Benchmarks: Difference between revisions

From CYANA Wiki
Jump to navigation Jump to search
 
(42 intermediate revisions by the same user not shown)
Line 3: Line 3:
{|border="1" cellpadding="5" cellspacing="0" style="text-align:center"
{|border="1" cellpadding="5" cellspacing="0" style="text-align:center"
|-
|-
|rowspan="2"|'''Computer system'''
|align="left" rowspan="2"|'''Computer system'''
|rowspan="2" width="15%"|'''Number of processors (cores)'''
|rowspan="2" width="15%"|'''Number of processors (cores) used'''
|colspan="2"|'''Computation time (s)'''
|colspan="2"|'''Computation time (s)'''
|rowspan="2" width="15%"|'''Relative performance'''
|rowspan="2" width="15%"|'''Relative performance'''
|-
|-
|width="15%"|'''Basic'''
|width="15%"|'''[[Basic structure calculation starting from given restraints|Basic]]'''
|width="15%"|'''Auto'''
|width="15%"|'''[[Structure calculation with automated NOESY assignment|Auto]]'''
|-
|-
|align="left"|Laptop|| 1 || 813 || ? || 0.26
|align="left"|Laptop|| 1 || 813 || ? || 0.26
|-
|-
|align="left" rowspan="2"|Desktop (intel) || 1 || 208 || ? || = 1
|align="left" rowspan="2"|Desktop (intel) || 1 || 208 || 9031 || = 1
|-
|-
| 4 || 57 || 2479 || 3.65
| 4 || 57 || 2479 || 3.65
|-
|-
|align="left"|Desktop (gfortran) || 1 || 258 || ? || 0.81
|align="left"|Desktop (gfortran) || 1 || 258 || 11705 || 0.77
|-
|-
|align="left"|Desktop (g95) || 1 || 336 || ? || 0.62
|align="left"|Desktop (g95) || 1 || 336 || 14898 || 0.61
|-
|-
|align="left" rowspan="2"|Mac mini (intel) || 1 || ? || ? || ?
|align="left" rowspan="2"|Mac mini || 1 || 157 || 7196|| 1.26
|-
|-
| 2 || 87 || 4092 || 2.39
| 2 || 87 || 4092 || 2.21
|-
|-
|align="left"|Mac mini (gfortran) || 2 || 104 || 4894 || 2.00
<!--
|align="left"|Mac mini (gfortran) || 2 || 104 || 4894 || 1.85
|-
|-
|align="left" rowspan="5"|Server ||  1 || 146 || ? || 1.42
-->
|align="left" rowspan="5"|Cluster ||  1 || 146 || 6756 || 1.34
|-
|-
| 10 || 16 || 745 || 13.0
| 10 || 16 || 745 || 12.1
|-
|-
| 20 || 11 || 412 || 18.9
| 20 || 11 || 412 || 21.9
|-
|-
| 50 || 4 || 209 || 52.0
| 50 || 4 || 209 || 43.2
|-
|-
|100 || 5 || 142 || 41.6
|100 || 5 || 142 || 63.6
|-
|-
|}
|}
Line 41: Line 43:
February 2010
February 2010


===== Computer systems =====


===== Computer systems: =====
* '''Laptop:''' Laptop computer with 1 Intel Core 2 Duo L7800 dual-core CPU, 2.00 Ghz, 2 GB memory, Cygwin under Windows 7, [http://gcc.gnu.org/fortran/ gfortran] (-O3) Fortran compiler 4.3.4.


* '''Laptop:''' Laptop computer with 1 Intel Core 2 Duo L7800 dual-core CPU, 2.00 Ghz, 2 GB memory, Cygwin under Windows 7, gfortran (-O3) Fortran compiler 4.3.4.
* '''Desktop:''' Desktop computer with 1 Intel Core 2 Q9400 quad-core CPU, 2.66 Ghz, 4 GB memory, Ubuntu 8.10 Linux, [http://software.intel.com/en-us/intel-compilers/ Intel Fortran compiler] 11.1, gfortran (-O3) Fortran compiler 4.3.2, or [http://www.g95.org/ g95] (-O5) Fortran compiler 0.91.  


* '''Desktop:''' Desktop computer with 1 Intel Core 2 Q9400 quad-core CPU, 2.66 Ghz, 4 GB memory, Ubuntu 8.10 Linux, Intel Fortran compiler 11.1, gfortran (-O3) Fortran compiler 4.3.2, or [http://www.g95.org/ g95] Fortran compiler 0.91.  
* '''Mac mini:''' Mac mini with 1 Intel Core 2 Duo P8700 dual-core CPU, 2.53 Ghz, 4 GB memory, Mac OS X 10.6.2, Intel Fortran compiler 11.1<!-- or gfortran (-O3) Fortran compiler 4.5.0-->.


* '''Mac mini:''' Mac mini with 1 Intel Core 2 Duo dual-core CPU, 2.53 Ghz, 4 GB memory, Mac OS X 10.6.2, Intel Fortran compiler 11.1 or gfortran (-O3) Fortran compiler 4.5.0.
* '''Cluster:''' Linux cluster system with 20 nodes, each having 2 Intel Xeon E5462 quad-core CPUs, 2.80 Ghz, 16 GB memory, Ubuntu 8.04 Linux, Intel Fortran compiler 10.1, [http://www.open-mpi.org/ OpenMPI].


* '''Server:''' Linux cluster system with 20 nodes, each having 2 Intel Xeon E5462 quad-core CPUs, 2.80 Ghz, 16 GB memory, Ubuntu 8.04 Linux, Intel Fortran compiler 10.1, OpenMPI.
===== Benchmark types =====


===== Benchmark types: =====
* '''[[Basic structure calculation starting from given restraints|Basic]]:''' Structure calculation of a protein with 114 amino acid residues, 1737 NOE distance restraints, 110 torsion angle restraints. 50 conformers are calculated using 4000 torsion angle dynamics steps per conformer.


* '''Basic:''' Structure calculation of a protein with 114 amino acid residues, 1737 NOE distance restraints, 110 torsion angle restraints. 50 conformers are calculated using 4000 torsion angle dynamics steps per conformer.
* '''[[Structure calculation with automated NOESY assignment|Auto]]''': Structure calculation with automated NOESY assignment of a protein with 114 amino acid residues, 4732 NOESY cross peaks, 88 torsion angle restraints. 8 x 100 = 800 conformers are calculated using 10000 torsion angle dynamics steps per conformer.


* '''Auto:''' Structure calculation with automated NOESY assignment of a protein with 114 amino acid residues, 4732 NOESY cross peaks, 88 torsion angle restraints. 8 x 100 = 800 conformers are calculated using 10000 torsion angle dynamics steps per conformer.
===== Relative performance =====


===== Relative performance =====
* Normalized to 1.0 for the calculation on the 'Desktop' computer system using 1 processor core.
* Based on the computation time for the '''Auto''' calculation
 
== Notes ==
 
* Running the complete CYANA 3.0 test suite, including the 'basic' and 'auto' calculations, as well as structure calculations for homodimers, structure calculations using RDCs, pseudocontact shifts, etc., takes about 3.65 times as long as the '''Auto''' benchmark.
 
* Doubling the number of torsion angle dynamics steps from 4000 to 8000 in the '''Basic''' calculation requires about 1.8 times more computation time.


* Normalized to 1.0 for the calculation with the 'Desktop' computer system using 1 processor core.
* Doubling the number of input distance and torsion angle restraints in the '''Basic''' calculation requires about 1.17 times more computation time.
* Based on the computation time for the '''Basic''' calculation


===== Remarks =====
* Performance with the gfortran compiler is 75-85% of the performance achieved with the Intel Fortran compiler.


* Running the complete CYANA 3.0 test suite, including the 'basic' and 'auto' calculations, as well as structure calculations for homodimers, structure calculations using RDCs, pseudocontact shifts, etc., takes about 3.7 times as long as the '''Auto''' benchmark.
* Performance with the g95 compiler is 60-65% of the performance achieved with the Intel Fortran compiler.

Latest revision as of 13:09, 9 March 2010

Benchmark results for CYANA 3.0:

Computer system Number of processors (cores) used Computation time (s) Relative performance
Basic Auto
Laptop 1 813 ? 0.26
Desktop (intel) 1 208 9031 = 1
4 57 2479 3.65
Desktop (gfortran) 1 258 11705 0.77
Desktop (g95) 1 336 14898 0.61
Mac mini 1 157 7196 1.26
2 87 4092 2.21
Cluster 1 146 6756 1.34
10 16 745 12.1
20 11 412 21.9
50 4 209 43.2
100 5 142 63.6

February 2010

Computer systems
  • Laptop: Laptop computer with 1 Intel Core 2 Duo L7800 dual-core CPU, 2.00 Ghz, 2 GB memory, Cygwin under Windows 7, gfortran (-O3) Fortran compiler 4.3.4.
  • Desktop: Desktop computer with 1 Intel Core 2 Q9400 quad-core CPU, 2.66 Ghz, 4 GB memory, Ubuntu 8.10 Linux, Intel Fortran compiler 11.1, gfortran (-O3) Fortran compiler 4.3.2, or g95 (-O5) Fortran compiler 0.91.
  • Mac mini: Mac mini with 1 Intel Core 2 Duo P8700 dual-core CPU, 2.53 Ghz, 4 GB memory, Mac OS X 10.6.2, Intel Fortran compiler 11.1.
  • Cluster: Linux cluster system with 20 nodes, each having 2 Intel Xeon E5462 quad-core CPUs, 2.80 Ghz, 16 GB memory, Ubuntu 8.04 Linux, Intel Fortran compiler 10.1, OpenMPI.
Benchmark types
  • Basic: Structure calculation of a protein with 114 amino acid residues, 1737 NOE distance restraints, 110 torsion angle restraints. 50 conformers are calculated using 4000 torsion angle dynamics steps per conformer.
  • Auto: Structure calculation with automated NOESY assignment of a protein with 114 amino acid residues, 4732 NOESY cross peaks, 88 torsion angle restraints. 8 x 100 = 800 conformers are calculated using 10000 torsion angle dynamics steps per conformer.
Relative performance
  • Normalized to 1.0 for the calculation on the 'Desktop' computer system using 1 processor core.
  • Based on the computation time for the Auto calculation

Notes

  • Running the complete CYANA 3.0 test suite, including the 'basic' and 'auto' calculations, as well as structure calculations for homodimers, structure calculations using RDCs, pseudocontact shifts, etc., takes about 3.65 times as long as the Auto benchmark.
  • Doubling the number of torsion angle dynamics steps from 4000 to 8000 in the Basic calculation requires about 1.8 times more computation time.
  • Doubling the number of input distance and torsion angle restraints in the Basic calculation requires about 1.17 times more computation time.
  • Performance with the gfortran compiler is 75-85% of the performance achieved with the Intel Fortran compiler.
  • Performance with the g95 compiler is 60-65% of the performance achieved with the Intel Fortran compiler.