'', T. Woolf and B. Roux, Proteins '''24''' (1996) 92-114 doi:[http://www3.interscience.wiley.com/journal/69546/abstract 10.1002/(SICI)1097-0134(199601)24:1<92::AID-PROT7>3.0.CO;2-Q]
; CHARMM-GUI [http://www.charmm-gui.org/?doc=input/membrane Membrane-builder]: automated web procedure (generate input files, run locally. ''Automated Builder and Database of Protein/Membrane Complexes for Molecular Dynamics Simulations'', S. Jo, T. Kim, and W. Im PLoS ONE '''2''' (2007) e880 . [http://dx.doi.org/10.1371/journal.pone.0000880 10.1371/journal.pone.0000880].

[[Category:MembraneInsertion]]
[[Category:Lipids]]
[[Category:Protocols]]
[[Category:Gromacs]]
[[Category:CHARMM]]

Membrane proteins

2010-05-22T10:13:24Z

Oliver: overview

[[Category:Lipids]]
[[Category:Protocols]]

== Setting up membrane protein simulations ==
# decide on the [[Protein/Membrane system size|size of the protein/membrane system]]
# obtain parameters (for lipids see, for instance, [http://lipidbook.bioch.ox.ac.uk Lipidbook])
# choose a [[transmembrane protein insertion methods|transmembrane protein insertion method]]
# run simulation

Transmembrane protein insertion methods

2010-05-22T10:12:53Z

Oliver: moved insertion methods to extra page

== Methods for setting up membrane protein simulations ==

Published methods:

; [http://wwwuser.gwdg.de/~ggroenh/membed.html g_membed]: Slowly growing a shrunken peptide into a bilayer. See ''g_membed: Efficient insertion of a membrane protein into an equilibrated lipid bilayer with minimal perturbation''. Maarten G. Wolf, Martin Hoefling, Camilo Aponte-Santamaría, Helmut Grubmüller, Gerrit Groenhof. J Comp Chem (2010). doi:[http://dx.doi.org/10.1002/jcc.21507 10.1002/jcc.21507]
; Tielemann's "shrinking an exploded bilayer": Scale all COM coordinates of lipids to increase space ("explode"), insert peptide, shrink via MD (see recent Tieleman review). ''Setting up and running molecular dynamics simulations of membrane proteins.'' Kandt C, Ash WL, Tieleman DP. Methods '''41''' (2007), 475-88. doi:[http://dx.doi.org/10.1016/j.ymeth.2006.08.006 10.1016/j.ymeth.2006.08.006]
; TaraGrid: an update on the Faraldo-Goméz/Smith "make hole" approach (?); ''Automated Protein-Insertion into Membranes for Molecular Dynamics Simulation Set-Up Using Taragrid'', René Staritzbichler, Lucy R. Forrest and José Faraldo-Gómez. Biophysics 2010 abstract [http://dx.doi.org/10.1016/j.bpj.2009.12.2105 10.1016/j.bpj.2009.12.2105]
; Griffin: Biophysics 2010 abstract ''Automated and Optimized Embedding of Proteins into Membranes for Molecular Dynamics Simulations using Griffin'', René Staritzbichler, Lucy R. Forresta and José D. Faraldo-Gómez doi:[http://dx.doi.org/10.1016/j.bpj.2009.12.3070 10.1016/j.bpj.2009.12.3070]
; gromacs make_hole: Faraldo-Goméz/Smith: use GRASP surface to gently generate a protein shaped hole in the bilayer; requires a special Gromacs binary [http://www.gromacs.org/@api/deki/files/48/=mdrun_make_hole.tar.gz mdrun_make_hole]. ''Setting up and optimization of membrane protein simulations.'' Faraldo-Gómez JD, Smith GR, Sansom MSP. Eur Biophys J. '''31'' (2002), 217-227. doi:[http://dx.doi.org/10.1007/s00249-002-0207-5 10.1007/s00249-002-0207-5]
; CHARMM style (Woolf & Roux): build bilayer from a library of lipid conformers+bound water that are translated and rotated as rigid units Files in the contributed section of CHARMM and at [http://thallium.bsd.uchicago.edu/RouxLab/membrane.html membrane builder]. ''Molecular dynamics simulation of the gramicidin channel in a phospholipid bilayer.'' Woolf TB, Roux B. Proc Natl Acad Sci U S A. '''91''' (1994), 11631-5 and ''Structure, energetics, and dynamics of lipid-protein interactions: A molecular dynamics study of the gramicidin A channel in a DMPC bilayer'', T. Woolf and B. Roux, Proteins '''24''' (1996) 92-114 doi:[http://www3.interscience.wiley.com/journal/69546/abstract 10.1002/(SICI)1097-0134(199601)24:1<92::AID-PROT7>3.0.CO;2-Q]
; CHARMM-GUI [http://www.charmm-gui.org/?doc=input/membrane Membrane-builder]: automated web procedure (generate input files, run locally. ''Automated Builder and Database of Protein/Membrane Complexes for Molecular Dynamics Simulations'', S. Jo, T. Kim, and W. Im PLoS ONE '''2''' (2007) e880 . [http://dx.doi.org/10.1371/journal.pone.0000880 10.1371/journal.pone.0000880].

[[Category:MembraneInsertion]]
[[Category:Lipids]]
[[Category:Protocols]]
[[Category:Gromacs]]
[[Category:CHARMM]]

Protein/Membrane system size

2010-05-22T10:11:49Z

Oliver: estimate num ber of lipids (from SBCBWiki, written by OB)

This page collects notes on how to decide how to initially setup a membrane protein simulation. ''Please do not follow these notes blindly — they are more [http://www.imdb.com/title/tt0325980/quotes guidelines] than actual rules.''

Once you've decided on box dimensions and number of lipids you can use any of the [[Membrane protein insertion]] methods to set up the system.

== System size ==
* Decide on the bilayer composition. Have you got [[Lipid FFDB|force field parameters for the lipids]]?
* System size:
** rule of thumb 1: have at least 2-3 lipid layers between protein and periodic box boundary in the x-y plane to ensure(?) membrane-like behaviour; more may be necessary: you will need to monitor the hydrophobic mismatch and make the system big enough so that the bilayer at the box edges has its natural thickness.
** rule of thumb 2: have at least 1-1.5nm of water between protein and z-boundary (so that Coulomb forces are effectively screened)
** Especially for coarse-grained self-assembly simulations you also need to take into account the [[BONDINI#A rough guide to System Content and Membrane Structure|lipid-to-water ratio and its affect on membrane structure]].
* Add ions at eg 100 mM (+counter ions) for additional Coulomb screening
* Remember that run time scales like <math>N \log N</math> (or even <math>N^2</math>) and <math>N</math> scales with the volume, i.e. <math>L^3</math>! Thus: keep your system small and pretty (ie the smallest system that still behaves like a big system... invariably you are trading size-artifacts versus speed and thus sampling)

== Number of lipids ==

Estimate the cross sectional area of the protein, eg using the radius of gyration

<math>A_{\mathrm{prot}} = \pi R_G</math>

and with the approximate area per lipid <math>A_l = 64</math> Å the number of lipids in two leaflets is for a box with dimensions <math>L_x</math> and <math>L_y</math>

<math>N = \frac{2(L_x L_y - A_\mathrm{prot})}{A_l}</math>.

If you are simulating a mixed bilayer, eg POPE:POPG 4:1 then you will need <math>4N/5</math> POPE lipids and <math>N/5</math> POPG ones.

== Number of ions ==

Once the system is solvated you can calculate the numbers of ions <math>N</math> needed for a given concentration <math>c</math> (in addition to any counter ions):

<math>c = \frac{n}{V} = \frac{N/N_A}{V}</math>

(where <math>N_A = 6.02214179 \times 10^{23} \mathrm{mol}^{-1}</math> is the [http://physics.nist.gov/cgi-bin/cuu/Value?na Avogadro constant]).

The problem is to get the water volume of the inhomogenous system, consisting of membrane and protein with the water. An approximation is to use the volume of a water molecule at standard conditions, <math>v_w = 30 \mathrm{\AA}^3</math> and calculate <math>V = N_w v_w</math> from the number of waters in the system, <math>N_w</math>. The equations become even simpler if we use the standard concentration of water, <math>c_w = n_w/(N_w v_w) = 55.5\, \mathrm{mol/l}</math>.

Using all this we get

<math>N_+ = \frac{c_+}{V} = \frac{c_+}{c_w} N_w</math>

for the (monovalent) cations, and correspondingly for monovalent anions

<math>N_- = N_+</math>

=== Example ===
For a concentration of 100 mM = 0.1 M and 8000 water molecules we will need to add

<math>N_+ = N_- = \frac{0.1\,\mathrm{M}}{55.5\,\mathrm{M}} \times 8000 = 14</math>

ions of each kind (eg 14 sodium and 14 chloride and whatever counterions are required to make the system charge neutral).

[[Category:MembraneInsertion]]
[[Category:Lipids]]
[[Category:Tutorial]]

Membrane proteins

2010-05-22T09:58:16Z

Oliver: methods for setting up membrane protein simulations

Membranes

2010-01-10T13:40:12Z

Oliver: /* Explicit lipids */

Simulations of proteins must also represent the environment faithfully (simulations in [[vacuum simulations|vacuum]] are ''not'' appropriate in most cases and really date back to the days when computer resources were severely limited). The environment of a protein consists of the [[solvent]] and, if it's a membrane protein, of the [[lipid bilayer]]. Here we discuss methods to represent the '''lipid membrane'''.

== Explicit lipids ==

Force-field parameters for lipids are often not distributed with the standard force field files but can be found on the [http://lipidbook.bioch.ox.ac.uk lipidbook] site.

=== All atom ===
All atom representations include heavy atoms and ''all'' hydrogens. See [[#United atom|united atom]] for the alternative.

=== United atom ===
A ''united atom'' representation combines carbons and aliphatic hydrogens into ''unified'' particles.

== Implicit membrane ==

=== Dielectric slab ===

=== Statistical potential-based membrane ===
The membrane is parametrized based on the distribution of amino acids along the bilayer normal <cite>Ulmschneider2005</cite>.

== References ==
<biblio>
#Ulmschneider2005 pmid=15723347
</biblio>

Membranes

2010-01-10T13:39:51Z

Oliver: /* Explicit lipids */ shameless plug: lipidbook

Simulations of proteins must also represent the environment faithfully (simulations in [[vacuum simulations|vacuum]] are ''not'' appropriate in most cases and really date back to the days when computer resources were severely limited). The environment of a protein consists of the [[solvent]] and, if it's a membrane protein, of the [[lipid bilayer]]. Here we discuss methods to represent the '''lipid membrane'''.

== Explicit lipids ==

Parameters for lipids are often not distributed with the standard force field files but can be found on the [http://lipidbook.bioch.ox.ac.uk lipidbook] site.

=== All atom ===
All atom representations include heavy atoms and ''all'' hydrogens. See [[#United atom|united atom]] for the alternative.

=== United atom ===
A ''united atom'' representation combines carbons and aliphatic hydrogens into ''unified'' particles.

== Implicit membrane ==

=== Dielectric slab ===

=== Statistical potential-based membrane ===
The membrane is parametrized based on the distribution of amino acids along the bilayer normal <cite>Ulmschneider2005</cite>.

== References ==
<biblio>
#Ulmschneider2005 pmid=15723347
</biblio>

Analysis software

2009-02-07T14:09:24Z

Oliver: /* Specialized tools */

Running simulations is often the easy bit. The hard bit is to extract meaningful information from the Gigabytes of trajectory data. This list can act as a starting point. For most advanced uses, however, one will probably have to write analysis code in [[python]], [[Perl]], [[tcl]], [[C/C++]], [[bash]] ... or any other language that "gets the job done".

== "Native" tools ==
Many of the [[Molecular dynamics software|MD packages]] come with their own analysis tools or scripting language. Sometimes it is possible to [[convert data formats]] between packages and use the other package's analysis tools.

;[[Gromacs]] analysis tools: oone of the strengths of Gromacs is that it comes with a large number of useful analysis tools that make many of the standard analysis tasks simple to perform

;NAMD/[[VMD]]: VMD can be used through its GUI or by scripting it in [[tcl]] to great effect

;[[Charmm]]: Charmm is feature-rich but its scripting language can cause a steep learning curve

;LAMMPS/[[pizza]]: pizza.py is a python library geared towards output from [[LAMMPS]]

;Amber/[[ptraj]]: command-line based analysis

== MD Analysis libraries ==

;[http://mdanalysis.googlecode.com/ MDAnalysis]: a python library to analyze dcd trajectories (in conjunction with a psf)

;[http://dirac.cnrs-orleans.fr/MMTK/ MMTK]: Another python-based framework for doing analysis is the ''Molecular Modelling Tool Kit''. However, it does not natively read [[Charmm]] dcd files and hence it can be cumbersome to use.

;[http://membrane.urmc.rochester.edu/Software/LOOS/Docs/main.html LOOS]: The ''Lightweight Object-Oriented Structure library (LOOS)'' from Alan Grossfield's lab provides a lightweight C++ library for analysis of molecular dynamics simulations. This includes parsing a number of PDB variants, as well as the native system description and trajectory formats for CHARMM, NAMD, and Amber. LOOS is not intended to be an all-encompassing library and it is primarily geared towards reading data in and processing rather than manipulating the files and structures and writing them out.

== Specialized tools ==

;[http://hole.biop.ox.ac.uk/hole HOLE]: Oliver Smart's program to trace out pore surfaces and estimate single channel conductances.

;CAVER: [http://loschmidt.chemi.muni.cz/caver/index.php CAVER] provides rapid, accurate and fully automated calculation of pathways leading from buried cavities to outside solvent in static and dynamic protein structures. Calculated pathways can be visualized by graphic program PyMol dissecting anatomy and dynamics of entrance tunnels. CAVER allows analysis of any molecular structure including proteins, nucleic acids, inorganic materials, etc. CAVER is available as [http://loschmidt.chemi.muni.cz/caver/online.php online version] or [[PyMol]] plugin suitable for calculation of pathways in discrete protein structures and stand alone version enabling analysis of trajectories from the molecular dynamics simulations.

;[http://swift.cmbi.ru.nl/gv/dssp/ dssp]: ''Definition of secondary structure of proteins given a set of 3D coordinates.'' The DSSP program defines secondary structure, geometrical features and solvent exposure of proteins, given atomic coordinates in Protein Data Bank format. The program does NOT PREDICT protein structure. According to the Science Citation Index (July 1995), the program has been cited in the scientific literature more than 1000 times.

; [http://www.compbio.dundee.ac.uk/manuals/stamp.4.2/stamp.html STAMP]: ''Structural Alignment of Multiple Proteins''. STAMP is a package for the alignment of protein sequence based on three-dimensional (3D) structure. It provides not only multiple alignments and the corresponding `best-fit' superimpositions, but also a systematic and reproducible method for assessing the quality of such alignments. It also provides a method for protein 3D structure data base scanning. In addition to structure comparison, the STAMP package provides input for programs to display and analyse protein sequence alignments and tertiary structures. Please note that, although STAMP outputs a sequence alignment, it is a program for 3D structures, and NOT sequences.

;[[swinker]]: finds and calculates helix hinges. It optionally finds the hinge point and calculates kink and swivel angles.

== General purpose mathematical packages ==
;[[Scientific Python]] and [[pylab]]: a [[matlab]]-like [[python]] module that has sophisticated analysis and plotting capabilities
;[[matlab]]:
;[[Mathematica]]:
;[[R]]: R is a language and environment for statistical computing and graphics. R provides a wide variety of statistical (linear and nonlinear modelling, classical statistical tests, time-series analysis, classification, clustering, ...) and graphical techniques, and is highly extensible. One of R's strengths is the ease with which well-designed publication-quality plots can be produced, including mathematical symbols and formulae where needed. Great care has been taken over the defaults for the minor design choices in graphics, but the user retains full control.

[[Category:Analysis]]
[[Category:Software]]

Analysis software

2009-02-07T14:07:57Z

Oliver: /* MD Analysis libraries */

Running simulations is often the easy bit. The hard bit is to extract meaningful information from the Gigabytes of trajectory data. This list can act as a starting point. For most advanced uses, however, one will probably have to write analysis code in [[python]], [[Perl]], [[tcl]], [[C/C++]], [[bash]] ... or any other language that "gets the job done".

== "Native" tools ==
Many of the [[Molecular dynamics software|MD packages]] come with their own analysis tools or scripting language. Sometimes it is possible to [[convert data formats]] between packages and use the other package's analysis tools.

;[[Gromacs]] analysis tools: oone of the strengths of Gromacs is that it comes with a large number of useful analysis tools that make many of the standard analysis tasks simple to perform

;NAMD/[[VMD]]: VMD can be used through its GUI or by scripting it in [[tcl]] to great effect

;[[Charmm]]: Charmm is feature-rich but its scripting language can cause a steep learning curve

;LAMMPS/[[pizza]]: pizza.py is a python library geared towards output from [[LAMMPS]]

;Amber/[[ptraj]]: command-line based analysis

== MD Analysis libraries ==

;[http://mdanalysis.googlecode.com/ MDAnalysis]: a python library to analyze dcd trajectories (in conjunction with a psf)

;[http://dirac.cnrs-orleans.fr/MMTK/ MMTK]: Another python-based framework for doing analysis is the ''Molecular Modelling Tool Kit''. However, it does not natively read [[Charmm]] dcd files and hence it can be cumbersome to use.

;[http://membrane.urmc.rochester.edu/Software/LOOS/Docs/main.html LOOS]: The ''Lightweight Object-Oriented Structure library (LOOS)'' from Alan Grossfield's lab provides a lightweight C++ library for analysis of molecular dynamics simulations. This includes parsing a number of PDB variants, as well as the native system description and trajectory formats for CHARMM, NAMD, and Amber. LOOS is not intended to be an all-encompassing library and it is primarily geared towards reading data in and processing rather than manipulating the files and structures and writing them out.

== Specialized tools ==

;[[HOLE]]: Oliver Smart's program to trace out pore surfaces and estimate single channel conductances.

;CAVER: [http://loschmidt.chemi.muni.cz/caver/index.php CAVER] provides rapid, accurate and fully automated calculation of pathways leading from buried cavities to outside solvent in static and dynamic protein structures. Calculated pathways can be visualized by graphic program PyMol dissecting anatomy and dynamics of entrance tunnels. CAVER allows analysis of any molecular structure including proteins, nucleic acids, inorganic materials, etc. CAVER is available as [http://loschmidt.chemi.muni.cz/caver/online.php online version] or [[PyMol]] plugin suitable for calculation of pathways in discrete protein structures and stand alone version enabling analysis of trajectories from the molecular dynamics simulations.

;[http://swift.cmbi.ru.nl/gv/dssp/ dssp]: ''Definition of secondary structure of proteins given a set of 3D coordinates.'' The DSSP program defines secondary structure, geometrical features and solvent exposure of proteins, given atomic coordinates in Protein Data Bank format. The program does NOT PREDICT protein structure. According to the Science Citation Index (July 1995), the program has been cited in the scientific literature more than 1000 times.

; [http://www.compbio.dundee.ac.uk/manuals/stamp.4.2/stamp.html STAMP]: ''Structural Alignment of Multiple Proteins''. STAMP is a package for the alignment of protein sequence based on three-dimensional (3D) structure. It provides not only multiple alignments and the corresponding `best-fit' superimpositions, but also a systematic and reproducible method for assessing the quality of such alignments. It also provides a method for protein 3D structure data base scanning. In addition to structure comparison, the STAMP package provides input for programs to display and analyse protein sequence alignments and tertiary structures. Please note that, although STAMP outputs a sequence alignment, it is a program for 3D structures, and NOT sequences.

;[[swinker]]: finds and calculates helix hinges. It optionally finds the hinge point and calculates kink and swivel angles.

== General purpose mathematical packages ==
;[[Scientific Python]] and [[pylab]]: a [[matlab]]-like [[python]] module that has sophisticated analysis and plotting capabilities
;[[matlab]]:
;[[Mathematica]]:
;[[R]]: R is a language and environment for statistical computing and graphics. R provides a wide variety of statistical (linear and nonlinear modelling, classical statistical tests, time-series analysis, classification, clustering, ...) and graphical techniques, and is highly extensible. One of R's strengths is the ease with which well-designed publication-quality plots can be produced, including mathematical symbols and formulae where needed. Great care has been taken over the defaults for the minor design choices in graphics, but the user retains full control.

[[Category:Analysis]]
[[Category:Software]]

Analysis software

2009-02-07T14:07:29Z

Oliver: /* MD Analysis libraries */

Running simulations is often the easy bit. The hard bit is to extract meaningful information from the Gigabytes of trajectory data. This list can act as a starting point. For most advanced uses, however, one will probably have to write analysis code in [[python]], [[Perl]], [[tcl]], [[C/C++]], [[bash]] ... or any other language that "gets the job done".

== "Native" tools ==
Many of the [[Molecular dynamics software|MD packages]] come with their own analysis tools or scripting language. Sometimes it is possible to [[convert data formats]] between packages and use the other package's analysis tools.

;[[Gromacs]] analysis tools: oone of the strengths of Gromacs is that it comes with a large number of useful analysis tools that make many of the standard analysis tasks simple to perform

;NAMD/[[VMD]]: VMD can be used through its GUI or by scripting it in [[tcl]] to great effect

;[[Charmm]]: Charmm is feature-rich but its scripting language can cause a steep learning curve

;LAMMPS/[[pizza]]: pizza.py is a python library geared towards output from [[LAMMPS]]

;Amber/[[ptraj]]: command-line based analysis

== MD Analysis libraries ==

;[[MDAnalysis]]: a python library to analyze dcd trajectories (in conjunction with a psf)

;[http://dirac.cnrs-orleans.fr/MMTK/ MMTK]: Another python-based framework for doing analysis is the ''Molecular Modelling Tool Kit''. However, it does not natively read [[Charmm]] dcd files and hence it can be cumbersome to use.

;[http://membrane.urmc.rochester.edu/Software/LOOS/Docs/main.html LOOS]: The ''Lightweight Object-Oriented Structure library (LOOS)'' from Alan Grossfield's lab provides a lightweight C++ library for analysis of molecular dynamics simulations. This includes parsing a number of PDB variants, as well as the native system description and trajectory formats for CHARMM, NAMD, and Amber. LOOS is not intended to be an all-encompassing library and it is primarily geared towards reading data in and processing rather than manipulating the files and structures and writing them out.

== Specialized tools ==

;[[HOLE]]: Oliver Smart's program to trace out pore surfaces and estimate single channel conductances.

;CAVER: [http://loschmidt.chemi.muni.cz/caver/index.php CAVER] provides rapid, accurate and fully automated calculation of pathways leading from buried cavities to outside solvent in static and dynamic protein structures. Calculated pathways can be visualized by graphic program PyMol dissecting anatomy and dynamics of entrance tunnels. CAVER allows analysis of any molecular structure including proteins, nucleic acids, inorganic materials, etc. CAVER is available as [http://loschmidt.chemi.muni.cz/caver/online.php online version] or [[PyMol]] plugin suitable for calculation of pathways in discrete protein structures and stand alone version enabling analysis of trajectories from the molecular dynamics simulations.

;[http://swift.cmbi.ru.nl/gv/dssp/ dssp]: ''Definition of secondary structure of proteins given a set of 3D coordinates.'' The DSSP program defines secondary structure, geometrical features and solvent exposure of proteins, given atomic coordinates in Protein Data Bank format. The program does NOT PREDICT protein structure. According to the Science Citation Index (July 1995), the program has been cited in the scientific literature more than 1000 times.

; [http://www.compbio.dundee.ac.uk/manuals/stamp.4.2/stamp.html STAMP]: ''Structural Alignment of Multiple Proteins''. STAMP is a package for the alignment of protein sequence based on three-dimensional (3D) structure. It provides not only multiple alignments and the corresponding `best-fit' superimpositions, but also a systematic and reproducible method for assessing the quality of such alignments. It also provides a method for protein 3D structure data base scanning. In addition to structure comparison, the STAMP package provides input for programs to display and analyse protein sequence alignments and tertiary structures. Please note that, although STAMP outputs a sequence alignment, it is a program for 3D structures, and NOT sequences.

;[[swinker]]: finds and calculates helix hinges. It optionally finds the hinge point and calculates kink and swivel angles.

== General purpose mathematical packages ==
;[[Scientific Python]] and [[pylab]]: a [[matlab]]-like [[python]] module that has sophisticated analysis and plotting capabilities
;[[matlab]]:
;[[Mathematica]]:
;[[R]]: R is a language and environment for statistical computing and graphics. R provides a wide variety of statistical (linear and nonlinear modelling, classical statistical tests, time-series analysis, classification, clustering, ...) and graphical techniques, and is highly extensible. One of R's strengths is the ease with which well-designed publication-quality plots can be produced, including mathematical symbols and formulae where needed. Great care has been taken over the defaults for the minor design choices in graphics, but the user retains full control.

[[Category:Analysis]]
[[Category:Software]]

Analysis software

2009-02-07T14:07:06Z

Oliver: /* MD Analysis libraries */ LOOS

Running simulations is often the easy bit. The hard bit is to extract meaningful information from the Gigabytes of trajectory data. This list can act as a starting point. For most advanced uses, however, one will probably have to write analysis code in [[python]], [[Perl]], [[tcl]], [[C/C++]], [[bash]] ... or any other language that "gets the job done".

== "Native" tools ==
Many of the [[Molecular dynamics software|MD packages]] come with their own analysis tools or scripting language. Sometimes it is possible to [[convert data formats]] between packages and use the other package's analysis tools.

;[[Gromacs]] analysis tools: oone of the strengths of Gromacs is that it comes with a large number of useful analysis tools that make many of the standard analysis tasks simple to perform

;NAMD/[[VMD]]: VMD can be used through its GUI or by scripting it in [[tcl]] to great effect

;[[Charmm]]: Charmm is feature-rich but its scripting language can cause a steep learning curve

;LAMMPS/[[pizza]]: pizza.py is a python library geared towards output from [[LAMMPS]]

;Amber/[[ptraj]]: command-line based analysis

== MD Analysis libraries ==

;[[MDAnalysis]]: a python library to analyze dcd trajectories (in conjunction with a psf)

;[http://dirac.cnrs-orleans.fr/MMTK/ MMTK]: Another python-based framework for doing analysis is the ''Molecular Modelling Tool Kit''. However, it does not natively read [[Charmm]] dcd files and hence it can be cumbersome to use.

;[http://membrane.urmc.rochester.edu/Software/LOOS/Docs/main.html LOOS}: The ''Lightweight Object-Oriented Structure library (LOOS)'' from Alan Grossfield's lab provides a lightweight C++ library for analysis of molecular dynamics simulations. This includes parsing a number of PDB variants, as well as the native system description and trajectory formats for CHARMM, NAMD, and Amber. LOOS is not intended to be an all-encompassing library and it is primarily geared towards reading data in and processing rather than manipulating the files and structures and writing them out.

== Specialized tools ==

;[[HOLE]]: Oliver Smart's program to trace out pore surfaces and estimate single channel conductances.

;CAVER: [http://loschmidt.chemi.muni.cz/caver/index.php CAVER] provides rapid, accurate and fully automated calculation of pathways leading from buried cavities to outside solvent in static and dynamic protein structures. Calculated pathways can be visualized by graphic program PyMol dissecting anatomy and dynamics of entrance tunnels. CAVER allows analysis of any molecular structure including proteins, nucleic acids, inorganic materials, etc. CAVER is available as [http://loschmidt.chemi.muni.cz/caver/online.php online version] or [[PyMol]] plugin suitable for calculation of pathways in discrete protein structures and stand alone version enabling analysis of trajectories from the molecular dynamics simulations.

;[http://swift.cmbi.ru.nl/gv/dssp/ dssp]: ''Definition of secondary structure of proteins given a set of 3D coordinates.'' The DSSP program defines secondary structure, geometrical features and solvent exposure of proteins, given atomic coordinates in Protein Data Bank format. The program does NOT PREDICT protein structure. According to the Science Citation Index (July 1995), the program has been cited in the scientific literature more than 1000 times.

; [http://www.compbio.dundee.ac.uk/manuals/stamp.4.2/stamp.html STAMP]: ''Structural Alignment of Multiple Proteins''. STAMP is a package for the alignment of protein sequence based on three-dimensional (3D) structure. It provides not only multiple alignments and the corresponding `best-fit' superimpositions, but also a systematic and reproducible method for assessing the quality of such alignments. It also provides a method for protein 3D structure data base scanning. In addition to structure comparison, the STAMP package provides input for programs to display and analyse protein sequence alignments and tertiary structures. Please note that, although STAMP outputs a sequence alignment, it is a program for 3D structures, and NOT sequences.

;[[swinker]]: finds and calculates helix hinges. It optionally finds the hinge point and calculates kink and swivel angles.

== General purpose mathematical packages ==
;[[Scientific Python]] and [[pylab]]: a [[matlab]]-like [[python]] module that has sophisticated analysis and plotting capabilities
;[[matlab]]:
;[[Mathematica]]:
;[[R]]: R is a language and environment for statistical computing and graphics. R provides a wide variety of statistical (linear and nonlinear modelling, classical statistical tests, time-series analysis, classification, clustering, ...) and graphical techniques, and is highly extensible. One of R's strengths is the ease with which well-designed publication-quality plots can be produced, including mathematical symbols and formulae where needed. Great care has been taken over the defaults for the minor design choices in graphics, but the user retains full control.

[[Category:Analysis]]
[[Category:Software]]

Help:Editing

2008-12-08T11:02:30Z

Oliver: /* Styles */

This page should tell you how to work with this Wiki – it's not complicated, promise! (In fact, the whole point of a wiki is to make it as easy for the user as possible to document whatever there is worth documenting).

==Editing an existing page==
# Select the ''Edit'' link at the top.
# Type or simply copy and paste text, eg from your editor, shell, or an email, and use the ''preview'' button.
# Once it looks sort-of right commit your changes by clicking ''save page''.

The Media Wiki page has a comprehensive list of the [http://meta.wikipedia.org/wiki/MediaWiki_User%27s_Guide:_Editing_overview#The_wiki_markup Wiki markup] but you can also look at the code of existing pages (edit the page but don't save it) or use the formatting buttons at the top of your editing box.

The most important thing is just to put something on the page and not to worry too much about formatting. If in doubt simply leave one initial space and everything gets formatted verbatim; anything resembling an URL will be turned into a link (or enclose it in <nowiki>'[' and ']'</nowiki>).

==Editing a new page==
If you click on a red link then you enter a non-existing page. This is not bad: Simply start editing it and ''write it yourself''. You can't make mistakes. Just do it.

==Creating a new page==
To make a non-existing page you simply insert the link into an existing page, it appears in red, you click it and then edit it.

== Signing your name ==
Sometimes this is useful: three tildes <nowiki>~~~</nowiki> signs your name like this: [[User:Oliver|Oliver]]; four <nowiki>~~~~</nowiki> dates it too: [[User:Oliver|Oliver]] 18:32, 24 January 2006 (EST)

== Categories ==
[http://meta.wikimedia.org/wiki/Help:Category '''Categories''' in MediaWiki] provide automatic indexes that are useful as tables of contents. See the Special Page '''[[Special:Categories]]''' for a list of all defined categories.

* You '''define a category''' by adding one or more special tags at the end of a page: add <tt><nowiki>[[</nowiki>Category:''Category name''<nowiki>]]</nowiki></tt> to the page's wikitext source. For instance, add to page that describes analysis scripts for Charmm
<nowiki>[[Category:Hippo]]</nowiki>
<nowiki>[[Category:Analysis]]</nowiki>
: as the last two lines. This will implicitly define the categories [[:Category:Analysis]] and [[:Category:Hippo]].

* The '''category pages''' can also be edited. In addition, the wiki software adds an alphabetically sorted list of all pages in the category. This makes a category useful as an entry point into a subject.
** One ''must'' edit a category page for the wiki to create the indexed list (even if it is just an empty edit).
** It is also possible to add another category to a category page: This will turn this category in a subcategory on the other category page.

* In order to '''reference a category within a page as a normal wiki link''' (without adding the page to the category) prefix the link name with a colon. For example: <tt><nowiki>[[:</nowiki>Category:Analysis<nowiki>]]</nowiki></tt>.

== Page maintenance ==
=== Redirection ===
A 'symbolic link' to another pages is created with the [http://meta.wikimedia.org/wiki/Help:Redirection REDIRECT] command:
<nowiki>#REDIRECT [[</nowiki>''page''<nowiki>]]</nowiki>

==Highlighting sourcecode==
Wikimedia can do syntax highlighting for over a dozen programming languages, including c, python, perl, fortran, and c++. Just surround the code with a tag of the name of the language. For example:
<pre>
<python>
import re

lines = file("ifconf.log").readlines()
l_iter = iter(lines)

nodes = []
for l in l_iter:
if l[:7] == "compute":
node = l.split('.')[0]
temp = l_iter.next().split()
iface, addr = temp[0], temp[4]
nodes.append([node, iface, addr])
</python>
</pre>

Gives you
<python>
import re

lines = file("ifconf.log").readlines()
l_iter = iter(lines)

nodes = []
for l in l_iter:
if l[:7] == "compute":
node = l.split('.')[0]
temp = l_iter.next().split()
iface, addr = temp[0], temp[4]
nodes.append([node, iface, addr])
</python>

==Inserting Gnuplot graphs==
You can insert gnuplot graphs directly into mediawiki by using the <tt>gnuplot</tt> tag. For example:
<pre>
<gnuplot>
set output 'func_approx.png'
plot '-' using 1:2 t 'quadratic approximation' with linesp lt 1 lw 3, \
'-' using 1:2 t 'cubic approximation' with linesp lt 2 lw 3
1 2
2 4
3 8
4 16
e
1 3
2 9
3 27
4 81
e
</gnuplot>
</pre>

Gives you the following:
<gnuplot>
set output 'func_approx.png'
plot '-' using 1:2 t 'quadratic approximation' with linesp lt 1 lw 3, \
'-' using 1:2 t 'cubic approximation' with linesp lt 2 lw 3
1 2
2 4
3 8
4 16
e
1 3
2 9
3 27
4 81
e
</gnuplot>

== Adding bibliographic references (specifically [http://www.pubmed.gov PubMed]) ==
You can insert bibliographic references into pages by using the <tt><nowiki><cite></nowiki></tt> and <tt><nowiki><biblio></nowiki></tt> tags. This uses the PubMed id number (pmid) found at the bottom of the abstract listing for a particular article. Mediawiki will go to [http://www.pubmed.gov PubMed] and pull the citation information for the reference.
Using the key in front of the pmid assignment within the <tt><nowiki><cite></nowiki></tt> tag (see example below) gives you a reference to the article within the current page. For example, although you can obmit the pmid and simply format the reference yourself.

<pre>
Recent papers from the Woolf lab <cite>jcp2005 proteins2005 jcp2004</cite>:
===Bibliography===
<biblio>
#jcp2005 pmid=15847458
#proteins2005 pmid=15828005
#jcp2004 pmid=15634036
</biblio>
</pre>

Will give you this:
Recent papers from the Woolf lab <cite>jcp2005 proteins2005 jcp2004</cite>:

Bibliography

<biblio>
#jcp2005 pmid=15847458
#proteins2005 pmid=15828005
#jcp2004 pmid=15634036
</biblio>

== Images ==

The following shows two methods of how to incorporate an image in a wiki page.

=== Upload ===
The image is uploaded to the webserver and resides in the wiki (somewhere... you don't need to know where, it's a secret)

# first [[Special:Upload|upload]] it (use the link in the toolbox on the left)
# enter a wiki link such as <tt><nowiki>[[Image:Gbim.jpg|200px]]
</nowiki></tt> into the text.

[[Image:Gbim.jpg|200px]]

[[Image:Gbim.jpg|thumb|right|150px|Insertion of a peptide into a Generalized Born implicit membrane.]]

An exhaustive description of the image capabilities are discussed in the [http://en.wikipedia.org/wiki/Wikipedia:Extended_image_syntax Wikipedia Extended Image Syntax]. Most importantly, if you want to change the size of the image, add a size option <tt>|''size''px</tt> option (<tt>|200px</tt> in the example above).

The new version of MediaWiki gives you a plethora of options to add captions, float the image to left or right, change sizes, show it as a (fast) thumbnail, ...
<tt><nowiki>[[Image:Gbim.jpg|thumb|right|150px|Insertion of a peptide into a Generalized Born implicit membrane.]]</nowiki></tt>
gives a thumbnail image floating on the right hand side.

== Styles ==
One can directly edit the [http://www.w3.org/Style/CSS/ Cascading Style Sheets] (CSS). The most important one is [[Mediawiki:Common.css]] – be careful, all changes directly affect every user.

See the [http://www.mediawiki.org/wiki/Manual:Interface/Stylesheets Mediawiki Stylesheet Manual] for details and there is also the useful [http://www.mediawiki.org/wiki/Manual:FAQ Mediawiki FAQ] on customizing the Interface. [http://en.wikipedia.org/wiki/Help:User_style Wikipedia's Help:User_style] has a lot of howto information.

== Links ==
If you want to gain an in-depth knowledge of working with this Wiki then have a look at these links:

=== Wikipedia edit help documents ===
* Wikipedia's [http://en.wikipedia.org/wiki/Wikipedia:How_to_edit_a_page full listing of wiki editing commands] (make sure you come back here to edit... otherwise you will be editing Wikipedia)
* Wikipedia [http://en.wikipedia.org/wiki/Wikipedia:Extended_image_syntax Extended Image Syntax] for all your image inclusion needs
* Syntax of [http://en.wikipedia.org/wiki/Help:Table Table] commands

=== MediaWiki documents ===
The [http://wiki.biowerkzeug.org/ Biowerkzeug Wiki] uses the [http://www.mediawiki.org/wiki MediaWiki] software. For more on MediaWiki see the following links:
* [http://meta.wikimedia.org/wiki/Help:Contents User's Guide]
* [http://www.mediawiki.org/wiki/Manual:Configuration_settings Configuration settings list]
* [http://www.mediawiki.org/wiki/Manual:FAQ MediaWiki FAQ]
* [http://lists.wikimedia.org/mailman/listinfo/mediawiki-announce MediaWiki release mailing list]

[[Category:Help]]

Performance

2008-12-08T11:00:46Z

Oliver: /* Single processor performance */

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
<li>The timings have errorbars of about ±2s</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|+ Benchmark (single core) and [[#Scaling|scaling]] results on multiple cores/cpus. '''cores''' indicates how many cores are available to the user on this cpu ''or'' the maximum number of cores on the node that were used for benchmarking scaling.
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
! scaling
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:05
| hippo
| rev35 
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 8 = 2x4
| 1:16
| hippo
| rev35 
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:37
| hippo_p3
| rev35 
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

In the table below, '''# cpus''' really means how many independent cores are available.
{| class="wikitable sortable"
|+ Scaling of the Hippo WALP test case using OpenMP on single-board machines
! vendor
! # cpus
! walltime and scaling
|-
| Intel
| 4
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| 8
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| AMD
| 4
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|}

Performance

2008-12-08T10:56:35Z

Oliver: /* Single processor performance */

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
<li>The timings have errorbars of about ±2s</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|+ Benchmark (single core) and [[#Scaling|scaling]] results on multiple cores/cpus.
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
! scaling
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:05
| hippo
| rev35 
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 8 = 2x4
| 1:16
| hippo
| rev35 
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:37
| hippo_p3
| rev35 
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

In the table below, '''# cpus''' really means how many independent cores are available.
{| class="wikitable sortable"
|+ Scaling of the Hippo WALP test case using OpenMP on single-board machines
! vendor
! # cpus
! walltime and scaling
|-
| Intel
| 4
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| 8
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| AMD
| 4
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|}

Performance

2008-12-08T10:54:38Z

Oliver: /* Single processor performance */

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
<li>The timings have errorbars of about ±2s</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|+ Benchmark (single core) and [[#Scaling|scaling]] results on multiple cores/cpus.
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
! scaling
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:05
| hippo
| rev35 
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:37
| hippo_p3
| rev35 
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

In the table below, '''# cpus''' really means how many independent cores are available.
{| class="wikitable sortable"
|+ Scaling of the Hippo WALP test case using OpenMP on single-board machines
! vendor
! # cpus
! walltime and scaling
|-
| Intel
| 4
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| 8
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| AMD
| 4
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|}

Performance

2008-12-08T10:53:27Z

Oliver: scaling in main table

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|+ Benchmark (single core) and [[#Scaling|scaling]] results on multiple cores/cpus.
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
! scaling
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:05
| hippo
| rev35 
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:37
| hippo_p3
| rev35 
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

In the table below, '''# cpus''' really means how many independent cores are available.
{| class="wikitable sortable"
|+ Scaling of the Hippo WALP test case using OpenMP on single-board machines
! vendor
! # cpus
! walltime and scaling
|-
| Intel
| 4
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| 8
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| AMD
| 4
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|}

MediaWiki:Common.css

2008-12-08T10:37:17Z

Oliver: table left aligned txt

/* CSS placed here will be applied to all skins */

/* Copied from http://de.wikipedia.org/wiki/MediaWiki:Common.css :
"+++++ 3. NEUE [[Hauptseite|HAUPTSEITE]] (ab 2006) +++++"
*/

/* Kategorie verbergen */
body.page-Hauptseite #catlinks {
display:none;
}

#hauptseite h2 {
background-color: #d8e8ff;
border: 1px solid #8898bf;
font-size: 1em;
font-weight: bold;
margin-top: 0;
margin-bottom: 0;
padding-top: 0.1em;
padding-bottom: 0.1em;
}
#hauptseite .inhalt {
background-color: #ffffff;
border: 1px solid #8898bf;
border-top: 0px solid white;
padding: 0.3em 0.8em 0.4em 0.8em;
}
#hauptseite .inhalt hr {
background-color: #8898bf;
color: #8898bf;
height: 1px;
margin:0.5em 0;
padding: 0;
}
#hauptseite .inhalt .mehr {
clear: both;
font-size: 95%;
margin-top: 0.8em;
text-align: right;
}
.hauptseite-oben,
.hauptseite-links,
.hauptseite-rechts {
margin-bottom: 1em;
}
.hauptseite-links {
margin-right: 0.5em;
}
.hauptseite-rechts {
margin-left: 0.5em;
}
.hauptseite-oben h2,
.hauptseite-unten h2 {
text-align: center;
}
.hauptseite-oben .inhalt .portale {
font-weight: bold;
margin-top: 0.2em;
margin-bottom: 0.2em;
}
.hauptseite-oben .inhalt .intern {
font-size: 90%;
text-align: center;
}
.hauptseite-links h2,
.hauptseite-rechts h2 {
text-indent: 0.8em;
}
#hauptseite-schwesterprojekte .inhalt a {
font-weight: bold;
}

/* p.catlinks span a[href*="/wiki/Kategorie:Arbeitskategorie"] { display:none } wegen HIDDENCAT nicht mehr notwendig */

/* Remove underline from IPA links */
.IPA a:link, .IPA a:visited {
text-decoration: none;
}

span.Unicode
{
font-family:
'Code2000',
'Sun-ExtA',
'Arial Unicode MS',
'NSimSun',
sans-serif;
}

span.Unicode1
{
font-family:
'Code2001',
'Quivira',
'MPH 2B Damase',
sans-serif;
}

span.Unicode2
{
font-family:
'Sun-ExtB',
'Code2002',
sans-serif;
}

span.IPA
{
font-family:
'Quivira',
'Code2000',
'Sun-ExtA',
'DejaVu Sans',
'Gentium',
'Arial Unicode MS',
'Lucida Sans Unicode',
sans-serif;
}

span.IAST
{
font-family:
'Code2000',
'SunExtA',
'Arial Unicode MS',
sans-serif;
}

span.altitalisch
{
font-family:
'Quivira',
'Code2001',
'MPH 2B Damase',
sans-serif;
}

span.gotisch
{
font-family:
'Quivira',
'Code2001',
'MPH 2B Damase',
sans-serif;
}

span.hebrew
{
font-family:
'Quivira',
'Sun-ExtA',
'Arial Unicode MS',
'SBL Hebrew',
'Code2000',
'MPH 2B Damase',
sans-serif;
}

span.spanAr
{
font-family:
'Arial Unicode MS',
'Code2000',
'MPH 2B Damase',
'DejaVu Sans',
sans-serif;
}

span.music-symbol
{
font-family:
'Musical Symbols',
'Euterpe',
'Code2001',
sans-serif;
}

/* Standardmäßige Ausblendung der Flagged-Revisions-Backlog-Sitenotice */
#mw-oldreviewed-notice {
display: none;
}

/* copied from http://meta.wikimedia.org/wiki/MediaWiki:Common.css */

/*****
** Table formatting
*****/
table.wikitable,
table.prettytable {
margin:1em 1em 1em 0;
background:#F9F9F9;
border:1px #AAA solid;
border-collapse:collapse;
}

table.wikitable th, table.wikitable td,
table.prettytable th, table.prettytable td {
border:1px #AAA solid;
padding:0.2em;
}

table.wikitable th,
table.prettytable th {
background:#F2F2F2;
text-align:center;
}

table.wikitable caption,
table.prettytable caption {
margin-left:inherit;
margin-right:inherit;
border:1px #AAA solid;
background: #E0E0E0;
text-align: left;
}

/*****
** box formatting
*****/
.infobox {
float:right;
clear:right;
margin-bottom:0.5em;
margin-left:1em;
padding:0.2em;
border:1px solid #AAA;
background:#F9F9F9;
color:black;

}

.infobox td,
.infobox th {
vertical-align:top;
}

.infobox caption {
font-size:larger;
margin-left:inherit;
}

.infobox.bordered {
border-collapse:collapse;
}

.infobox.bordered td,
.infobox.bordered th {
border:1px solid #AAA;
}

.infobox.bordered .borderless td,
.infobox.bordered .borderless th {
border:0;
}

/* Change the external link icon to an Adobe icon for all PDF files */
/* (in browsers that support these CSS selectors, like Mozilla and Opera) */
#bodyContent a[href$=".pdf"].external,
#bodyContent a[href*=".pdf?"].external,
#bodyContent a[href*=".pdf#"].external,
#bodyContent a[href$=".PDF"].external,
#bodyContent a[href*=".PDF?"].external,
#bodyContent a[href*=".PDF#"].external {
background: url(http://upload.wikimedia.org/wikipedia/commons/thumb/2/23/Icons-mini-file_acrobat.gif/15px-Icons-mini-file_acrobat.gif) center right no-repeat;
padding-right: 16px;
}

/* Change the external link icon to an Adobe icon anywhere the PDFlink class */
/* is used (notably Template:PDFlink). This works in IE, unlike the above. */
span.PDFlink a {
background: url(http://upload.wikimedia.org/wikipedia/commons/thumb/2/23/Icons-mini-file_acrobat.gif/15px-Icons-mini-file_acrobat.gif) center right no-repeat !important;
padding-right: 17px !important;
}

MediaWiki:Common.css

2008-12-08T10:35:12Z

Oliver: table caption

/* CSS placed here will be applied to all skins */

/* Copied from http://de.wikipedia.org/wiki/MediaWiki:Common.css :
"+++++ 3. NEUE [[Hauptseite|HAUPTSEITE]] (ab 2006) +++++"
*/

/* Kategorie verbergen */
body.page-Hauptseite #catlinks {
display:none;
}

#hauptseite h2 {
background-color: #d8e8ff;
border: 1px solid #8898bf;
font-size: 1em;
font-weight: bold;
margin-top: 0;
margin-bottom: 0;
padding-top: 0.1em;
padding-bottom: 0.1em;
}
#hauptseite .inhalt {
background-color: #ffffff;
border: 1px solid #8898bf;
border-top: 0px solid white;
padding: 0.3em 0.8em 0.4em 0.8em;
}
#hauptseite .inhalt hr {
background-color: #8898bf;
color: #8898bf;
height: 1px;
margin:0.5em 0;
padding: 0;
}
#hauptseite .inhalt .mehr {
clear: both;
font-size: 95%;
margin-top: 0.8em;
text-align: right;
}
.hauptseite-oben,
.hauptseite-links,
.hauptseite-rechts {
margin-bottom: 1em;
}
.hauptseite-links {
margin-right: 0.5em;
}
.hauptseite-rechts {
margin-left: 0.5em;
}
.hauptseite-oben h2,
.hauptseite-unten h2 {
text-align: center;
}
.hauptseite-oben .inhalt .portale {
font-weight: bold;
margin-top: 0.2em;
margin-bottom: 0.2em;
}
.hauptseite-oben .inhalt .intern {
font-size: 90%;
text-align: center;
}
.hauptseite-links h2,
.hauptseite-rechts h2 {
text-indent: 0.8em;
}
#hauptseite-schwesterprojekte .inhalt a {
font-weight: bold;
}

/* p.catlinks span a[href*="/wiki/Kategorie:Arbeitskategorie"] { display:none } wegen HIDDENCAT nicht mehr notwendig */

/* Remove underline from IPA links */
.IPA a:link, .IPA a:visited {
text-decoration: none;
}

span.Unicode
{
font-family:
'Code2000',
'Sun-ExtA',
'Arial Unicode MS',
'NSimSun',
sans-serif;
}

span.Unicode1
{
font-family:
'Code2001',
'Quivira',
'MPH 2B Damase',
sans-serif;
}

span.Unicode2
{
font-family:
'Sun-ExtB',
'Code2002',
sans-serif;
}

span.IPA
{
font-family:
'Quivira',
'Code2000',
'Sun-ExtA',
'DejaVu Sans',
'Gentium',
'Arial Unicode MS',
'Lucida Sans Unicode',
sans-serif;
}

span.IAST
{
font-family:
'Code2000',
'SunExtA',
'Arial Unicode MS',
sans-serif;
}

span.altitalisch
{
font-family:
'Quivira',
'Code2001',
'MPH 2B Damase',
sans-serif;
}

span.gotisch
{
font-family:
'Quivira',
'Code2001',
'MPH 2B Damase',
sans-serif;
}

span.hebrew
{
font-family:
'Quivira',
'Sun-ExtA',
'Arial Unicode MS',
'SBL Hebrew',
'Code2000',
'MPH 2B Damase',
sans-serif;
}

span.spanAr
{
font-family:
'Arial Unicode MS',
'Code2000',
'MPH 2B Damase',
'DejaVu Sans',
sans-serif;
}

span.music-symbol
{
font-family:
'Musical Symbols',
'Euterpe',
'Code2001',
sans-serif;
}

/* Standardmäßige Ausblendung der Flagged-Revisions-Backlog-Sitenotice */
#mw-oldreviewed-notice {
display: none;
}

/* copied from http://meta.wikimedia.org/wiki/MediaWiki:Common.css */

/*****
** Table formatting
*****/
table.wikitable,
table.prettytable {
margin:1em 1em 1em 0;
background:#F9F9F9;
border:1px #AAA solid;
border-collapse:collapse;
}

table.wikitable th, table.wikitable td,
table.prettytable th, table.prettytable td {
border:1px #AAA solid;
padding:0.2em;
}

table.wikitable th,
table.prettytable th {
background:#F2F2F2;
text-align:center;
}

table.wikitable caption,
table.prettytable caption {
margin-left:inherit;
margin-right:inherit;
border:1px #AAA solid;
background: #F0F0F0;
}

/*****
** box formatting
*****/
.infobox {
float:right;
clear:right;
margin-bottom:0.5em;
margin-left:1em;
padding:0.2em;
border:1px solid #AAA;
background:#F9F9F9;
color:black;

}

.infobox td,
.infobox th {
vertical-align:top;
}

.infobox caption {
font-size:larger;
margin-left:inherit;
}

.infobox.bordered {
border-collapse:collapse;
}

.infobox.bordered td,
.infobox.bordered th {
border:1px solid #AAA;
}

.infobox.bordered .borderless td,
.infobox.bordered .borderless th {
border:0;
}

/* Change the external link icon to an Adobe icon for all PDF files */
/* (in browsers that support these CSS selectors, like Mozilla and Opera) */
#bodyContent a[href$=".pdf"].external,
#bodyContent a[href*=".pdf?"].external,
#bodyContent a[href*=".pdf#"].external,
#bodyContent a[href$=".PDF"].external,
#bodyContent a[href*=".PDF?"].external,
#bodyContent a[href*=".PDF#"].external {
background: url(http://upload.wikimedia.org/wikipedia/commons/thumb/2/23/Icons-mini-file_acrobat.gif/15px-Icons-mini-file_acrobat.gif) center right no-repeat;
padding-right: 16px;
}

/* Change the external link icon to an Adobe icon anywhere the PDFlink class */
/* is used (notably Template:PDFlink). This works in IE, unlike the above. */
span.PDFlink a {
background: url(http://upload.wikimedia.org/wikipedia/commons/thumb/2/23/Icons-mini-file_acrobat.gif/15px-Icons-mini-file_acrobat.gif) center right no-repeat !important;
padding-right: 17px !important;
}

Help:Editing

2008-12-08T10:30:30Z

Oliver: CSS

This page should tell you how to work with this Wiki – it's not complicated, promise! (In fact, the whole point of a wiki is to make it as easy for the user as possible to document whatever there is worth documenting).

==Editing an existing page==
# Select the ''Edit'' link at the top.
# Type or simply copy and paste text, eg from your editor, shell, or an email, and use the ''preview'' button.
# Once it looks sort-of right commit your changes by clicking ''save page''.

The Media Wiki page has a comprehensive list of the [http://meta.wikipedia.org/wiki/MediaWiki_User%27s_Guide:_Editing_overview#The_wiki_markup Wiki markup] but you can also look at the code of existing pages (edit the page but don't save it) or use the formatting buttons at the top of your editing box.

The most important thing is just to put something on the page and not to worry too much about formatting. If in doubt simply leave one initial space and everything gets formatted verbatim; anything resembling an URL will be turned into a link (or enclose it in <nowiki>'[' and ']'</nowiki>).

==Editing a new page==
If you click on a red link then you enter a non-existing page. This is not bad: Simply start editing it and ''write it yourself''. You can't make mistakes. Just do it.

==Creating a new page==
To make a non-existing page you simply insert the link into an existing page, it appears in red, you click it and then edit it.

== Signing your name ==
Sometimes this is useful: three tildes <nowiki>~~~</nowiki> signs your name like this: [[User:Oliver|Oliver]]; four <nowiki>~~~~</nowiki> dates it too: [[User:Oliver|Oliver]] 18:32, 24 January 2006 (EST)

== Categories ==
[http://meta.wikimedia.org/wiki/Help:Category '''Categories''' in MediaWiki] provide automatic indexes that are useful as tables of contents. See the Special Page '''[[Special:Categories]]''' for a list of all defined categories.

* You '''define a category''' by adding one or more special tags at the end of a page: add <tt><nowiki>[[</nowiki>Category:''Category name''<nowiki>]]</nowiki></tt> to the page's wikitext source. For instance, add to page that describes analysis scripts for Charmm
<nowiki>[[Category:Hippo]]</nowiki>
<nowiki>[[Category:Analysis]]</nowiki>
: as the last two lines. This will implicitly define the categories [[:Category:Analysis]] and [[:Category:Hippo]].

* The '''category pages''' can also be edited. In addition, the wiki software adds an alphabetically sorted list of all pages in the category. This makes a category useful as an entry point into a subject.
** One ''must'' edit a category page for the wiki to create the indexed list (even if it is just an empty edit).
** It is also possible to add another category to a category page: This will turn this category in a subcategory on the other category page.

* In order to '''reference a category within a page as a normal wiki link''' (without adding the page to the category) prefix the link name with a colon. For example: <tt><nowiki>[[:</nowiki>Category:Analysis<nowiki>]]</nowiki></tt>.

== Page maintenance ==
=== Redirection ===
A 'symbolic link' to another pages is created with the [http://meta.wikimedia.org/wiki/Help:Redirection REDIRECT] command:
<nowiki>#REDIRECT [[</nowiki>''page''<nowiki>]]</nowiki>

==Highlighting sourcecode==
Wikimedia can do syntax highlighting for over a dozen programming languages, including c, python, perl, fortran, and c++. Just surround the code with a tag of the name of the language. For example:
<pre>
<python>
import re

lines = file("ifconf.log").readlines()
l_iter = iter(lines)

nodes = []
for l in l_iter:
if l[:7] == "compute":
node = l.split('.')[0]
temp = l_iter.next().split()
iface, addr = temp[0], temp[4]
nodes.append([node, iface, addr])
</python>
</pre>

Gives you
<python>
import re

lines = file("ifconf.log").readlines()
l_iter = iter(lines)

nodes = []
for l in l_iter:
if l[:7] == "compute":
node = l.split('.')[0]
temp = l_iter.next().split()
iface, addr = temp[0], temp[4]
nodes.append([node, iface, addr])
</python>

==Inserting Gnuplot graphs==
You can insert gnuplot graphs directly into mediawiki by using the <tt>gnuplot</tt> tag. For example:
<pre>
<gnuplot>
set output 'func_approx.png'
plot '-' using 1:2 t 'quadratic approximation' with linesp lt 1 lw 3, \
'-' using 1:2 t 'cubic approximation' with linesp lt 2 lw 3
1 2
2 4
3 8
4 16
e
1 3
2 9
3 27
4 81
e
</gnuplot>
</pre>

Gives you the following:
<gnuplot>
set output 'func_approx.png'
plot '-' using 1:2 t 'quadratic approximation' with linesp lt 1 lw 3, \
'-' using 1:2 t 'cubic approximation' with linesp lt 2 lw 3
1 2
2 4
3 8
4 16
e
1 3
2 9
3 27
4 81
e
</gnuplot>

== Adding bibliographic references (specifically [http://www.pubmed.gov PubMed]) ==
You can insert bibliographic references into pages by using the <tt><nowiki><cite></nowiki></tt> and <tt><nowiki><biblio></nowiki></tt> tags. This uses the PubMed id number (pmid) found at the bottom of the abstract listing for a particular article. Mediawiki will go to [http://www.pubmed.gov PubMed] and pull the citation information for the reference.
Using the key in front of the pmid assignment within the <tt><nowiki><cite></nowiki></tt> tag (see example below) gives you a reference to the article within the current page. For example, although you can obmit the pmid and simply format the reference yourself.

<pre>
Recent papers from the Woolf lab <cite>jcp2005 proteins2005 jcp2004</cite>:
===Bibliography===
<biblio>
#jcp2005 pmid=15847458
#proteins2005 pmid=15828005
#jcp2004 pmid=15634036
</biblio>
</pre>

Will give you this:
Recent papers from the Woolf lab <cite>jcp2005 proteins2005 jcp2004</cite>:

Bibliography

<biblio>
#jcp2005 pmid=15847458
#proteins2005 pmid=15828005
#jcp2004 pmid=15634036
</biblio>

== Images ==

The following shows two methods of how to incorporate an image in a wiki page.

=== Upload ===
The image is uploaded to the webserver and resides in the wiki (somewhere... you don't need to know where, it's a secret)

# first [[Special:Upload|upload]] it (use the link in the toolbox on the left)
# enter a wiki link such as <tt><nowiki>[[Image:Gbim.jpg|200px]]
</nowiki></tt> into the text.

[[Image:Gbim.jpg|200px]]

[[Image:Gbim.jpg|thumb|right|150px|Insertion of a peptide into a Generalized Born implicit membrane.]]

An exhaustive description of the image capabilities are discussed in the [http://en.wikipedia.org/wiki/Wikipedia:Extended_image_syntax Wikipedia Extended Image Syntax]. Most importantly, if you want to change the size of the image, add a size option <tt>|''size''px</tt> option (<tt>|200px</tt> in the example above).

The new version of MediaWiki gives you a plethora of options to add captions, float the image to left or right, change sizes, show it as a (fast) thumbnail, ...
<tt><nowiki>[[Image:Gbim.jpg|thumb|right|150px|Insertion of a peptide into a Generalized Born implicit membrane.]]</nowiki></tt>
gives a thumbnail image floating on the right hand side.

== Styles ==
One can directly edit the [http://www.w3.org/Style/CSS/ Cascading Style Sheets] (CSS). The most important one is [[Mediawiki:Common.css]] – be careful, all changes directly affect every user.

See the [http://www.mediawiki.org/wiki/Manual:Interface/Stylesheets Mediawiki Stylesheet Manual] for details and there is also the useful [http://www.mediawiki.org/wiki/Manual:FAQ Mediawiki FAQ] on customizing the Interface.

== Links ==
If you want to gain an in-depth knowledge of working with this Wiki then have a look at these links:

=== Wikipedia edit help documents ===
* Wikipedia's [http://en.wikipedia.org/wiki/Wikipedia:How_to_edit_a_page full listing of wiki editing commands] (make sure you come back here to edit... otherwise you will be editing Wikipedia)
* Wikipedia [http://en.wikipedia.org/wiki/Wikipedia:Extended_image_syntax Extended Image Syntax] for all your image inclusion needs
* Syntax of [http://en.wikipedia.org/wiki/Help:Table Table] commands

=== MediaWiki documents ===
The [http://wiki.biowerkzeug.org/ Biowerkzeug Wiki] uses the [http://www.mediawiki.org/wiki MediaWiki] software. For more on MediaWiki see the following links:
* [http://meta.wikimedia.org/wiki/Help:Contents User's Guide]
* [http://www.mediawiki.org/wiki/Manual:Configuration_settings Configuration settings list]
* [http://www.mediawiki.org/wiki/Manual:FAQ MediaWiki FAQ]
* [http://lists.wikimedia.org/mailman/listinfo/mediawiki-announce MediaWiki release mailing list]

[[Category:Help]]

Performance

2008-12-08T10:23:26Z

Oliver: /* Scaling */

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

In the table below, '''# cpus''' really means how many independent cores are available.
{| class="wikitable sortable"
|+ Scaling of the Hippo WALP test case using OpenMP on single-board machines
! vendor
! # cpus
! walltime and scaling
|-
| Intel
| 4
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| 8
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| AMD
| 4
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|}

Performance

2008-12-08T10:21:08Z

Oliver: /* Scaling */

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

{| class="wikitable sortable"
! vendor
! # cpus or cores
! walltime and scaling
|-
| Intel
| 4
| [[Image:Scaling Q9550.png|thumb|none|Intel Quad Core Q9550 2.8 GHz]]
|-
| Intel
| 8
| [[Image:Scaling E5420.png|thumb|none|Dual Intel Quad Core Xeon E5420 2.5 GHz]]
|-
| AMD
| 4
| [[Image:Scaling AMD X4 9840.png|thumb|none|AMD Phenom X4 9850 Quad Core 2.5 GHz]]
|}

Performance

2008-12-07T21:15:37Z

Oliver: fixed Phenom GHz

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.5
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

[[Image:Scaling Q9550.png|thumb|400px|left|Intel Quad Core Q9550 2.8 GHz]]
[[Image:Scaling E5420.png|thumb|400px|right|Dual Intel Quad Core Xeon E5420 2.5 GHz]]

[[Image:Scaling AMD X4 9840.png|thumb|400px|left|AMD Phenom X4 9850 Quad Core 2.5 GHz]]

Performance

2008-12-07T21:15:08Z

Oliver: /* Scaling */ Phenom

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.75
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

[[Image:Scaling Q9550.png|thumb|400px|left|Intel Quad Core Q9550 2.8 GHz]]
[[Image:Scaling E5420.png|thumb|400px|right|Dual Intel Quad Core Xeon E5420 2.5 GHz]]

[[Image:Scaling AMD X4 9840.png|thumb|400px|left|AMD Phenom X4 9850 Quad Core 2.5 GHz]]

File:Scaling AMD X4 9840.png

2008-12-07T21:12:29Z

Oliver: Scaling of Hippo rev 35 walp test on AMD Phenom X4 9840

Scaling of Hippo rev 35 walp test on AMD Phenom X4 9840

Calc testjobs linux.sh

2008-12-07T16:13:33Z

Oliver:

This script allows running Hippo benchmarks without having to think too much about finding the correct binary and supplemental files. In addition it gives the total wall time required to run each test case; this can be used for crude benchmarks.

* The testing of the mpi binaries is not tested and probably does not work. However, they are ''only'' needed for replica exchange simulations and there are no test cases for those anyway.
* [http://forums.biowerkzeug.org/viewtopic.php?f=22&t=6&p=15&hilit=OpenMP#p15 Not all tests support multi-threaded runs] (with OpenMP); these tests ignore the NSLOTS argument.

<pre>#!/bin/bash
# $Id: calc_testjobs_linux.sh 2535 2008-12-07 16:07:19Z www-data $
# Running Hippo tests (Linux)
# Copyright (c) 2008 Biowerkzeug
# Oliver Beckstein <orbeckst@gmail.com>
#set -x

prog=$(basename $0)
CURDIR=${PWD}

# defaults (: can be set in environment)
: ${HIPPO_DIR:="${CURDIR}/.."}
HIPPO_TESTS="hexane_NVT_dp_MD octane_NPT_sp_MC pentadecane_NPT_sp_MD tip3p_NPT_sp_MD trpzip2_GBSA_MC vpu_GBIM_MC walp_octane_NPT_sp_MD"
RUN_TESTS=${HIPPO_TESTS}
USE_MPI=0

usage="usage $prog [opts] [tests]

Run Hippo tests. By default it runs all of them:
${HIPPO_TESTS}

OPTIONS:
-h help
-n number of threads (not possible for all tests)
-D directory where we look for Hippo executables [${HIPPO_DIR}]
-M use mpi (replica exchange) binaries [${USE_MPI}]

Environment variables:
HIPPO_DIR overrides -D [${HIPPO_DIR}]
"

function die () {
local msg="$1" err=${2:-1}
echo 1>&2 "ERROR: failed in $PWD: ${msg}"
cd ${CURDIR}
exit $err
}

NSLOTS=1
# opt processing
while getopts hn:D:M: opt; do
case $opt in
h) echo "$usage"; exit 0;;
n) NSLOTS=${OPTARG};;
D) HIPPO_DIR=${OPTARG};;
M) USE_MPI=${OPTARG};;
*) die "Unknown option" 1;;
esac
done

#echo "OPTIND=$OPTIND OPTARG=$OPTARG argv=$*"
shift $((OPTIND - 1))
if [ -n "$*" ]; then
RUN_TESTS="$*"
fi

echo "Running the following tests using ${NSLOTS} threads: ${RUN_TESTS}"

# find working executable
# we'll use the first one that only complain about missing input file
#
if [ ${USE_MPI} = 0 ]; then
echo "Trying standard binaries (with OpenMP)"
_HIPPO_BINARIES="hippo hippo_p3"
else
echo "Testing mpi binaries. Note: these are ONLY needed for replica exchange simulations"
_HIPPO_BINARIES="hippo_mpi hippo_p3_mpi"
fi
HIPPO="not_found"
rm -f hippo_input.txt # clean any input files
for h in ${_HIPPO_BINARIES}; do
exe="${HIPPO_DIR}/${h}"
if ${exe} 2>&1 | egrep "^Hippo.*Copyright.*Biowerkzeug" >/dev/null; then
HIPPO=${exe}
break
fi
done
if [ "${HIPPO}" = "not_found" ]; then
ARCH=$(uname -m);
OS=$(uname -s);
die "No usable hippo executable found; see if there is one at
http://www.biowerkzeug.com for your architecture ${ARCH} and operating
system ${OS}. "
fi

echo "Using executable ${HIPPO}"

TOPOLOGY=${HIPPO_DIR}/hippo_protein_database.dat
FF=${HIPPO_DIR}/oplsaa_forcefield.dat

echo "Setting up test directory"
rm -rf test
mkdir test
cd test

topdir="${CURDIR}/test"

function setup_hippo () {
local numthreads=${1:-1}
local input=hippo_input.txt
cp ${HIPPO} ./hippo || return $?
cp ${TOPOLOGY} . || return $?
cp ${FF} . || return $?
test -e $input || die "Missing run input file $input in $PWD"
if [ $NSLOTS -gt 1 ]; then
# adjusting for OpenMP run
sed -i.orig -e "s/[[:space:]]*openMP numthreads.*/openMP numthreads ${numthreads}/" $input
fi
return 0
}

function run_test () {
local testdir="$1" numthreads="${2:-1}"
echo "---------------------------------------------------------"
cd ${testdir} || die "Cannot 'cd ${testdir}'"
setup_hippo ${numthreads} || die "setup_hippo() failed"
echo "Set up all files for NSLOTS=${numthreads}"
echo "Running hippo test case ${testdir}..."
t_start=$(date +%s)
./hippo
t_stop=$(date +%s)
delta_t=$(( t_stop-t_start ))
echo "Completed hippo test case ${testdir} in ${delta_t} seconds, running ${numthreads} threads"
echo "BENCHMARK: ${testdir} ${numthreads} ${delta_t}"
cd ${topdir}
}

cp -r ../jobs/* .

for t in ${RUN_TESTS};
do run_test $t ${NSLOTS}
done

echo "Finished running hippo test suite"
</pre>

Calc testjobs linux.sh

2008-12-07T16:13:14Z

Oliver:

This script allows running Hippo benchmarks without having to think too much about finding the correct binary and supplemental files. In addition it gives the total wall time required to run each test case; this can be used for crude benchmarks.

* The testing of the mpi binaries is not tested and probably does not work. However, they are ''only'' needed for replica exchange simulations and there are no test cases for those anyway.
* [http://forums.biowerkzeug.org/viewtopic.php?f=22&t=6&p=15&hilit=OpenMP#p15 Not all tests support multi-threaded runs] (with Opn)n; these tests ignore the NSLOTS argument.

<pre>#!/bin/bash
# $Id: calc_testjobs_linux.sh 2535 2008-12-07 16:07:19Z www-data $
# Running Hippo tests (Linux)
# Copyright (c) 2008 Biowerkzeug
# Oliver Beckstein <orbeckst@gmail.com>
#set -x

prog=$(basename $0)
CURDIR=${PWD}

# defaults (: can be set in environment)
: ${HIPPO_DIR:="${CURDIR}/.."}
HIPPO_TESTS="hexane_NVT_dp_MD octane_NPT_sp_MC pentadecane_NPT_sp_MD tip3p_NPT_sp_MD trpzip2_GBSA_MC vpu_GBIM_MC walp_octane_NPT_sp_MD"
RUN_TESTS=${HIPPO_TESTS}
USE_MPI=0

usage="usage $prog [opts] [tests]

Run Hippo tests. By default it runs all of them:
${HIPPO_TESTS}

OPTIONS:
-h help
-n number of threads (not possible for all tests)
-D directory where we look for Hippo executables [${HIPPO_DIR}]
-M use mpi (replica exchange) binaries [${USE_MPI}]

Environment variables:
HIPPO_DIR overrides -D [${HIPPO_DIR}]
"

function die () {
local msg="$1" err=${2:-1}
echo 1>&2 "ERROR: failed in $PWD: ${msg}"
cd ${CURDIR}
exit $err
}

NSLOTS=1
# opt processing
while getopts hn:D:M: opt; do
case $opt in
h) echo "$usage"; exit 0;;
n) NSLOTS=${OPTARG};;
D) HIPPO_DIR=${OPTARG};;
M) USE_MPI=${OPTARG};;
*) die "Unknown option" 1;;
esac
done

#echo "OPTIND=$OPTIND OPTARG=$OPTARG argv=$*"
shift $((OPTIND - 1))
if [ -n "$*" ]; then
RUN_TESTS="$*"
fi

echo "Running the following tests using ${NSLOTS} threads: ${RUN_TESTS}"

# find working executable
# we'll use the first one that only complain about missing input file
#
if [ ${USE_MPI} = 0 ]; then
echo "Trying standard binaries (with OpenMP)"
_HIPPO_BINARIES="hippo hippo_p3"
else
echo "Testing mpi binaries. Note: these are ONLY needed for replica exchange simulations"
_HIPPO_BINARIES="hippo_mpi hippo_p3_mpi"
fi
HIPPO="not_found"
rm -f hippo_input.txt # clean any input files
for h in ${_HIPPO_BINARIES}; do
exe="${HIPPO_DIR}/${h}"
if ${exe} 2>&1 | egrep "^Hippo.*Copyright.*Biowerkzeug" >/dev/null; then
HIPPO=${exe}
break
fi
done
if [ "${HIPPO}" = "not_found" ]; then
ARCH=$(uname -m);
OS=$(uname -s);
die "No usable hippo executable found; see if there is one at
http://www.biowerkzeug.com for your architecture ${ARCH} and operating
system ${OS}. "
fi

echo "Using executable ${HIPPO}"

TOPOLOGY=${HIPPO_DIR}/hippo_protein_database.dat
FF=${HIPPO_DIR}/oplsaa_forcefield.dat

echo "Setting up test directory"
rm -rf test
mkdir test
cd test

topdir="${CURDIR}/test"

function setup_hippo () {
local numthreads=${1:-1}
local input=hippo_input.txt
cp ${HIPPO} ./hippo || return $?
cp ${TOPOLOGY} . || return $?
cp ${FF} . || return $?
test -e $input || die "Missing run input file $input in $PWD"
if [ $NSLOTS -gt 1 ]; then
# adjusting for OpenMP run
sed -i.orig -e "s/[[:space:]]*openMP numthreads.*/openMP numthreads ${numthreads}/" $input
fi
return 0
}

function run_test () {
local testdir="$1" numthreads="${2:-1}"
echo "---------------------------------------------------------"
cd ${testdir} || die "Cannot 'cd ${testdir}'"
setup_hippo ${numthreads} || die "setup_hippo() failed"
echo "Set up all files for NSLOTS=${numthreads}"
echo "Running hippo test case ${testdir}..."
t_start=$(date +%s)
./hippo
t_stop=$(date +%s)
delta_t=$(( t_stop-t_start ))
echo "Completed hippo test case ${testdir} in ${delta_t} seconds, running ${numthreads} threads"
echo "BENCHMARK: ${testdir} ${numthreads} ${delta_t}"
cd ${topdir}
}

cp -r ../jobs/* .

for t in ${RUN_TESTS};
do run_test $t ${NSLOTS}
done

echo "Finished running hippo test suite"
</pre>

Calc testjobs linux.sh

2008-12-07T16:12:54Z

Oliver: no mpi

This script allows running Hippo benchmarks without having to think too much about finding the correct binary and supplemental files. In addition it gives the total wall time required to run each test case; this can be used for crude benchmarks.

* The testing of the mpi binaries is not tested and probably does not work. However, they are ''only'' needed for replica exchange simulations and there are no test cases for those anyway.
* [http://forums.biowerkzeug.org/viewtopic.php?f=22&t=6&p=15&hilit=OpenMP#p15 Not all tests support multi-threaded runs] (with OpneMP); these tests ignore the NSLOTS argument.

<pre>#!/bin/bash
# $Id: calc_testjobs_linux.sh 2535 2008-12-07 16:07:19Z www-data $
# Running Hippo tests (Linux)
# Copyright (c) 2008 Biowerkzeug
# Oliver Beckstein <orbeckst@gmail.com>
#set -x

prog=$(basename $0)
CURDIR=${PWD}

# defaults (: can be set in environment)
: ${HIPPO_DIR:="${CURDIR}/.."}
HIPPO_TESTS="hexane_NVT_dp_MD octane_NPT_sp_MC pentadecane_NPT_sp_MD tip3p_NPT_sp_MD trpzip2_GBSA_MC vpu_GBIM_MC walp_octane_NPT_sp_MD"
RUN_TESTS=${HIPPO_TESTS}
USE_MPI=0

usage="usage $prog [opts] [tests]

Run Hippo tests. By default it runs all of them:
${HIPPO_TESTS}

OPTIONS:
-h help
-n number of threads (not possible for all tests)
-D directory where we look for Hippo executables [${HIPPO_DIR}]
-M use mpi (replica exchange) binaries [${USE_MPI}]

Environment variables:
HIPPO_DIR overrides -D [${HIPPO_DIR}]
"

function die () {
local msg="$1" err=${2:-1}
echo 1>&2 "ERROR: failed in $PWD: ${msg}"
cd ${CURDIR}
exit $err
}

NSLOTS=1
# opt processing
while getopts hn:D:M: opt; do
case $opt in
h) echo "$usage"; exit 0;;
n) NSLOTS=${OPTARG};;
D) HIPPO_DIR=${OPTARG};;
M) USE_MPI=${OPTARG};;
*) die "Unknown option" 1;;
esac
done

#echo "OPTIND=$OPTIND OPTARG=$OPTARG argv=$*"
shift $((OPTIND - 1))
if [ -n "$*" ]; then
RUN_TESTS="$*"
fi

echo "Running the following tests using ${NSLOTS} threads: ${RUN_TESTS}"

# find working executable
# we'll use the first one that only complain about missing input file
#
if [ ${USE_MPI} = 0 ]; then
echo "Trying standard binaries (with OpenMP)"
_HIPPO_BINARIES="hippo hippo_p3"
else
echo "Testing mpi binaries. Note: these are ONLY needed for replica exchange simulations"
_HIPPO_BINARIES="hippo_mpi hippo_p3_mpi"
fi
HIPPO="not_found"
rm -f hippo_input.txt # clean any input files
for h in ${_HIPPO_BINARIES}; do
exe="${HIPPO_DIR}/${h}"
if ${exe} 2>&1 | egrep "^Hippo.*Copyright.*Biowerkzeug" >/dev/null; then
HIPPO=${exe}
break
fi
done
if [ "${HIPPO}" = "not_found" ]; then
ARCH=$(uname -m);
OS=$(uname -s);
die "No usable hippo executable found; see if there is one at
http://www.biowerkzeug.com for your architecture ${ARCH} and operating
system ${OS}. "
fi

echo "Using executable ${HIPPO}"

TOPOLOGY=${HIPPO_DIR}/hippo_protein_database.dat
FF=${HIPPO_DIR}/oplsaa_forcefield.dat

echo "Setting up test directory"
rm -rf test
mkdir test
cd test

topdir="${CURDIR}/test"

function setup_hippo () {
local numthreads=${1:-1}
local input=hippo_input.txt
cp ${HIPPO} ./hippo || return $?
cp ${TOPOLOGY} . || return $?
cp ${FF} . || return $?
test -e $input || die "Missing run input file $input in $PWD"
if [ $NSLOTS -gt 1 ]; then
# adjusting for OpenMP run
sed -i.orig -e "s/[[:space:]]*openMP numthreads.*/openMP numthreads ${numthreads}/" $input
fi
return 0
}

function run_test () {
local testdir="$1" numthreads="${2:-1}"
echo "---------------------------------------------------------"
cd ${testdir} || die "Cannot 'cd ${testdir}'"
setup_hippo ${numthreads} || die "setup_hippo() failed"
echo "Set up all files for NSLOTS=${numthreads}"
echo "Running hippo test case ${testdir}..."
t_start=$(date +%s)
./hippo
t_stop=$(date +%s)
delta_t=$(( t_stop-t_start ))
echo "Completed hippo test case ${testdir} in ${delta_t} seconds, running ${numthreads} threads"
echo "BENCHMARK: ${testdir} ${numthreads} ${delta_t}"
cd ${topdir}
}

cp -r ../jobs/* .

for t in ${RUN_TESTS};
do run_test $t ${NSLOTS}
done

echo "Finished running hippo test suite"
</pre>

Calc testjobs linux.sh

2008-12-07T16:08:31Z

Oliver: calc_testjobs_linux.sh 2535 2008-12-07 16:07:19Z

This script allows running Hippo benchmarks without having to think too much about finding the correct binary and supplemental files. In addition it gives the total wall time required to run each test case; this can be used for crude benchmarks.

<pre>#!/bin/bash
# $Id: calc_testjobs_linux.sh 2535 2008-12-07 16:07:19Z www-data $
# Running Hippo tests (Linux)
# Copyright (c) 2008 Biowerkzeug
# Oliver Beckstein <orbeckst@gmail.com>
#set -x

prog=$(basename $0)
CURDIR=${PWD}

# defaults (: can be set in environment)
: ${HIPPO_DIR:="${CURDIR}/.."}
HIPPO_TESTS="hexane_NVT_dp_MD octane_NPT_sp_MC pentadecane_NPT_sp_MD tip3p_NPT_sp_MD trpzip2_GBSA_MC vpu_GBIM_MC walp_octane_NPT_sp_MD"
RUN_TESTS=${HIPPO_TESTS}
USE_MPI=0

usage="usage $prog [opts] [tests]

Run Hippo tests. By default it runs all of them:
${HIPPO_TESTS}

OPTIONS:
-h help
-n number of threads (not possible for all tests)
-D directory where we look for Hippo executables [${HIPPO_DIR}]
-M use mpi (replica exchange) binaries [${USE_MPI}]

Environment variables:
HIPPO_DIR overrides -D [${HIPPO_DIR}]
"

function die () {
local msg="$1" err=${2:-1}
echo 1>&2 "ERROR: failed in $PWD: ${msg}"
cd ${CURDIR}
exit $err
}

NSLOTS=1
# opt processing
while getopts hn:D:M: opt; do
case $opt in
h) echo "$usage"; exit 0;;
n) NSLOTS=${OPTARG};;
D) HIPPO_DIR=${OPTARG};;
M) USE_MPI=${OPTARG};;
*) die "Unknown option" 1;;
esac
done

#echo "OPTIND=$OPTIND OPTARG=$OPTARG argv=$*"
shift $((OPTIND - 1))
if [ -n "$*" ]; then
RUN_TESTS="$*"
fi

echo "Running the following tests using ${NSLOTS} threads: ${RUN_TESTS}"

# find working executable
# we'll use the first one that only complain about missing input file
#
if [ ${USE_MPI} = 0 ]; then
echo "Trying standard binaries (with OpenMP)"
_HIPPO_BINARIES="hippo hippo_p3"
else
echo "Testing mpi binaries. Note: these are ONLY needed for replica exchange simulations"
_HIPPO_BINARIES="hippo_mpi hippo_p3_mpi"
fi
HIPPO="not_found"
rm -f hippo_input.txt # clean any input files
for h in ${_HIPPO_BINARIES}; do
exe="${HIPPO_DIR}/${h}"
if ${exe} 2>&1 | egrep "^Hippo.*Copyright.*Biowerkzeug" >/dev/null; then
HIPPO=${exe}
break
fi
done
if [ "${HIPPO}" = "not_found" ]; then
ARCH=$(uname -m);
OS=$(uname -s);
die "No usable hippo executable found; see if there is one at
http://www.biowerkzeug.com for your architecture ${ARCH} and operating
system ${OS}. "
fi

echo "Using executable ${HIPPO}"

TOPOLOGY=${HIPPO_DIR}/hippo_protein_database.dat
FF=${HIPPO_DIR}/oplsaa_forcefield.dat

echo "Setting up test directory"
rm -rf test
mkdir test
cd test

topdir="${CURDIR}/test"

function setup_hippo () {
local numthreads=${1:-1}
local input=hippo_input.txt
cp ${HIPPO} ./hippo || return $?
cp ${TOPOLOGY} . || return $?
cp ${FF} . || return $?
test -e $input || die "Missing run input file $input in $PWD"
if [ $NSLOTS -gt 1 ]; then
# adjusting for OpenMP run
sed -i.orig -e "s/[[:space:]]*openMP numthreads.*/openMP numthreads ${numthreads}/" $input
fi
return 0
}

function run_test () {
local testdir="$1" numthreads="${2:-1}"
echo "---------------------------------------------------------"
cd ${testdir} || die "Cannot 'cd ${testdir}'"
setup_hippo ${numthreads} || die "setup_hippo() failed"
echo "Set up all files for NSLOTS=${numthreads}"
echo "Running hippo test case ${testdir}..."
t_start=$(date +%s)
./hippo
t_stop=$(date +%s)
delta_t=$(( t_stop-t_start ))
echo "Completed hippo test case ${testdir} in ${delta_t} seconds, running ${numthreads} threads"
echo "BENCHMARK: ${testdir} ${numthreads} ${delta_t}"
cd ${topdir}
}

cp -r ../jobs/* .

for t in ${RUN_TESTS};
do run_test $t ${NSLOTS}
done

echo "Finished running hippo test suite"
</pre>

Calc testjobs linux.sh

2008-12-07T16:03:47Z

Oliver:

This script allows running Hippo benchmarks without having to think too much about finding the correct binary and supplemental files. In addition it gives the total wall time required to run each test case; this can be used for crude benchmarks.

<pre>#!/bin/bash
# Running Hippo tests (Linux)
# Copyright (c) 2008 Biowerkzeug
# Oliver Beckstein <orbeckst@gmail.com>
#set -x

prog=$(basename $0)
CURDIR=${PWD}

# defaults (: can be set in environment)
: ${HIPPO_DIR:="${CURDIR}/.."}
HIPPO_TESTS="hexane_NVT_dp_MD octane_NPT_sp_MC pentadecane_NPT_sp_MD tip3p_NPT_sp_MD trpzip2_GBSA_MC vpu_GBIM_MC walp_octane_NPT_sp_MD"
RUN_TESTS=${HIPPO_TESTS}
USE_MPI=0

usage="usage $prog [opts] [tests]

Run Hippo tests. By default it runs all of them:
${HIPPO_TESTS}

OPTIONS:
-h help
-n number of threads (not possible for all tests)
-D directory where we look for Hippo executables [${HIPPO_DIR}]
-M use mpi (replica exchange) binaries [${USE_MPI}]

Environment variables:
HIPPO_DIR overrides -D [${HIPPO_DIR}]
"

function die () {
local msg="$1" err=${2:-1}
echo 1>&2 "ERROR: failed in $PWD: ${msg}"
cd ${CURDIR}
exit $err
}

NSLOTS=1
# opt processing
while getopts hn:D:M: opt; do
case $opt in
h) echo "$usage"; exit 0;;
n) NSLOTS=${OPTARG};;
D) HIPPO_DIR=${OPTARG};;
M) USE_MPI=${OPTARG};;
*) die "Unknown option" 1;;
esac
done

#echo "OPTIND=$OPTIND OPTARG=$OPTARG argv=$*"
shift $((OPTIND - 1))
if [ -n "$*" ]; then
RUN_TESTS="$*"
fi

echo "Running the following tests using ${NSLOTS} threads: ${RUN_TESTS}"

# find working executable
# we'll use the first one that only complain about missing input file
#
if [ ${USE_MPI} = 0 ]; then
echo "Only trying single cpu binaries"
_HIPPO_BINARIES="hippo hippo_p3"
else
_HIPPO_BINARIES="hippo_mpi hippo hippo_p3_mpi hippo_p3"
fi
HIPPO="not_found"
rm -f hippo_input.txt # clean any input files
for h in ${_HIPPO_BINARIES}; do
exe="${HIPPO_DIR}/${h}"
if ${exe} 2>&1 | egrep "^Hippo.*Copyright.*Biowerkzeug" >/dev/null; then
HIPPO=${exe}
break
fi
done
if [ "${HIPPO}" = "not_found" ]; then
ARCH=$(uname -m);
OS=$(uname -s);
die "No usable hippo executable found; see if there is one at
http://www.biowerkzeug.com for your architecture ${ARCH} and operating
system ${OS}. "
fi

echo "Using executable ${HIPPO}"

TOPOLOGY=${HIPPO_DIR}/hippo_protein_database.dat
FF=${HIPPO_DIR}/oplsaa_forcefield.dat

echo "Setting up test directory"
rm -rf test
mkdir test
cd test

topdir="${CURDIR}/test"

function setup_hippo () {
local numthreads=${1:-1}
local input=hippo_input.txt
cp ${HIPPO} ./hippo || return $?
cp ${TOPOLOGY} . || return $?
cp ${FF} . || return $?
test -e $input || die "Missing run input file $input in $PWD"
if [ $NSLOTS -gt 1 ]; then
# adjusting for OpenMP run
sed -i.orig -e "s/[[:space:]]*openMP numthreads.*/openMP numthreads ${numthreads}/" $input
fi
return 0
}

function run_test () {
local testdir="$1" numthreads="${2:-1}"
echo "---------------------------------------------------------"
cd ${testdir} || die "Cannot 'cd ${testdir}'"
setup_hippo ${numthreads} || die "setup_hippo() failed"
echo "Set up all files for NSLOTS=${numthreads}"
echo "Running hippo test case ${testdir}..."
t_start=$(date +%s)
./hippo
t_stop=$(date +%s)
delta_t=$(( t_stop-t_start ))
echo "Completed hippo test case ${testdir} in ${delta_t} seconds, running ${numthreads} threads"
echo "BENCHMARK: ${testdir} ${numthreads} ${delta_t}"
cd ${topdir}
}

cp -r ../jobs/* .

for t in ${RUN_TESTS};
do run_test $t ${NSLOTS}
done

echo "Finished running hippo test suite"
</pre>

Test scaling.py

2008-12-07T16:00:51Z

Oliver: test_scaling.py 2534 2008-12-07 15:59:59Z

Python script that uses [[calc_testjobs_linux.sh]] to benchmark scaling using the ''walp_octane_NPT_sp_MD'' test case. See [[Talk:Performance#Scaling]] for details.

<pre>#!/usr/bin/env python
# $Id: test_scaling.py 2534 2008-12-07 15:59:59Z www-data $
# Testing scaling of hippo
# Copyright (c) 2008 Biowerkzeug
# Oliver Beckstein <orbeckst@gmail.com>

from subprocess import Popen,PIPE
import sys,re

calc_test_jobs = '/home/oliver/Library/Hippo/Benchmark/calc_testjobs_linux.sh'
hippo_test_case = 'walp_octane_NPT_sp_MD'
filename = "scaling.xvg"
figname = "scaling.png"

try:
maxslots = int(sys.argv[1])
except:
print "usage: %s NSLOTS" % sys.argv[0]
sys.exit(1)

slotrange = (1,maxslots+1) # <--- 1-4 !!

benchmark_pattern = re.compile(r'BENCHMARK:\s*(\w+)\s+(?P<NUMTHREADS>[0-9]+)\s+(?P<T_SECONDS>[0-9.]+)')

runtime = {}
out = open(filename,'w')
out.write("# scaling for Hippo\n# numthreads walltime/s scaling\n")
for NSLOTS in xrange(*slotrange):
print "-- running NSLOTS = %(NSLOTS)d" % vars()
p1 = Popen([calc_test_jobs, '-n', str(NSLOTS), hippo_test_case],stdout=PIPE)
p2 = Popen(['grep','BENCHMARK:'],stdin=p1.stdout,stdout=PIPE)
output = p2.communicate()[0]
m = benchmark_pattern.match(output)
print "output: ",output,
if not m:
print "ERROR: no benchmark data found"
continue
numthreads = int(m.group('NUMTHREADS'))
walltime = float(m.group('T_SECONDS'))
runtime[numthreads] = walltime
scaling = runtime[1]/walltime # runtime[1] is known after the first iteration!
out.write("%(numthreads)d %(walltime)f %(scaling)f\n" % vars())
out.close()

# Analysis
import numpy
import pylab
N = numpy.sort(runtime.keys())
T = numpy.array([runtime[n] for n in N],dtype=float)
S = T[0]/T

pylab.clf()

pylab.subplot(211)
pylab.title('Hippo test case: '+hippo_test_case)
pylab.xlabel('cpus')
pylab.ylabel('walltime/s')
pylab.plot(N,T,'ro-')

pylab.subplot(212)
pylab.xlabel('cpus')
pylab.ylabel('scaling')
pylab.plot(N,S,'ro-')
pylab.plot([N[0],N[-1]], [1,N[-1]], 'k--')

pylab.savefig(figname)
print "Saved figure "+figname
</pre>

File:Scaling Q9550.png

2008-12-07T15:53:52Z

Oliver: uploaded a new version of "Image:Scaling Q9550.png"

Scaling of WALP test case on Intel Q9550. Hippo rev35.

File:Scaling Q9550.png

2008-12-07T15:52:01Z

Oliver: uploaded a new version of "Image:Scaling Q9550.png"

Scaling of WALP test case on Intel Q9550. Hippo rev35.

Talk:Performance

2008-12-07T15:41:53Z

Oliver: /* Scaling */

== Integration with tests ==
We could write a script that does the benchmark while running the test. It would even be possible to automatically post it (with the user's consent, of course). — [[User:Oliver|Oli]] 16:03, 14 November 2008 (UTC)

== Scaling ==
How to run scaling tests/testing methodology:

Use [[test_scaling.py]] which in turn uses [[calc_testjobs_linux.sh]].
cd testjobs
test_scaling.py ''NSLOTS''
where ''NSLOTS'' is the maximum number of available cpus/cores. Results are the files
scaling.xvg # numbers
scaling.png # graphs
* Only measures ''wall time'', hence must be run on an empty machine.
* Uses <code>date +%s</code> so accuracy is limited to seconds.
* No repeat runs; results may vary slightly.

Manually:
Use updated [[calc_testjobs_linux.sh]]
DATA=scaling.xvg
MAXSLOTS=8
echo -e "# scaling for Hippo\n# numthreads walltime/s" > $DATA
for NSLOTS in `seq $MAXSLOTS`; do
echo "-- NSLOTS = $NSLOTS";
../../calc_testjobs_linux.sh -n $NSLOTS walp_octane_NPT_sp_MD \
| awk '/BENCHMARK/ {print $3, $4}' >> $DATA;
done
All the above remarks apply here, too.

Performance

2008-12-07T15:39:04Z

Oliver: /* Scaling */

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.75
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

[[Image:Scaling Q9550.png|thumb|400px|left|Intel Quad Core Q9550 2.8 GHz]]
[[Image:Scaling E5420.png|thumb|400px|right|Dual Intel Quad Core Xeon E5420 2.5 GHz]]

File:Scaling E5420.png

2008-12-07T15:35:56Z

Oliver: walp test rev 35 on dual quad core xeon E5420 2.5 GHz

walp test rev 35 on dual quad core xeon E5420 2.5 GHz

Performance

2008-12-07T15:34:33Z

Oliver: /* Scaling */

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.75
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

[[Image:Scaling Q9550.png|frame|Intel Quad Core Q9550]]

Performance

2008-12-07T15:33:21Z

Oliver: /* Scaling */

== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.75
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

== Scaling ==
All tests were run with Hippo beta rev35 on empty systems. See remarks on the [[Talk:Performance#Scaling|test methodology]].

[[Image:Scaling Q9550.png|Intel Quad Core Q9550]]

File:Scaling Q9550.png

2008-12-07T15:30:24Z

Oliver: Scaling of WALP test case on Intel Q9550. Hippo rev35.

Scaling of WALP test case on Intel Q9550. Hippo rev35.

Test scaling.py

2008-12-07T15:26:28Z

Oliver:

Python script that uses [[calc_testjobs_linux.sh]] to benchmark scaling using the ''walp_octane_NPT_sp_MD'' test case. See [[Talk:Performance#Scaling]] for details.

<pre>#!/usr/bin/env python
# $Id: test_scaling.py 2533 2008-12-07 15:11:10Z www-data $
# Testing scaling of hippo
# Copyright (c) 2008 Biowerkzeug
# Oliver Beckstein <orbeckst@gmail.com>

from subprocess import Popen,PIPE
import sys,re

calc_test_jobs = '/home/oliver/Library/Hippo/Benchmark/calc_testjobs_linux.sh'
hippo_test_case = 'walp_octane_NPT_sp_MD'
filename = "scaling.xvg"
figname = "scaling.png"

try:
maxslots = int(sys.argv[1])
except:
print "usage: %s NSLOTS" % sys.argv[0]
sys.exit(1)

slotrange = (1,maxslots+1) # <--- 1-4 !!

benchmark_pattern = re.compile(r'BENCHMARK:\s*(\w+)\s+(?P<NUMTHREADS>[0-9]+)\s+(?P<T_SECONDS>[0-9.]+)')

runtime = {}
out = open(filename,'w')
out.write("# scaling for Hippo\n# numthreads walltime/s scaling\n")
for NSLOTS in xrange(*slotrange):
print "-- running NSLOTS = %(NSLOTS)d" % vars()
p1 = Popen([calc_test_jobs, '-n', str(NSLOTS), hippo_test_case],stdout=PIPE)
p2 = Popen(['grep','BENCHMARK:'],stdin=p1.stdout,stdout=PIPE)
output = p2.communicate()[0]
m = benchmark_pattern.match(output)
print "output: ",output,
if not m:
print "ERROR: no benchmark data found"
continue
numthreads = int(m.group('NUMTHREADS'))
walltime = float(m.group('T_SECONDS'))
runtime[numthreads] = walltime
scaling = walltime/runtime[1] # runtime[1] is known after the first iteration!
out.write("%(numthreads)d %(walltime)f %(scaling)f\n" % vars())
out.close()

# Analysis
import numpy
import pylab
N = numpy.sort(runtime.keys())
T = numpy.array([runtime[n] for n in N],dtype=float)
S = T[0]/T

pylab.clf()

pylab.subplot(211)
pylab.title('Hippo test case: '+hippo_test_case)
pylab.xlabel('cpus')
pylab.ylabel('walltime/s')
pylab.plot(N,T,'ro-')

pylab.subplot(212)
pylab.xlabel('cpus')
pylab.ylabel('scaling')
pylab.plot(N,S,'ro-')
pylab.plot([N[0],N[-1]], [1,N[-1]], 'k--')

pylab.savefig(figname)
print "Saved figure "+figname
</pre>

Test scaling.py

2008-12-07T15:25:50Z

Oliver: test_scaling.py 2533 2008-12-07 15:11:10Z

Python script that uses [[calc_testjobs_linux.sh]] to benchmark scaling using the ''walp_octane_NPT_sp_MD'' test case. See [[Talk:Performace#Scaling]] for details.

<pre>#!/usr/bin/env python
# $Id: test_scaling.py 2533 2008-12-07 15:11:10Z www-data $
# Testing scaling of hippo
# Copyright (c) 2008 Biowerkzeug
# Oliver Beckstein <orbeckst@gmail.com>

from subprocess import Popen,PIPE
import sys,re

calc_test_jobs = '/home/oliver/Library/Hippo/Benchmark/calc_testjobs_linux.sh'
hippo_test_case = 'walp_octane_NPT_sp_MD'
filename = "scaling.xvg"
figname = "scaling.png"

try:
maxslots = int(sys.argv[1])
except:
print "usage: %s NSLOTS" % sys.argv[0]
sys.exit(1)

slotrange = (1,maxslots+1) # <--- 1-4 !!

benchmark_pattern = re.compile(r'BENCHMARK:\s*(\w+)\s+(?P<NUMTHREADS>[0-9]+)\s+(?P<T_SECONDS>[0-9.]+)')

runtime = {}
out = open(filename,'w')
out.write("# scaling for Hippo\n# numthreads walltime/s scaling\n")
for NSLOTS in xrange(*slotrange):
print "-- running NSLOTS = %(NSLOTS)d" % vars()
p1 = Popen([calc_test_jobs, '-n', str(NSLOTS), hippo_test_case],stdout=PIPE)
p2 = Popen(['grep','BENCHMARK:'],stdin=p1.stdout,stdout=PIPE)
output = p2.communicate()[0]
m = benchmark_pattern.match(output)
print "output: ",output,
if not m:
print "ERROR: no benchmark data found"
continue
numthreads = int(m.group('NUMTHREADS'))
walltime = float(m.group('T_SECONDS'))
runtime[numthreads] = walltime
scaling = walltime/runtime[1] # runtime[1] is known after the first iteration!
out.write("%(numthreads)d %(walltime)f %(scaling)f\n" % vars())
out.close()

# Analysis
import numpy
import pylab
N = numpy.sort(runtime.keys())
T = numpy.array([runtime[n] for n in N],dtype=float)
S = T[0]/T

pylab.clf()

pylab.subplot(211)
pylab.title('Hippo test case: '+hippo_test_case)
pylab.xlabel('cpus')
pylab.ylabel('walltime/s')
pylab.plot(N,T,'ro-')

pylab.subplot(212)
pylab.xlabel('cpus')
pylab.ylabel('scaling')
pylab.plot(N,S,'ro-')
pylab.plot([N[0],N[-1]], [1,N[-1]], 'k--')

pylab.savefig(figname)
print "Saved figure "+figname
</pre>

Talk:Performance

2008-12-07T15:23:25Z

Oliver: /* Scaling */

== Integration with tests ==
We could write a script that does the benchmark while running the test. It would even be possible to automatically post it (with the user's consent, of course). — [[User:Oliver|Oli]] 16:03, 14 November 2008 (UTC)

== Scaling ==
Use updated [[calc_testjobs_linux.sh]]
DATA=scaling.xvg
echo -e "# scaling for Hippo\n# numthreads walltime/s" > $DATA
for NSLOTS in `seq 8`; do
echo "-- NSLOTS = $NSLOTS";
../../calc_testjobs_linux.sh -n $NSLOTS walp_octane_NPT_sp_MD \
| awk '/BENCHMARK/ {print $3, $4}' >> $DATA;
done
Note:
* only run this on an empty machine: we are using ''wall time'' for timing

Slightly more sophisticated: use [[test_scaling.py]].
cd testjobs
test_scaling.py ''NSLOTS''
where ''NSLOTS'' is the maximum number of available cpus/cores. Results are the files
scaling.xvg # numbers
scaling.png # graphs

Talk:Performance

2008-12-07T13:41:02Z

Oliver: /* Scaling */

Talk:Performance

2008-12-07T13:38:51Z

Oliver: benchmarking scaling

== Integration with tests ==
We could write a script that does the benchmark while running the test. It would even be possible to automatically post it (with the user's consent, of course). — [[User:Oliver|Oli]] 16:03, 14 November 2008 (UTC)

== Scaling ==
Use updated [[calc_testjobs_linux.sh]]
for NSLOTS in `seq 8`; do
echo "-- NSLOTS = $NSLOTS";
../../calc_testjobs_linux.sh -n $NSLOTS walp_octane_NPT_sp_MD \
| awk '/BENCHMARK/ {print $3, $4}' >> scaling.dat;
done
Note:
* only run this on an empty machine: we are using ''wall time'' for timing

Calc testjobs linux.sh

2008-12-07T13:37:17Z

Oliver: benchmarking and testing script

This script allows running Hippo benchmarks.

<pre>#!/bin/bash
# Running Hippo tests (Linux)
# Copyright (c) 2008 Biowerkzeug
# Oliver Beckstein <orbeckst@gmail.com>
#set -x

prog=$(basename $0)
CURDIR=${PWD}

# defaults (: can be set in environment)
: ${HIPPO_DIR:="${CURDIR}/.."}
HIPPO_TESTS="hexane_NVT_dp_MD octane_NPT_sp_MC pentadecane_NPT_sp_MD tip3p_NPT_sp_MD trpzip2_GBSA_MC vpu_GBIM_MC walp_octane_NPT_sp_MD"
RUN_TESTS=${HIPPO_TESTS}
USE_MPI=0

usage="usage $prog [opts] [tests]

Run Hippo tests. By default it runs all of them:
${HIPPO_TESTS}

OPTIONS:
-h help
-n number of threads (not possible for all tests)
-D directory where we look for Hippo executables [${HIPPO_DIR}]
-M use mpi (replica exchange) binaries [${USE_MPI}]

Environment variables:
HIPPO_DIR overrides -D [${HIPPO_DIR}]
"

function die () {
local msg="$1" err=${2:-1}
echo 1>&2 "ERROR: failed in $PWD: ${msg}"
cd ${CURDIR}
exit $err
}

NSLOTS=1
# opt processing
while getopts hn:D:M: opt; do
case $opt in
h) echo "$usage"; exit 0;;
n) NSLOTS=${OPTARG};;
D) HIPPO_DIR=${OPTARG};;
M) USE_MPI=${OPTARG};;
*) die "Unknown option" 1;;
esac
done

#echo "OPTIND=$OPTIND OPTARG=$OPTARG argv=$*"
shift $((OPTIND - 1))
if [ -n "$*" ]; then
RUN_TESTS="$*"
fi

echo "Running the following tests using ${NSLOTS} threads: ${RUN_TESTS}"

# find working executable
# we'll use the first one that only complain about missing input file
#
if [ ${USE_MPI} = 0 ]; then
echo "Only trying single cpu binaries"
_HIPPO_BINARIES="hippo hippo_p3"
else
_HIPPO_BINARIES="hippo_mpi hippo hippo_p3_mpi hippo_p3"
fi
HIPPO="not_found"
rm -f hippo_input.txt # clean any input files
for h in ${_HIPPO_BINARIES}; do
exe="${HIPPO_DIR}/${h}"
if ${exe} 2>&1 | egrep "^Hippo.*Copyright.*Biowerkzeug" >/dev/null; then
HIPPO=${exe}
break
fi
done
if [ "${HIPPO}" = "not_found" ]; then
ARCH=$(uname -m);
OS=$(uname -s);
die "No usable hippo executable found; see if there is one at
http://www.biowerkzeug.com for your architecture ${ARCH} and operating
system ${OS}. "
fi

echo "Using executable ${HIPPO}"

TOPOLOGY=${HIPPO_DIR}/hippo_protein_database.dat
FF=${HIPPO_DIR}/oplsaa_forcefield.dat

echo "Setting up test directory"
rm -rf test
mkdir test
cd test

topdir="${CURDIR}/test"

function setup_hippo () {
local numthreads=${1:-1}
local input=hippo_input.txt
cp ${HIPPO} ./hippo || return $?
cp ${TOPOLOGY} . || return $?
cp ${FF} . || return $?
test -e $input || die "Missing run input file $input in $PWD"
if [ $NSLOTS -gt 1 ]; then
# adjusting for OpenMP run
sed -i.orig -e "s/[[:space:]]*openMP numthreads.*/openMP numthreads ${numthreads}/" $input
fi
return 0
}

function run_test () {
local testdir="$1" numthreads="${2:-1}"
echo "---------------------------------------------------------"
cd ${testdir} || die "Cannot 'cd ${testdir}'"
setup_hippo ${numthreads} || die "setup_hippo() failed"
echo "Set up all files for NSLOTS=${numthreads}"
echo "Running hippo test case ${testdir}..."
t_start=$(date +%s)
./hippo
t_stop=$(date +%s)
delta_t=$(( t_stop-t_start ))
echo "Completed hippo test case ${testdir} in ${delta_t} seconds, running ${numthreads} threads"
echo "BENCHMARK: ${testdir} ${numthreads} ${delta_t}"
cd ${topdir}
}

cp -r ../jobs/* .

for t in ${RUN_TESTS};
do run_test $t ${NSLOTS}
done

echo "Finished running hippo test suite"
</pre>

Performance

2008-12-07T12:20:18Z

Oliver: rev35

As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List of Intelmicroprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 CPUs]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List of AMD microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD CPUs]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable sortable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev35 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.75
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

Performance

2008-11-21T02:13:01Z

Oliver:

As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List_of_Intel_microprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 cpus]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List_of_AMD_microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD cpus]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Phenom_microprocessors#.22Agena.22_.28B2_.26_B3.2C_65_nm.29 Phenom X4 9850]
| 2.75
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

Performance

2008-11-21T02:10:37Z

Oliver:

As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List_of_Intel_microprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 cpus]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List_of_AMD_microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD cpus]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28standard-voltage.2C_45_nm.29 Quad Core Xeon E5420]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| Phenom 9850 Quad-Core
| 2.75
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.

Performance

2008-11-21T02:08:52Z

Oliver:

As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
cd testjobs
time ./calc_testjobs
<ul>
<li>We report the ''user'' time.</li>
<li>CPU properties are found with
<pre>cat /proc/cpuinfo</pre>
and model names from [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors List_of_Intel_microprocessors] for Intel (also see the [http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#Detailed_x86_architecture_microprocessor_lists detailed lists of Intel x86 cpus]) and [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors List_of_AMD_microprocessors] for AMD cpus (also see the [http://en.wikipedia.org/wiki/List_of_AMD_microprocessors#Detailed_microprocessor_release_lists detailed lists of AMD cpus]).</li>
<li>Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt>) and the revision.</li>
<li>Note that these tests only utilize a ''single core''.</li>
<li>These results only give a rough idea of the ''relative'' performance of different architectures.</li>
</ul>

Please add your own results.
{| class="wikitable"
|-
! vendor
! model
! GHz
! cores
! time/min
! binary
! revision
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:04
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Core_2_microprocessors#.22Yorkfield.22_.2845_nm.29 Core 2 Quad Q9550]
| 2.83
| 4
| 1:06
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon E4520]
| 2.5
| 4
| 1:16
| hippo
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:27
| hippo
| rev32
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 4
| 1:31
| hippo_p3
| rev32 
|-
| AMD
| Phenom 9850 Quad-Core
| 2.75
| 4
| 1:36
| hippo_p3
| rev32 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Opteron_microprocessors#Opteron_200-series_.22SledgeHammer.22_.28B3_.26_C0_.26_CG.2C_130_nm.29 Opteron 248]
| 2.2
| 1
| 2:18
| hippo_p3
| rev32 
|-
| Intel
| Core Duo T2300 (Mac Mini)
| 1.66
| 2
| 2:41
| hippo_p3
| rev32 
|-
| Intel
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Prestonia.22_.28standard-voltage.2C_130_nm.29 Xeon 2.4]
| 2.4
| 1
| 3:03
| hippo_p3
| rev26 
|-
| AMD
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 Athlon XP 1800+]
| 1.53
| 1
| 4:30
| hippo_p3
| rev26
|}

(You don't ''have'' to link to the CPU wikipedia page; just put down whatever you got and even if you're unsure. If in doubt just put the output from
cat /proc/cpuinfo | \
awk 'BEGIN {FS=":"}; \
/vendor_id/ {vendor=$2}; \
/model name/ {model=$2}; \
/cpu MHz/ {GHz=$2/1000}; \
/siblings/ {cores=$2}; \
END {printf("|-\n| %s\n| %s\n| %.1f\n| %d\n| TIME\n| ?\n| ?\n", vendor, model, GHz, cores)}'
into the wiki table and replace ''TIME'' by what you measured.