La version du navigateur que vous utilisez n'est pas recommandée pour ce site.
Nous vous conseillons de mettre à niveau vers la version la plus récente de votre navigateur en cliquant sur l'un des liens suivants.

Intel® Xeon Phi™ Product Family

Power your breakthrough innovations with the highly parallel processing of the Intel® Xeon Phi™ coprocessor. We have packed over a teraFLOPS of double-precision peak performance into every chip.

Life science applications

Life sciences is one of the largest commercial market segments for high-performance computing. Life sciences includes many disciplines, such as chemistry, bio-chemistry, molecular modeling, protein folding, genomics, and statistical analysis to name a few. Quick simulation and analysis leads to breakthroughs in all areas of the life sciences. BWA maps human genomes. Gromacs simulates particle molecular dynamics. LAMMPS is a versatile large scale atomic simulation package. MPI-HMMER is an open source MPI implementation of the HMMER protein sequence analysis suite. NAMD is a parallel molecular dynamics code for large bio molecular systems.

Developer starter kits >

Configuration Details:

BLASTp & BLASTn : (v.30) (Heterogeneous)

Platform hosting the coprocessor and platform for two-socket Intel® Xeon® processor baseline:

Two-socket Intel® Software Development Platform: two Intel Xeon processors E5-2697v2 (12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 130W TDP, Intel® Turbo Boost Technology on, Intel® Hyper-Threading Technology on) 64GB Memory at 1600 MHz, RHEL 6.4

Coprocessor details:

Intel® Xeon Phi™ coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology off, error-correcting code (ECC) on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 3.3.1, Intel® Compiler 15.0.0

 

  BLASTp

BLASTn

Two-socket Intel Xeon processor score:

56.0 sec

52.5 sec (lower is better)

Two-socket Intel Xeon processor + Intel® Xeon Phi™ coprocessor score:

48.5 sec 

41.8 sec (lower is better)

Source: Intel Internal Testing TR 2110

 

Soft Sphere Simulation (Pre Beta version) (3D Molecular Dynamic application of CAS IPC (Institute of Process Engineering, Chinese Academy of Sciences): Workload 90*90*90 (NDA Result) (Native)

Platform hosting the coprocessor and platform for two-socket Intel Xeon processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 130W TDP, Intel Turbo Boost Technology on, Intel Hyper-Threading Technology on) 64GB Memory at 1600MHz, RHEL 6.2

Coprocessor Details:

Intel Xeon Phi coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Turbo on, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 2.1.6720-19, Intel Composer XE sp1.1.106, Intel MPI 4.1.0.030
Two-socket Intel® Xeon® processor score: 14.6 seconds (lower is better)
Two-socket Intel Xeon processor + Intel® Xeon Phi™ coprocessor score: 8.75 seconds (lower is better)

Source: Intel Internal Testing TR 2103

 

BWA:  (Burrow-Wheeler Aligner Release bwa-0.5.10)

Platform hosting the coprocessor and platform for two-socket Intel Xeon processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 130W TDP, Intel Turbo Boost Technology on, Intel® Hyper-Threading Technology on) 64GB Memory at 1600MHz, RHEL 6.4

Coprocessor Details:

Intel® Xeon Phi™ coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology off, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 2.1.6720-16 (Flash: 2.1.03.0386; coprocessor OS: 2.6.38.8-gefd324e)

Intel Composed XE 13.1.3 163.0, Intel MPI 4.1.1.030

Two-socket Intel® Xeon® processor score Xeon Score: 178.73 seconds (lower is better)

Two-socket Intel Xeon processor + Intel® Xeon Phi™ coprocessor score: 119.17 seconds (lower is better)

Two-socket Intel Xeon processor + two Intel® Xeon Phi™ coprocessor score: 87.83 seconds (lower is better)

Source: Intel Internal Testing TR 2093

 

NAMD : ( STMV single node, 23ppn, 2.10 Pre-Release)

Platform hosting the coprocessor and platform for two-socket Intel® Xeon® processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 130W TDP, Intel Turbo Boost Technology on, Intel® Hyper-Threading Technology on) 64GB Memory at 1600MHz, RHEL 6.2

Coprocessor Details:

Intel® Xeon Phi™ coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 2.1.6720-21 (Flash: 2.1.03.0386; coprocessor OS: 2.6.38.8-gefd324e)

Intel Compiler 13.1.3 20130607

Two-socket Intel Xeon processor score Xeon Score: 0.158 ns/day (higher is better)
Two-socket Intel Xeon processor + two Intel Xeon Phi coprocessor score: 0.513 ns/day (higher is better)

Source: Intel Internal Testing TR 2100A

 

AMBER 14: ( Cellulose NPT, 408K Atoms) (Sept 2014)

Platform hosting the coprocessor and platform for two-socket Intel Xeon processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 130W TDP, Intel Turbo Boost Technology on, Intel Hyper-Threading Technology on) 64GB Memory at 1600MHz, RHEL 6.5

Coprocessor Details:

Intel Xeon Phi coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology off, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 2.1.6720-19 (Flash: 2.1.03.0386)

Intel Compiler 14.0 U1 , Intel MPI 4.1.1.036

Two-socket Intel Xeon processor score Xeon Score: 1.55 ns/day (higher is better)

Two-socket Intel Xeon processor + Intel Xeon Phi coprocessor score:  1.98 ns/day (higher is better)

Source: Intel Internal Testing TR 2108A

 

MPIHammer:

Platform hosting the coprocessor and platform for two-socket Intel® Xeon® processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2670 (8Core, 20M cache, 2.6GHz, 8.0GT/s Intel QPI, 115W TDP, Intel Turbo Boost Technology on, Intel Hyper-Threading Technology on) 64GB Memory @ 1600MHz, RHEL 6.4

Coprocessor Details:

Intel Xeon Phi coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology off, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 2.1.6720-13 (Flash: 2.1.02.0386; coprocessor OS: 2.6.38.8-g5f2543d)

Intel Composed XE 2013.3.163

Two-socket Intel Xeon processor score Xeon Score: 89 seconds

Two-socket Intel Xeon processor + Intel Xeon Phi coprocessor score: 57 Seconds

Source: Intel Internal Testing TR 2051

 

Bowtie 2 (John Hopkins – version 2.2.3)  (workload = TGen)  (Jan 2015) (Heterogeneous)

Platform hosting the coprocessor and platform for two-socket Intel Xeon processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (2 x 12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 135W TDP, Intel Turbo Boost Technology on, Intel Hyper-Threading Technology on)  64GB Memory at 1600MHz, RHEL 6.4

Coprocessor Details:

Intel Xeon Phi coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Hyper-Threading Technology off, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 3.4.1, Intel Compiler 15.0.09

Two-socket Intel Xeon processor score Xeon Score: 1721.78 Seconds (run time) lower is better

Two-socket Intel Xeon processor + Intel Xeon Phi coprocessor score: 1253.78 Seconds (run time) lower is better

Source: Intel Internal Testing TR 2118

 

LAMMPS (Rhodopsin Benchmark, 512K atoms, August 2014 Main LAMMPS repository) (Sept 2014)

Platform hosting the coprocessor and platform for two-socket Intel® Xeon® processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (2 x 12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 135W TDP, Intel Turbo Boost Technology on, Intel Hyper-Threading Technology on)  64GB Memory at 1600MHz, RHEL 6.5

Coprocessor Details:

Intel Xeon Phi coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology off, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 2.1.6720-13, Flash 2.1.03.0386

Intel Compiler 14.0.1; Intel MPI5.0.028

Two-socket Intel Xeon processor score: 16.94 Seconds (simulation time) lower is better

Two-socket Intel Xeon processor + Intel Xeon Phi coprocessor score: 11.40 seconds (lower is better)

Source: Intel Internal Testing TR 2104B

 

Quantum Espresso (QE) DEISA Benchmark – AUSRUF112 (version 5.0.3 - July 2014)

(http://qe-forge.org/gf/download/frsrelease/45/78/AUSURF112.tgz)

 http://www.quantum-espresso.org

Platform hosting the coprocessor and platform for two-socket Intel Xeon processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (2 x 12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 135W TDP, Intel Turbo Boost Technology on, Intel Hyper-Threading Technology on) 64GB Memory at 1600MHz, RHEL 6.4

Coprocessor Details:

Intel Xeon Phi coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology off, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 2.1.6720-16 (Flash:  2.1.02.0386)

Intel Compiler 14.0.1; Intel MPI 4.1.1.036

 

1 Node

16 Nodes

Two-socket Intel Xeon processor score Xeon Score:

26.84 seconds

4.33 seconds

Two-socket Intel Xeon processor + Intel Xeon Phi coprocessor score:

19.10 seconds

2.63 seconds

Source: Intel Internal Testing TR 2092

 

NWCHEM CCSD(T) (www.nwchem-sw.org / https://svn.pnl.gov/svn/nwchem/trunk ) - Computational chemistry software package (Quantum chemistry  and Molecular dynamics), (code 6.5 in development) July 2014.

Platform hosting the coprocessor and platform for two-socket Intel® Xeon® processor baseline: 

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (2 x 12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 135W TDP, Intel Turbo Boost Technology on, Intel Hyper-Threading Technology on) 64GB Memory at 1600MHz, RHEL 6.4

Coprocessor Details:

Intel Xeon Phi coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 2.1.6720-16 /Flash Version: 2.1.03.0386

Intel Compiler 14.0.0 U1; Intel MPI 4.1.1.036

Fabrics: FDR Mellanox

Two-socket Intel Xeon processor score: 747.0 sec (lower is better)

Two-socket Intel Xeon processor + two Intel Xeon Phi coprocessors score: 552.6 sec (lower is better)

Source: Intel Internal Testing TR 2105

 

GROMACS (Version 5.0-rc1workload 512K H20 with RF method)

Platform hosting the coprocessor and platform for two-socket Intel Xeon processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 130W TDP, Intel Turbo Boost Technology on, Intel Hyper-Threading Technology on) 64GB Memory at 1600MHz, RHEL 6.4

Coprocessor Details:

Intel Xeon Phi coprocessor 7120A: 61cores, 1.238GHz, 16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology off, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 2.1.6720-13 (Flash: 2.1.03.0386; coprocessor OS: 2.6.38.8-gefd324e)

Intel Composed XE 13.3.163.0, Intel MPI 4.1.1.030

Two-socket Intel Xeon processor score:  2.312 ns/day

Two-socket Intel Xeon processor + Intel Xeon Phi coprocessor score:  3.605 ns/day seconds

Source: Intel Internal Testing TR 2101

 

Open LB (Lattice Boltzmann Method) (Release olb-0.8r0)

Platform hosting the coprocessor and platform for two-socket Intel Xeon processor baseline:

Two-socket Intel Software Development Platform: two Intel Xeon processors E5-2697v2 (12Core, 30M cache, 2.7GHz, 8.0GT/s Intel QPI, 130W TDP, Intel Turbo Boost Technology on, Intel Hyper-Threading Technology on)  64GB Memory at 1600MHz, RHEL 6.2

Coprocessor Details:

Intel Xeon Phi coprocessor 7120A: 61cores, 1.238GHz,  16 memory channels, 16GB Memory at 5.5GT/s, 300W TDP C-step (Intel Turbo Boost Technology, ECC on)

Software Stack (Intel Xeon Phi coprocessor):

MPSS 3.3 (Flash: 2.1.02.0390; coprocessor OS: 2.6.38.8)

Intel Composed XE 15.0.0.90, Intel MPI 5.0.1.035

Two-socket Intel Xeon processor score: 0.27 seconds (time/step, lower is better)

Two-socket Intel Xeon processor + Intel Xeon Phi coprocessor score: 0.179 seconds (time/step, lower is better)

Source: Intel Internal Testing TR 2113

Additional information: 1 2 3 4 5

Vidéos associées

Infos sur le produit et ses performances

1

Les logiciels et charges de travail utilisés dans les tests de performance ont peut-être été optimisés uniquement pour les microprocesseurs Intel®. Les tests de performance tels que SYSmark* et MobileMark* portent sur des configurations, composants, logiciels, opérations et fonctions spécifiques. Les résultats peuvent varier en fonction de ces facteurs. Pour l'évaluation d'un produit, il convient de consulter d'autres tests et d'autres sources d'informations, notamment pour connaître le comportement de ce produit avec d'autres composants. Informations complémentaires : www.intel.com/performance.

2

Intel ne maîtrise et ne vérifie ni la mise en place, ni la mise en œuvre des bancs d’essai cités ici en référence et effectués par des tiers, que ce soit directement ou à partir des sites Internet sur lesquels ils sont publiés. Intel incite l’ensemble de ses clients à consulter eux-mêmes ces sites ou bien d’autres sites qui publient les résultats de bancs d’essai similaires ainsi qu’à juger par eux-mêmes de l’exactitude des résultats de ceux-ci et de leur pertinence par rapport aux performances des configurations du marché.

3

La numérotation des processeurs Intel® ne constitue pas une indication quantitative de leurs performances. Elle permet de différencier des modèles appartenant à une même famille (ligne) de processeurs, mais non pas à des familles différentes. Voir www.intel.com/content/www/fr/fr/processors/processor-numbers.html pour davantage d'informations.

4

Les compilateurs d'Intel peuvent optimiser ou non au même degré les microprocesseurs non Intel pour les optimisations qui ne sont pas uniques aux microprocesseurs Intel®. Ces optimisations comprennent les jeux d’instructions SSE2 et SSE3 ainsi que d’autres optimisations. Intel ne fournit aucune garantie quant à la disponibilité, la fonctionnalité ou l'efficacité des optimisations sur des microprocesseurs d'autres fabricants.

Dans ce produit, les optimisations dépendantes du processeur sont conçues pour les processeurs Intel. Certaines optimisations non spécifiques à la microarchitecture Intel® sont réservées aux microprocesseurs Intel. Reportez-vous au guide de l'utilisateur et au guide de référence applicable pour obtenir davantage d'informations concernant les jeux d'instructions spécifiques couverts par cet avis.

Avis révision nº 20110804

5

Les différentes architectures matérielles peuvent nécessiter un code source différent. Les résultats sont basés sur les meilleurs efforts d'Intel pour utiliser un code optimisé pour une exécution sur toutes les architectures en réalisant le même travail. Les optimisations de code à venir peuvent entraîner des résultats différents.

Dans ce produit, les optimisations dépendantes du processeur sont conçues pour les microprocesseurs Intel®. Certaines optimisations non spécifiques à la microarchitecture Intel® sont réservées aux microprocesseurs Intel. Reportez-vous au guide de l'utilisateur et au guide de référence applicable pour obtenir davantage d'informations concernant les jeux d'instructions spécifiques couverts par cet avis.

Avis révision nº 20110804