Performance: Difference between revisions

From Biowerkzeug Wiki
Jump to navigationJump to search
hippo_p3 on Q
Kaihsu (talk | contribs)
use <source/>
 
(32 intermediate revisions by 2 users not shown)
Line 1: Line 1:
== Single processor performance ==
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
As a basic performance test we run the simulations in the <tt>test</tt> directory with these commands (on Linux):
  cd testjobs
  cd testjobs
  time ./calc_testjobs
  time ./calc_testjobs
We report the ''user'' time. CPU properties are found with
<ul>
cat /proc/cpuinfo
<li>We report the ''user'' time.</li>
Note that these tests only utilize a single core. Also report the binary used (<tt>hippo</tt> or <tt>hippo_p3</tt> and the revision).
<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 feel free to add your own results.
Please add your own results.
{| class="wikitable"
{| 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.
! CPU make
! vendor
! model
! model
! GHz
! GHz
Line 15: Line 23:
! time/min
! time/min
! binary
! binary
! rev
! 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 <!-- deathspud -->
|-
| 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 <!-- deathspud -->
| [[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 <!-- deathspud -->
|-
|-
| Intel Xeon Quad
| Intel
| L5410
|  [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 <!-- darthtater -->
| [[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 <!-- darthtater -->
|-
| 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
| 2.33
| 4
| 4
| 1:27
| 1:27
| hippo
| hippo
| beta rev32
| rev32
|-
|-
| Intel Xeon Quad
| Intel  
| L5410
| [http://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Harpertown.22_.28low-voltage.2C_45_nm.29 Quad Core Xeon L5410]
| 2.33
| 2.33
| 4
| 4
| 1:31
| 1:31
| hippo_p3
| hippo_p3
| beta rev32
| rev32 <!-- tinman -->
|-
| 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 <!-- greenwulf -->
|-
|-
| Intel Core Duo
| AMD
| T2300 (Mac Mini)
| [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 <!-- greenwulf -->
| [[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 <!-- gltph -->
|-
| Intel  
| Core Duo T2300 (Mac Mini)
| 1.66
| 1.66
| 2
| 2
| 2:41
| 2:41
| hippo_p3
| hippo_p3
| beta rev32
| rev32 <!-- abbondanza -->
|-
| 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 <!-- timberwulf -->
|-
|-
| [http://en.wikipedia.org/wiki/List_of_AMD_Athlon_XP_microprocessors#Athlon_XP_.22Palomino.22_.28Model_6.2C_180_nm.29 AMD Athlon XP 1800+]
| AMD
| AX1800DMT3C
| [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
| 1.53
| 1
| 1
| 4:30
| 4:30
| hippo_p3
| hippo_p3
| beta rev26
| rev26
|}
 
<small>(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
<source lang="bash">
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)}'
</source>
into the wiki table and replace ''TIME'' by what you measured.</small>
 
== 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]]
|}
|}

Latest revision as of 15:49, 8 December 2008

Single processor performance

As a basic performance test we run the simulations in the test directory with these commands (on Linux):

cd testjobs
time ./calc_testjobs

Please add your own results.

Benchmark (single core) and 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 Core 2 Quad Q9550 2.83 4 1:04 hippo rev32
Intel Core 2 Quad Q9550 2.83 4 1:05 hippo rev35
Error creating thumbnail: sh: 1: /usr/bin/convert: not found Error code: 127
Intel Quad Core Q9550 2.8 GHz
Intel Core 2 Quad Q9550 2.83 4 1:06 hippo_p3 rev32
Intel Quad Core Xeon E5420 2.5 8 = 2x4 1:16 hippo rev35
Error creating thumbnail: sh: 1: /usr/bin/convert: not found Error code: 127
Dual Intel Quad Core Xeon E5420 2.5 GHz
Intel Quad Core Xeon E5420 2.5 4 1:16 hippo rev32
Intel Quad Core Xeon L5410 2.33 4 1:27 hippo rev32
Intel Quad Core Xeon L5410 2.33 4 1:31 hippo_p3 rev32
AMD Phenom X4 9850 2.5 4 1:36 hippo_p3 rev32
AMD Phenom X4 9850 2.5 4 1:37 hippo_p3 rev35
Error creating thumbnail: sh: 1: /usr/bin/convert: not found Error code: 127
AMD Phenom X4 9850 Quad Core 2.5 GHz
AMD 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 Xeon 2.4 2.4 1 3:03 hippo_p3 rev26
AMD 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 test methodology.

In the table below, # cpus really means how many independent cores are available.

Scaling of the Hippo WALP test case using OpenMP on single-board machines
vendor # cpus walltime and scaling
Intel 4
Error creating thumbnail: sh: 1: /usr/bin/convert: not found Error code: 127
Intel Quad Core Q9550 2.8 GHz
Intel 8
Error creating thumbnail: sh: 1: /usr/bin/convert: not found Error code: 127
Dual Intel Quad Core Xeon E5420 2.5 GHz
AMD 4
Error creating thumbnail: sh: 1: /usr/bin/convert: not found Error code: 127
AMD Phenom X4 9850 Quad Core 2.5 GHz