World Record Benchmark Performance

Bring your business’s best ideas to life by transforming big data and real-time analytics into new business opportunities while ensuring the reliability and uptime of the most business-critical services with the Intel® Xeon® processor E7-8800/4800 v4 product families.

Turn Real-Time Analytics into Real Business Advantage

Leadership Performance

Accelerate performance across the data center to deliver real insights with real value in real time for the largest workloads running on fast, high-performing infrastructure. Faster transaction speeds and accelerated operations mean real-time responsiveness, quick results, and increased productivity. Offering up to 24 cores and 48 threads per processor, a 33 percent increase compared to the prior generation, and up to 60MB last level cache along with core improvements deliver:

  • Up to double the queries-per-hour answered enabling smarter decisions for businesses analyzing their sales and customer data.1 Ad-hoc cost per query has dropped from $21 to just $0.38 in just the past six-years—a 98% cost reduction, while queries-per-hour capability has increased by 19x.2
  • Up to 1.3x average performance across key industry-standard workloads.3
  • Up to 35% more virtual machines (VMs) and infrastructure applications throughput supported with same service-level agreement level to help IT grow line-of-business (LOB) heterogeneous needs.4
  • Use 1/3rd the servers for equivalent performance to lower operational expenses by replacing 4–5 year old installed platforms5, savings include lower network and server maintenance costs by up to 92%, lower utilities costs by up to 73%, and lower annual software licensing fees by up to 67%.

Platform Performance Highlights

Over 40 new world record performance benchmark results with the best Intel® Xeon® processor E7-8800/4800 v4 product families-based platforms (as of 30 March 2017).

Partner Platform Segment - Benchmark Importance
 

Cisco* UCS C460 M4

General Computing:
SPECint*_base2006

4-socket world record

Cisco UCS B260 M4

Technical Computing: 
SPECompG*_2012

2-socket world record

Cisco UCS C460 M4

Big Data Analytics:
TPC Benchmark* H @ 10000GB non-cluster

4-socket world record

Partner Platform Segment - Benchmark

Importance

Dell PowerEdge* R930 Business Processing:
SAP Sales and Distribution 2-tier*
4-socket world record Linux*
Dell PowerEdge R930 Big Data Analytics:
SAP BW-AML* @ 2 billion (2B) initial
records – Adv. Query Nav Steps/Hour
World record (Adv. Query Nav Steps/Hour) @ 2B
Dell PowerEdge R930 Big Data Analytics:
SAP BW-AML* @ 4 billion (4B) initial
records – Adv Query Nav Steps/Hour
and Total Runtime Delta Load/Trans.
Test
World record (Adv Query Nav Steps/Hour
and Total Runtime Delta Load/Trans. Test) @ 4B

Partner Platform Segment - Benchmark Importance
Fujitsu PRIMEQUEST* 2800E3 General Computing:
SPECint*_rate_base2006
2-socket world record
8-socket world record
Fujitsu PRIMEQUEST 2800E3 Technical Computing:
SPECfp*_rate_base2006
8-socket world record
Fujitsu PRIMERGY* RX4770 M3 Technical Computing:
SPECfp*_base2006
4-socket world record tie
Fujitsu PRIMEQUEST 2800E3 Infrastructure/Virtualization:
VMware* VMmark* V2.5.2 Performance 
Overall world record
Fujitsu PRIMERGY* RX4770 M3 Infrastructure/Virtualization:
VMware* VMmark* V2.5.2 Performance (matched pair)
4-socket world record
Fujitsu PRIMERGY RX4770 M3 Infrastructure/Virtualization:
VMware* VMmark* V2.5.2 Performance with Server Power
4-socket world record

Partner Platform Segment - Benchmark Importance
HPE ProLiant* DL580 Gen9 Technical Computing:
SPECfp*_base2006
4-socket world record tie
HPE ProLiant DL580 Gen9 Actian Vector 5.0 Big Data Analytics:
TPC Benchmark* H @ 3000GB non-cluster
4-socket world record
HPE Integrity* Superdome X Business Processing:
SAP* Sales and Distribution 2-tier
16-socket world record
HPE Integrity Superdome X Server-side Java*:
SPECjbb*2015 MultiJVM Max-jOPS
overall world record
16-socket world record
HPE CS500 (DL580 GEN9) Big Data Analytics:
SAP BW-AML* @ 2 billion (2B) initial
records – Normalized Mean Runtime Single
Query Test, Total Runtime Delta Load/Trans. Test
World records (Normalized
Mean Runtime Single Query Test,
Total Runtime Delta Load/Trans. Test)
@ 2B
HPE Integrity Superdome X
HPE Integrity Superdome X
HPE Synergy* 680 Gen 9
Server-side Java*:
SPECjbb*2015 MultiJVM Critical-jOPS
Overall world record
8 & 16-socket world record
4-socket world record

HPE ProLiant DL580 Gen9
HPE Synergy* 680 Gen9
Server-side Java*:
SPECjbb*2015 Composite Max-jOPS
Overall world records (tied)
HPE Synergy 680 Gen9 Server-side Java*:
SPECjbb*2015 Composite Critical-jOPS
Overall world record
HPE Synergy 680 Gen9 Server-side Java*:
SPECjbb*2015 Distributed Max-jOPS
Overall world record
HPE Synergy 680 Gen9 Server-side Java*:
SPECjbb*2015 Distributed Critical-jOPS

Overall world record

HPE Integrity MC990 X Technical Computing:
SPECompG*_2012
8-socket world record

Partner Platform Segment - Benchmark Importance
Hitachi Compute Blade* 520XB3 Big Data Analytics: 
SAP BW-AML* @ 4 billion (4B) initial
records – Normalized Mean Runtime Single Query Test
World record (Normalized Mean
Runtime Single Query Test) @ 4B 

Partner

Platform

Segment - Benchmark

Importance

Huawei* RH5885H V3

General Computing:
SPECint*_rate_base2006
4-socket world record
Huawei* Kunlun 9016 General Computing:
SPECint*_rate_base2006
16-socket world record
Huawei RH5885H V3 Technical Computing:
SPECfp*_rate_base2006
2-socket world record
4-socket world record
Huawei RH8100 V3 Server-side Java*:
SPECjbb*2015 Multi-JVM Max-jOPS
8-socket world record

Partner

Platform

Segment - Benchmark

Importance

Lenovo System* x3950 X6 General Computing:
SPECint*_base2006

8-socket world record

Lenovo System x3950 X6 Technical Computing:
SPECfp*_base2006
8-socket world record
Lenovo System x3950 X6 Business Processing:
SAP Sales and Distribution 2-tier*
8-socket world record Windows*
Lenovo System x3850 X6 Big Data Analytics:
SAP* BW edition for SAP HANA* @
1.3 billion (1.3B) initial records -
(all 3 key performance Indicators)
Overall world record (Total Runtime
of Data Load/Transformation, Query
Executions Per Hour, Total Runtime
of Complex Query Phase) @ 1.3B
Lenovo System x3950 X6 Infrastructure/Virtualization:
SPECvirt_sc*2013
Overall world record
Lenovo System x3850 X6 Infrastructure/Virtualization:
SPECvirt_sc*2013
4-socket world record
Lenovo System x3850 X6 Business Processing:
TPC Benchmark* E
4-socket world record
Lenovo System x3850 X6 Big Data Analytics:
STAC-M3* Shasta Suite
Overall world record
Shasta Suite (15/15 required)

Partner Platform Segment - Benchmark Importance

SGI* UV 300

Technical Computing: 
SPECompG*_2012

Overall world record
32-socket world record
16-socket world record

World Record Configurations

Results and configurations as of 30 March 2017

Cisco:

Claim based on best-published four-socket SPECint*_base2006 result published at https://www.spec.org/cpu2006/results/ as of 30 March 2017. New configuration: 1-Node, 4 x Intel® Xeon® processor E7-8891 v4 on Cisco Systems with 512 GB total memory on SUSE Linux Enterprise Server* 12 using C/C++: Version 16.0.0.101 of Intel® C++ Studio XE for Linux. Source: https://www.spec.org/cpu2006/results/res2016q2/cpu2006-20160517-41361.html, SPECint*_base2006 score: 71.5.

Claim based on best-published two-socket SPECompG*_base2012 result published at https://www.spec.org/omp2012/results as of 30 March 2017. New configuration: 1-Node, 2 x Intel® Xeon® processor E7-8890 v4 on Cisco* Systems UCS B260 M4 with 256 GB total memory on Red Hat Enterprise Linux* 6.7 kernel 2.6.32-573 using C/C++/Fortran: Version 16.0.0.109 of Intel® Composer for Linux Build 20150815. Source: https://www.spec.org/omp2012/results/res2016q2/omp2012-20160511-00073.html, SPECompG*_base2012 score: 13.4 with 96 base OpenMP threads.

Claim based on best-published four-socket TPC Benchmark* H @ 10,000GB non-cluster results published at http://www.tpc.org/tpch/results/tpch_perf_results.asp as of 30 March 2017. New configuration: 1-Node, 4 x Intel® Xeon® processor E7-8890 v4  (4P, 96C, 192T) on Cisco UCS* C460 M4 server with 6 TB total memory running Windows* Server 2016 Standard Edition and Microsoft* SQL Server 2016. Source: http://www.tpc.org/3326, Score: 1,115,298 QphH@10,000GB, $0.87/QphH@10,000GB available 28 November 2016.

Dell:

Claim based on best-published four-socket SAP SD 2-Tier on Linux* result published at http://global.sap.com/solutions/benchmark/sd2tier.epx as of 30 March 2017. New configuration: 2-tier, 4 x Intel® Xeon® processor E7-8894 v4 (96 cores/192 threads) on Dell PowerEdge* R930 with 1024 GB total memory on Red Hat Enterprise Linux* 7.2-kernel 3.10.0-327 using SAP Enhancement Package 5 for SAP ERP 6.0, SAP NetWeaver 7.20 pl510. Benchmark: SAP* SD 2-Tier enhancement package 5 for SAP ERP 6.0, source: Certificate #2017001, http://global.sap.com/solutions/benchmark/sd2tier.epx. Score: 43,300 benchmark users.

Claim based on best-published SAP BW-AML* @ 2 billion initial records result published at http://www.sap.com/solution/benchmark/appbm/netweaver.html as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processors E7-8890 v4 (96 cores/192 threads) on Dell PowerEdge* R930 with 1024 GB total memory on SUSE Linux Enterprise Server* 12 for SAP Applications using SAP HANA* 1.0, SAP NetWeaver* 7.50. Benchmark: SAP BW Advanced Mixed Load (BW-AML) @ 2 billion initial records, source: SAP Certificate #2016059, http://www.sap.com/solution/benchmark/appbm/netweaver.advanced-mixed-load-bw-aml.html. Score: Best advanced query navigation steps/hour (40,020).

Claim based on best-published SAP BW-AML* @ 4 Billion initial records result published at http://www.sap.com/solution/benchmark/appbm/netweaver.html as of 30 March 2017. 1-Node, 4 x Intel® Xeon® processors E7-8890 v4 (96 cores/192 threads) on Dell PowerEdge* R930 with 1536 GB total memory on SUSE Linux Enterprise Server* 11 SP4 for SAP Applications using SAP HANA* 1.0, SAP NetWeaver* 7.50. Benchmark: SAP BW Advanced Mixed Load (BW-AML) @ 4 Billion initial records. Source: SAP Certificate #2016054, http://www.sap.com/solution/benchmark/appbm/netweaver.advanced-mixed-load-bw-aml.html. Score: Best total advanced query navigation steps/hour (21,580) and best total runtime delta load/transformation test (294 seconds).

Fujitsu:

Claim based on best-published 2-socket SPECint*_rate_base2006 published at https://www.spec.org/cpu2006/results/ as of 30 March 2017. New configuration: 1-Node, 2x Intel® Xeon® processor E7-8894 v4 on Fujitsu PRIMEQUEST* 2800E3 with 256 GB total memory on SUSE Linux Enterprise Server* 12 SP2 4.4.21-68-default using Version 17.0.0.098 of Intel® C/C++ Compiler for Linux, source: https://www.spec.org/cpu2006/results/res2017q1/cpu2006-20170221-46674.pdf SPECint*_rate_base2006 score: 1860.

Claim based on best-published 8-socket SPECint*_rate_base2006 result published at https://www.spec.org/cpu2006/results/ as of 30 March 2017. New configuration: 1-Node, 8x Intel® Xeon® processor E7-8894 v4 on Fujitsu PRIMEQUEST* 2800E3 with 1024 GB total memory on SUSE Linux Enterprise Server* 12 SP2 4.4.21-68-default using Version 17.0.0.098 of Intel® C/C++ Compiler for Linux, source: https://www.spec.org/cpu2006/results/res2017q1/cpu2006-20170221-46671.pdf. SPECint*_rate_base2006 score: 7390.

Claim based on best-published 8-socket SPECfp*_rate_base2006 result published at https://www.spec.org/cpu2006/results/ as of 30 March 2017. New configuration: 1-Node, 8x Intel® Xeon® processor E7-8894 v4 on Fujitsu PRIMEQUEST* 2800E3 with 1024 GB total memory on SUSE Linux Enterprise Server* 12 SP2 4.4.21-68-default using Version 17.0.0.098 of Intel® C/C++ Compiler for Linux, Version 17.0.0.098 of Intel Fortran Compiler, source: https://www.spec.org/cpu2006/results/res2017q1/cpu2006-20170123-46533.pdf. SPECfp*_rate_base2006 score: 4770.

Claim based on best-published four-socket SPECfp*_base2006 result published at https://www.spec.org/cpu2006/results/ as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8891 v4 on Fujitsu Server PRIMERGY* RX4770 M3 with 512 GB total memory on SUSE Linux Enterprise Server* 12 SP1 (X86_64) Kernel 3.12.49-11-default using Version 16.0.0.101 of Intel® C/C++ Studio XE for Linux, Version 16.0.0.101 of Intel Fortran Studio XE for Linux, source: https://www.spec.org/cpu2006/results/res2017q1/cpu2006-20170307-46790.pdf. SPECfp*_base2006 score: 132.

Claim based on best-published VMmark* 2.5.2 performance result published at http://www.vmware.com/a/vmmark as of 30 March 2017. New configuration: Four hosts (realized as completely independent system partitions in two Fujitsu Server PRIMEQUEST 2800E3* systems) with each partition containing 4x Intel® Xeon® processor E7-8890 v4 (96 cores, 192 threads), 2048 GB memory, running VMware* ESXi 6.0.0 U2 Build 3825889 and vCenter Server 6.0.0 U1 Build 3018524, source: http://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vmmark/2016-08-16-Fujitsu-PRIMEQUEST2800E3.pdf. VMmark 2.5.2 Performance score: 114.47 @ 100 Tiles.

Claim based on best-published four-socket VMmark* 2.5.2 performance (matched pair) result published at http://www.vmware.com/a/vmmark as of 30 March 2017. New configuration: Fujitsu Server PRIMERGY RX4770 M3 with 4x Intel® Xeon® processor E7-8890 v4 (96 cores, 192 threads), 1024 GB memory, running VMware* ESXi 6.0.0 U2 Build 3620759 and vCenter Server 6.0.0 Build 3018524. Benchmark: VMmark 2.5.2 Performance, source: http://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vmmark/2016-06-21-Fujitsu-RX4770M3.pdf. VMmark 2.5.2 Performance score: 61.32 @ 52 Tiles.

Claim based on best-published four-socket VMmark* 2.5.2 performance with server power (matched pair) result published at http://www.vmware.com/a/vmmark as of 30 March 2017. New configuration: Fujitsu Server PRIMERGY RX4770 M3 in a matched pair configuration, each with four Intel® Xeon® processor E7-8890 v4 2.2 GHz (96 cores, 192 threads), 1024 GB memory, running VMware* ESXi 6.0.0 U2 Build 3620759 and vCenter Server 6.0.0 Build 3018524; Fibre Channel SAN storage; measured with Hioki* 3334 power meter. Benchmark: VMmark 2.5.2 Performance Server Power. Source: http://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vmmark/2016-06-21-Fujitsu-RX4770M3-serverPPKW.pdf. VMmark 2.5.2 Performance PPKW score: 30.4559 @ 52 Tiles.

HPE:

Claim based on best-published four-socket SPECfp*_base2006 results published at https://www.spec.org/cpu2006/results/ as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8891 v4 on Hewlett Packard Enterprise ProLiant* DL580 Gen9 with 512 GB total memory on SUSE Linux Enterprise Server* 12 SP1 using C/C++: Version 16.0.0.101 of Intel® C++ Studio XE for Linux; Fortran: Version 16.0.0.101 of Intel® Fortran Studio XE for Linux. Source: https://www.spec.org/cpu2006/results/res2016q2/cpu2006-20160517-41419.html, SPECfp_base2006 score: 132.

Claim based on best-published four-socket TPC Benchmark* H @ 3000GB non-cluster results published at http://www.tpc.org/tpch/results/tpch_perf_results.asp as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8890 v4 (4P, 96C, 192T) on Hewlett Packard Enterprise ProLiant* DL580 Gen9 with 3072 GB total memory on Red Hat Enterprise Linux* 7.2-kernel 3.10.0-327 using Actian* Vector Capacity Based VECTOR-PPL VW 5.0. Source: http://www.tpc.org/3323, Scores: 2,140,307 QphH @ 3000GB, $0.38/QphH@3000GB available 6 June 2016.

Claim based on best-published 16-socket SAP SD Standard Application Benchmark* result on Linux* published at http://global.sap.com/solutions/benchmark as of 30 March 2017. New configuration: 1-Node, 16x Intel® Xeon® processor E7-8890 v4 (384 cores/768 threads) on HPE Integrity* Superdome X with 4096 GB total memory on Red Hat Enterprise Linux* 7.2 using SAP Enhancement Package 5 for SAP ERP 6.0, SAP ASE 16. Benchmark: SAP* SD 2-Tier enhancement package 5 for SAP ERP 6.0, Source: Certificate #2016044, http://global.sap.com/solutions/benchmark/sd2tier.epx. Score: 117,611 benchmark users.

Claim based on best-published SAP BW-AML* @ 2 billion initial records results published at http://www.sap.com/solution/benchmark/appbm/netweaver.html as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processors E7-8894 v4 (96 cores/192 threads) on Hewlett Packard Enterprise CS500 (DL580 Gen9) with 1024 GB total memory on SUSE Linux Enterprise Server* 12 using SAP HANA* 1.0, SAP NetWeaver* 7.50. Benchmark: SAP BW Advanced Mixed Load (BW-AML) @ 2 billion initial records, source: SAP Certificate #2017008, http://www.sap.com/solution/benchmark/appbm/netweaver.advanced-mixed-load-bw-aml.html. Score: Best Normalized Mean Runtime Single Query Test of 6.66 seconds/billion records and Total Runtime Delta Load/Transformation Test of 234 seconds @ 2B.

Claim based on best-published 16-socket SPECjbb*2015 MultiJVM max-jOPS results published at https://www.spec.org/jbb2015/results/jbb2015multijvm.html as of 30 March 2017. New configuration: 1-Node, 16x Intel® Xeon® Processor E7-8890 v4 on Hewlett Packard Enterprise Integrity* Superdome X with 8192 GB total memory on SUSE Linux Enterprise Server* 12 SP1 using Java* HotSpot 64-bit Server VM, version 1.8.0_102. Source: https://www.spec.org/jbb2015/results/res2016q3/jbb2015-20160830-00113.html, SPECjbb2015 MultiJVM scores: 776,269 max-jOPS and 84,557 critical-jOPS.  

Claim based on best-published 16-socket SPECjbb*2015 MultiJVM critical-jOPS results published at https://www.spec.org/jbb2015/results/jbb2015multijvm.html as of 30 March 2017. New configuration: 1-Node, 16x Intel® Xeon® Processor E7-8890 v4 on Hewlett Packard Enterprise Integrity* Superdome X with 8192 GB total memory on SUSE Linux Enterprise Server* 12 SP1 using Java* HotSpot 64-bit Server VM, version 1.8.0_102. Source: https://www.spec.org/jbb2015/results/res2016q3/jbb2015-20160830-00114.html, SPECjbb2015 MultiJVM scores: 439,937 critical-jOPS and 700,536 max-jOPS.

Claim based on best-published 8-socket SPECjbb*2015 MultiJVM critical-jOPS results published at https://www.spec.org/jbb2015/results/jbb2015multijvm.html as of 30 March 2017. New configuration: 1-Node, 8x Intel® Xeon® Processor E7-8890 v4 on Hewlett Packard Enterprise Integrity* Superdome X with 4096 GB total memory on SUSE Linux Enterprise Server* 12 SP1 using Java* HotSpot 64-bit Server VM, version 1.8.0_102. Source: https://www.spec.org/jbb2015/results/res2016q3/jbb2015-20160830-00115.html, SPECjbb2015 MultiJVM scores: 237,725 critical-jOPS and 359,735 max-jOPS.

Claim based on best-published four-socket SPECjbb*2015 MultiJVM critical-jOPS results published at http://www.spec.org/jbb2015/results/jbb2015composite.html as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® Processor E7-8894 v4 on Hewlett Packard Enterprise Synergy 680 Gen9 with 2048 GB total memory on Red Hat Enterprise Linux* Server 7.3 using Java* HotSpot 64-bit Server VM, version 1.8.0_121. Source: http://www.spec.org/jbb2015/results/res2017q1/jbb2015-20170315-00154.html, SPECjbb2015 MultiJVM scores 138,466 critical-jOPS and 194,068 max-jOPS.

Claim based on best-published four-socket SPECjbb*2015 Composite max-jOPS results published at http://www.spec.org/jbb2015/results/jbb2015composite.html as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® Processor E7-8894 v4 on Hewlett Packard Enterprise ProLiant DL580 Gen9 with 512 GB total memory on SUSE Linux Enterprise Server* 12 SP1 using Java* HotSpot 64-bit Server VM, version 1.8.0_121. Source: http://www.spec.org/jbb2015/results/res2017q1/jbb2015-20170315-00150.html, SPECjbb2015 Composite scores 186,337 max-jOPS and 86,878 critical-jOPS.

 

Claim based on best-published four-socket SPECjbb*2015 Composite max-jOPS and critical-jOPS results published at https://www.spec.org/jbb2015/results/jbb2015composite.html as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® Processor E7-8894 v4 on Hewlett Packard Enterprise Synergy 680 Gen9 with 512 GB total memory on SUSE Linux Enterprise Server* 12 SP1 using Java* HotSpot 64-bit Server VM, version 1.8.0_121. Source: https://www.spec.org/jbb2015/results/res2017q1/jbb2015-20170315-00152.html, SPECjbb2015 Composite scores 186,337 max-jOPS and 88,538 critical-jOPS.

 

Claim based on best overall published SPECjbb*2015 Distributed max-jOPS results published at https://www.spec.org/jbb2015/results/jbb2015distributed.html as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8894 v4 on Hewlett Packard Enterprise Synergy 680 Gen9 with 512 GB total memory on SUSE Linux Enterprise Server* 12 SP2 using Java* HotSpot 64-bit Server VM, version 1.8.0_121. Source: https://www.spec.org/jbb2015/results/res2017q1/jbb2015-20170315-00155.html, SPECjbb2015 Distributed scores: 227,634 max-jOPS and 75,399 critical-jOPS.

Claim based on best overall published SPECjbb*2015 Distributed critical-jOPS results published at https://www.spec.org/jbb2015/results/jbb2015distributed.html as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8894 v4 on Hewlett Packard Enterprise Synergy 680 Gen9 with 2048 GB total memory on Rd Hat Enterprise Linux* Server 7.3, kernel 3.10.0-513 using Java HotSpot 64-bit Server VM, version 1.8.0_121. Source: https://www.spec.org/jbb2015/results/res2017q1/jbb2015-20170315-00153.html, SPECjbb2015 Distributed scores: 142,026 critical-jOPS and 196,434 max-jOPS.

Claim based on best-published 8-socket SPECompG*_base2012 results published at https://www.spec.org/omp2012/results as of 30 March 2017. New configuration: 1-Node, 8x Intel® Xeon® processor E7-8890 v4 on HPE Integrity MC990 X with 2048 GB total memory on Red Hat Enterprise Linux* Server 7.2 kernel 3.10.0-327.18.2.el7.x86_64 using Version 16.0.3.210 of Intel® C++ Studio XE for Linux, Version 16.0.3.210 of Intel® Fortran. Source: https://www.spec.org/omp2012/results/res2016q3/omp2012-20160718-00079.pdf. SPECompG_base2012 score: 37.5 with 384 base OpenMP threads.

Hitachi:

Claim based on best-published SAP BW-AML* @ 4 Billion initial records result published at http://www.sap.com/solution/benchmark/appbm/netweaver.html as of 30 March 2017. 1-Node, 4 x Intel® Xeon® processors E7-8890 v4 (96 cores/192 threads) on Hitachi Compute Blade 520XB3, with 1536 GB total memory on SUSE Linux Enterprise Server* 12 for SAP Applications using SAP HANA* 1.0, SAP NetWeaver* 7.50. Benchmark: SAP BW Advanced Mixed Load (BW-AML) @ 4 Billion initial records, Source: SAP Certificate #2016052, http://www.sap.com/solution/benchmark/appbm/netweaver.advanced-mixed-load-bw-aml.html. Score: Best normalized mean runtime single query test (5.56 seconds / billion records).

Huawei:

Claim based on best-published four-socket SPECint*_rate_base2006 result published at https://www.spec.org/cpu2006/results as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8894 v4 (24 cores, 48 threads) on Huawei* RH5885H V3 with 512 GB total memory on SUSE* Linux Enterprise Server 12 (x86_64) SP1, Kernel 3.12.49-11-default, Version 16.0.0.101 of Intel® C++ Studio XE for Linux, source: https://www.spec.org/cpu2006/results/res2017q1/cpu2006-20170123-46508.pdf. SPECint_rate_base2006 score: 3800.

Claim based on best-published 16-socket SPECint*_rate_base2006 result published at https://www.spec.org/cpu2006/results as of 30 March 2017. New configuration: 1-Node, 16x Intel® Xeon® processor E7-8890 v4 (24 cores, 48 threads) on Huawei Kunlun 9016 with 2048 GB total memory on SUSE* Linux Enterprise Server 12 (x86_64) SP1, Kernel 3.12.49-11-default, Version 16.0.0.101 of Intel® C++ Studio XE for Linux*, source: https://www.spec.org/cpu2006/results/res2017q1/cpu2006-20170302-46704.pdf. SPECint_rate_base score: 13800.

Claim based on best-published two-socket SPECfp*_rate_base2006 result published at https://www.spec.org/cpu2006/results as of 30 March 2017. New configuration: 1-Node, 2x Intel® Xeon® processor E7-8890 v4 (24 cores, 48 threads) on Huawei* RH5885H V3 with 256 GB total memory, Red Hat Enterprise Linux Server release 7.2 (Maipo) 3.10.0-327.e17.x86_64, Version 16.0.0.101 of Intel® C++ Studio XE for Linux*, Version 16.0.0.101 of Intel® Fortran Studio XE for Linux, source: https://www.spec.org/cpu2006/results/res2016q2/cpu2006-20160517-41370.pdf. SPECfp_rate_base2006 score: 1240.

Claim based on best-published four-socket SPECfp*_rate_base2006 result published at https://www.spec.org/cpu2006/results as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8894 v4 (24 cores, 48 threads) on Huawei* RH5885H V3 with 512 GB total memory on SUSE* Linux Enterprise Server 12 (x86_64) SP1, Kernel 3.12.49-11-default, Version 16.0.0.101 of Intel® C++ Studio XE for Linux*, Version 16.0.0.101 of Intel® Fortran Studio XE for Linux, source: https://www.spec.org/cpu2006/results/res2017q1/cpu2006-20170123-46525.pdf.. SPECfp_rate_base2006 score: 2460.

Claim based on best-published 8-socket SPECjbb*2015 MultiJVM Critical-jOPS result published at https://www.spec.org/jbb2015/results/ as of 30 March 2017. New configuration: 1-Node, 8x Intel® Xeon® processor E7-8890 v4 (24 cores, 48 threads) on Huawei* RH8100 V3 with 1024 GB total memory on SUSE* Linux Enterprise Server* 12 using Oracle Java* SE 8u92, Java HotSpot 64-bit Server VM, version 1.8.0_92. Source: https://www.spec.org/jbb2015/results/res2016q2/jbb2015-20160511-00092.html. SPECjbb*2015 - MultiJVM score: 411,801 max-jOPS and 119,520 critical-jOPS.

Lenovo:

Claim based on best-published 8-socket SPECint*_base2006 benchmark published at https://www.spec.org/cpu2006/results as of 30 March 2017. New configuration: 1-Node, 8x Intel® Xeon® processor E7-8891 v4 (10 cores, 20 threads) on Lenovo* System x*3950 X6 with 1024 GB memory on SUSE* Linux Enterprise Server 12 SP1, using Version 16.0.0.101 of Intel® C/C++ Studio XE for Linux*, Source: https://www.spec.org/cpu2006/results/res2016q3/cpu2006-20160809-43451.pdf. SPECint_base2006 score: 71.1.

Claim based on best-published 8-socket SPECfp*_base2006 benchmark published at https://www.spec.org/cpu2006/results as of 30 March 2017. New configuration: 1-Node, 8x Intel® Xeon® processor E7-8891 v4 (10 cores, 20 threads) on Lenovo* System x*3950 X6 with 1024 GB memory on SUSE* Linux Enterprise Server 12 SP1, using Version 16.0.0.101 of Intel® C/C++ Studio XE for Linux*, Version 16.0.0.101 of Intel® Fortran Studio XE for Linux, Source: https://www.spec.org/cpu2006/results/res2016q3/cpu2006-20160809-43447.pdf. SPECfp_base2006 score: 129.

Claim based on best-published 8-socket SAP SD Standard Application Benchmark* result on Microsoft Windows* published at http://global.sap.com/solutions/benchmark as of 30 March 2017. New configuration: 2-tier, 8x Intel® Xeon® processor E7-8894 v4 (192 cores/384 threads) on Lenovo System x3950 X6 with 2048 GB total memory on Windows Server* 2012 R2 Standard Edition using SAP Enhancement Package 5 for SAP ERP 6.0 and IBM DB2 10.1. Benchmark: SAP* SD 2-Tier enhancement package 5 for SAP ERP 6.0. Source: Certification #2017002, http://global.sap.com/solutions/benchmark/sd2tier.epx. Score: 77,098 benchmark users.

Claim based on best-published SAP* BW edition for SAP HANA Standard Application Benchmark* @ 1.3 billion initial records result published at http://global.sap.com/solutions/benchmark as of 30 March 2017. New configuration: 4x Intel® Xeon® processor E7-8894 v4 (96 cores/192 threads) on Lenovo System x3850 X6 with 2048 GB total memory on SUSE Linux Enterprise Server 11 using SAP HANA 1.0, SAP NetWeaver 7.50. Benchmark: SAP BW for SAP HANA @ 1.3B initial records, Source: Certification #2017003, http://www.sap.com/solution/benchmark/appbm/netweaver.sap-bw-edition-for-sap-hana-standard-application.html. Score: Runtime of Data Load/Trans (14,939 secs), Query Executions per Hour (4,273), Runtime of Complex Query (154 seconds).

Claim based on best-published 8-socket SPECvirt_sc*2013 results published at https://www.spec.org/virt_sc2013/results/ as of 30 March 2017. New configuration: 1-Node, 8x Intel® Xeon® processor E7-8890 v4 on Lenovo* System x3950 X6 with 4096 GB total memory on Red Hat Enterprise Linux* 7.2 (KVM). Source: https://www.spec.org/virt_sc2013/results/res2016q3/virt_sc2013-20160830-00062-perf.html, SPECvirt_sc2013 score: 8284 @ 462 VMs.

Claim based on best-published SPECvirt_sc*2013 results published at https://www.spec.org/virt_sc2013/results/ as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8890 v4 on Lenovo* System x3850 X6 with 2048 GB total memory on Red Hat Enterprise Linux* 7.2 (KVM). Source: https://www.spec.org/virt_sc2013/results/res2016q3/virt_sc2013-20160913-00063-perf.html, SPECvirt_sc2013 score: 4377 @ 247 VMs.

Claim based on best-published four-socket TPC Benchmark* E results published at http://www.tpc.org/tpce/results/tpce_perf_results.asp as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8890 v4 on Lenovo* System x3850 X6 with 4096 GB total memory on Windows Server* 2012 R2 Standard using SQL Server* 2016 Enterprise Edition. Availability: July 31, 2016. Source: www.tpc.org/4078. Score: 9,068 tpsE, at a price/performance ratio of $139.85 USD/tpsE*.

Claim based on best-published STAC-M3* 15 world records (Shasta Suite) at http://www.stacresearch.com/m3 as of 30 March 2017. New configuration: 1-Node, 4x Intel® Xeon® processor E7-8890 v4 on Lenovo* System x3850 X6 with 6144 GB total memory on Red Hat Enterprise Linux* 7.2-kernel 3.10.0-327 using Kx Systems* kdb+ 3.3 software. STAC-M3 Tick Analytics Benchmarks, Shasta Suite (15 of 15 benchmarks required) source: http://www.STACresearch.com/KDB160425.

SGI:

Claim based on best-published 32-socket SPECompG*_base2012 results published at https://www.spec.org/omp2012/results as of 30 March 2017. New configuration: 1-Node, 32x Intel® Xeon® processor E7-8867 v4 on SGI* UV 300 with 8192 GB total memory on SUSE Linux Enterprise Server* 12 SP1 using C/C++/Fortran: Version 16.0.3.210 of Intel® Composer XE for Linux, Build 20150415. Source: https://www.spec.org/omp2012/results/res2016q3/omp2012-20160613-00077.pdf, SPECompG_base2012 score: 84.5 with 513 base OpenMP threads.

Claim based on best-published 16-socket SPECompG*_base2012 results published at https://www.spec.org/omp2012/results as of 30 March 2017. New configuration: 1-Node, 16x Intel® Xeon® processor E7-8867 v4 on SGI* UV 300 with 4096 GB total memory on SUSE Linux Enterprise Server* 12 SP1 using C/C++/Fortran: Version 16.0.3.210 of Intel® Composer XE for Linux, Build 20160415, source: https://www.spec.org/omp2012/results/res2016q3/omp2012-20160613-00078.pdf, SPECompG_base2012 score: 57.0 with 512 base OpenMP threads. 

Additional information: 6 7 8 9 10 11 12 13 14

Infos sur le produit et ses performances

1

Requêtes jusqu'à 2,1 fois plus rapides reposant sur un facteur d'échelle TPC-H à 3 000 Go, comparant un système HPE ProLiant* DL580 9e génération équipé de processeurs Intel® Xeon® E7-8890 v4 (4P, 96C, 192T), d'une base de données Actian* Vector 5.0 enregistrant 2 140 307 QphH (requêtes par heure) à 3 000 Go, en comptant 0,38 $/QphH à 3 000 Go en date du 06/06/16 (source : http://www.tpc.org/3323) avec un système équipé de processeurs E7-8890 v3 (4P, 72C, 144T), enregistrant 1 071 018 QphH à 3 000 Go, en comptant 0,60 $/QphH à 3 000 Go en date du 01/06/16 (source : http://www.tpc.org/3322)

2

Jusqu'à 19 requêtes par heure avec un coût réduit de 98 % par requête, en comparant un serveur HPE* DL580 de 9e génération équipé de processeurs Intel® Xeon® E7-8890 v4 (4P, 96C, 192T), d'une base de donnée Actian* enregistrant 2 140 307 QphH à 3 000 Go, en comptant 0,38 $/QphH à 3 000 Go en date du 06/06/16 (source : http://www.tpc.org/3323) avec un système équipé de processeurs Intel® Xeon® X7460 (16P, 96C, 96T), d'une base de données SQL Server* 2008 Datacenter Edition R2 enregistrant 102 778 QphH à 3 000 Go, en comptant 21,05 $/QphH à 3 000 Go en date du 06/05/10 (source historique : http://www.tpc.org/3245)

3

Gain moyen de performances pouvant atteindre 1,3 selon les bancs d'essai standard (SPECjbb*2015 Multi-JVM Critical et Max jOPS, SPECint*_rate_base2006, SAP SD* 2-tier, SPECvirt_sc*2013 et TPC-E*), en comparant les meilleures publications de serveurs OEM 4 sockets sur SPEC.org, SAP.com/benchmarks et TPC.org. Consultez www.intel.com/E7v4Record pour plus d'informations.

4

Hypothèse de jusqu'à 1,35 machines virtuelles supplémentaires selon le banc d'essai SPECvirt_sc*2013, comparant 4 processeurs Intel® Xeon® E7-8890 v3 avec 1 nœud et mémoire totale de 2 To sur un système Red Hat Enterprise Linux* 6.6 utilisant Red Hat Enterprise Linux 6.6 (KVM). Source de données : http://www.spec.org/virt_sc2013/results/res2016q1/virt_sc2013-20160126-00041-perf.html, Banc d'essai : SPECvirt_sc* 2013, Score : 3 118 à 174 machines virtuelles pour 4 processeurs Intel® Xeon® E7-8890 v4 avec 1 nœud sur un système Hewlett Packard Enterprise avec une mémoire totale de 2 048 Go sur un noyau Red Hat Enterprise Linux* 7.2, 3.10.0-327. Source : https://www.spec.org/virt_sc2013/results/res2016q2/virt_sc2013-20160517-00050-perf.html, Banc d'essai : SPECvirt_sc* 2013, Score : 4 231 à 237 machines virtuelles, le score le plus élevé étant le meilleur

5

Réduction du déploiement de serveurs à hauteur de 1/3 avec estimation de la réduction des frais d'exploitation basée sur des hypothèses et des estimations formulées par l'outil Server Refresh Savings Estimator, comparant des serveurs à quatre sockets avec 8 ports GbE par serveur, un coût d'assistance logicielle de 1 399 $ par serveur et par an, un coût de validation des logiciels de 1 000 $ par serveur, un coût de 0,10 $/kWh avec une utilisation moyenne de 10 %, une réduction de 50 % des coûts de main-d'œuvre, 2 399 $ de frais de maintenance par serveur et par an, et 15 $ de maintenance réseau par serveur et par an :

  • 100x : processeur Intel® Xeon® E7-4870 (30 Mo de cache, 2,40 GHz, ancien nom de code « Westmere-EX »), score SPECint*_rate_base2006 estimé de 1 100 avec compilateur Intel® 12.1, puissance estimée de 392 W au repos/692 W en activité
  • à 33x : E7-8890 v4 (60 Mo de cache, 2,20 GHz, ancien nom de code « Broadwell-EX »), score SPECint*_rate_base2006 estimé de 1 100 avec compilateur Intel® 12.1, puissance estimée de 392 W au repos/692 W en activité
6

Les résultats des bancs d'essai ont été obtenus avant le déploiement de récents correctifs logiciels et mises à jour micrologicielles destinés à faire face aux failles de sécurité « Spectre » et « Meltdown ». L'installation de ces mises à jour peut rendre ces résultats inapplicables à votre appareil ou système. 

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 complète d'un produit, il convient de consulter d'autres tests et d'autres sources d'informations, notamment pour connaître le comportement de ce produit lorsqu'il est associé à d'autres composants. Consultez http://www.intel.fr/benchmarks à ce sujet.

7

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.

8

Avis sur l'optimisation : 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 Intel® Streaming SIMD Extensions 2 (Intel® SSE2), Intel® SSE3, et Supplemental Streaming SIMD Extensions 3 (SSSE3) et 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. Révision d'avis nº 20110804

9

Les fonctionnalités et avantages des technologies Intel® dépendent de la configuration du système et peuvent nécessiter du matériel et des logiciels compatibles, ou l'activation de services. Les performances varient en fonction de la configuration du système. Aucun ordinateur ne saurait être totalement sécurisé. Consultez l'éditeur de vos logiciels, le constructeur ou le revendeur de votre système ou apprenez-en plus à http://www.intel.com/software/tsx.

10

Les processeurs Intel® de même référence peuvent présenter des variations en termes de fréquence ou de puissance, dû au processus de production.

11

SPEC, SPECint, SPECfp, SPECjbb, SPECjEnterprise, SPECvirt_sc, SPECpower_ssj, SPECompM et SPECompL sont des marques commerciales de la SPEC (Standard Performance Evaluation Corporation).

12

TPC Benchmark, TPC-C, tpmC, TPC-H, QphH, TPC-E et tpsE sont des marques commerciales du Transaction Processing Performance Council. Consultez www.tpc.org à ce sujet.

13

SAP et SAP NetWeaver sont des marques déposées de SAP AG en Allemagne et dans plusieurs autres pays. Consultez www.sap.com/benchmark pour plus d'informations.

14

VMware est une marque déposée et VMmark est une marque commerciale de VMware, Inc.