We’ve often talked about how Red Hat Enterprise Linux was fueling top benchmark results on the latest generation of Intel Xeon processors.
Today, we’re pleased to say that Red Hat Enterprise Linux has again demonstrated how an OS platform can handle the rigors of highly multi-threaded processing and on-demand scaling. In the most recent set of benchmarks done with one of the latest cutting edge processors, the Intel Xeon processor E7 v2 family, Red Hat Enterprise Linux claims 10 new world records. Here are these leading results at a glance:
Benchmark Name |
Benchmark Result |
OS version |
System Description |
Significance of Results |
SPECint_rate_base2006 |
1170 |
Red Hat Enterprise Linux 6.5 |
Cisco UCS B260 M4 (2 x Intel Xeon processor E7-2890 v2, 2.8 GHz) |
World Record 2-socket |
SPECint_rate_base2006 |
2320 |
Red Hat Enterprise Linux 6.5 |
Cisco UCS C460 M4 (4 x Intel Xeon processor E7-4890 v2, 2.8 GHz)Dell PowerEdge* R920 (4 x Intel Xeon processor E7-4890 v2, 2.8 GHz)Huawei RH5885H* v3 (4 x Intel® Xeon® processor E7-4890 v2, 2.8 GHz) |
World Record 4-socket (3 way tie) |
SPECint_rate_base2006 |
4570 |
Red Hat Enterprise Linux 6.5 |
Fujitsu PRIMEQUEST 2800E (8 x Intel Xeon processor E7-8890 v2, 2.8 GHz) |
World Record 8-socket |
SPECfp_rate_base2006 |
865 |
Red Hat Enterprise Linux 6.4 |
Cisco UCS B260 M4 (2 x Intel Xeon processor E7-2890 v2, 2.8 GHz) |
World Record 2-socket |
SPECfp_rate_base2006 |
1730 |
Red Hat Enterprise Linux 6.5 |
Huawei RH5885H* v3 (4 x Intel® Xeon® processor E7-4890 v2, 2.8 GHz) |
World Record 4-socket |
SPECfp_rate_base2006 |
3240 |
Red Hat Enterprise Linux 6.5 |
Fujitsu PRIMEQUEST 2800E (8x Intel Xeon processor E7-8890 v2, 2.8 GHz) |
World Record 8-socket |
SPECompG2012 | 8.91 baseline score60 OpenMP threads |
Red Hat Enterprise Linux 6.4 |
Cisco UCS B260 M4 (2 x Intel Xeon processor E7-2890 v2, 2.8 GHz) |
World Record 2-socket |
SPECompG2012 | 17.9 baseline score120 OpenMP threads |
Red Hat Enterprise Linux 6.4 |
Cisco UCS C460 M4 (4 x Intel Xeon processor E7-4890 v2, 2.8 GHz) |
World Record 4-socket |
SPECjbb2013 Multi-JVM | 177,753 max jOPS65,529 critical-jOPS |
Red Hat Enterprise Linux 6.4 |
NEC Express5800*/A2040b (4 x Intel® Xeon® processor E7-4890 v2, 2.8 GHz) |
World Record 4-socket |
SPECvirt_sc2013 | 2,081 @ 116VMs |
Red Hat Enterprise Linux 6.5 with KVM |
IBM System X3850 X6 (4 x Intel Xeon processor E7-4890 v2, 2.8 GHz) |
Overall World Record |
Red Hat Enterprise Linux has distanced itself from the rest of the platform pack when it comes to highly resource-intensive and scalable workloads that most modern enterprises use on a daily basis. While these workloads are generally diverse, they fundamentally are performing two types of operations: integer or floating point. Integer-intensive applications are usually transactional in nature and include applications such as databases, accounting, and financial software. On the other hand, engineering and science applications, as well as graphics intensive applications for modeling and computer aided engineering, all almost exclusively rely upon floating point operations.
Regardless of application type, Red Hat Enterprise Linux distinguished itself by showing industry leading capability and excellent scalability on both integer and floating point intensive suites of SPEC CPU2006 benchmarks while setting six world records (see table above). The two charts below offer another view into these six data points, depicting how the superior memory management and scheduling algorithms inherent to Red Hat Enterprise Linux enabled near linear scalability across systems of increasing size.
Another worthwhile observation is around the reproduction and repetition of these results. As evident from a 3-way tie on SPECint_rate_base2006 benchmark, Red Hat Enterprise Linux delivered the same performance across similarly configured systems from multiple vendors demonstrating that OS capabilities are uniformly available across multiple system designs.
While all of these world record numbers are impressive, it’s important to note that Red Hat Enterprise Linux isn’t just an extreme computing platform – we believe that it is the perfect platform for every enterprise workload, meeting the needs of today while scaling to meet even the most extreme computing demands of tomorrow.
All results as of February 20th 2014. SPEC and the benchmark names SPECcpu, SPECjbb, SPEComp and SPECvirt_sc are registered trademarks of the Standard Performance Evaluation Corporation. For more information about SPEC and its benchmarks see: www.spec.org
关于作者
Yan Fisher is a Global evangelist at Red Hat where he extends his expertise in enterprise computing to emerging areas that Red Hat is exploring.
Fisher has a deep background in systems design and architecture. He has spent the past 20 years of his career working in the computer and telecommunication industries where he tackled as diverse areas as sales and operations to systems performance and benchmarking.
Having an eye for innovative approaches, Fisher is closely tracking partners' emerging technology strategies as well as customer perspectives on several nascent topics such as performance-sensitive workloads and accelerators, hardware innovation and alternative architectures, and, exascale and edge computing.
产品
工具
试用购买与出售
沟通
关于红帽
我们是世界领先的企业开源解决方案供应商,提供包括 Linux、云、容器和 Kubernetes。我们致力于提供经过安全强化的解决方案,从核心数据中心到网络边缘,让企业能够更轻松地跨平台和环境运营。