此網站不建議使用您目前的瀏覽器版本。
請考慮透過下列其中一個連結,升級到最新版的瀏覽器。

Competitive Performance Summary with Intel® Xeon Phi™ Product Family

Unlock deeper insights to solve your most complex data challenges faster with the Intel® Xeon Phi™ processor, a foundational element of Intel® Scalable System Framework (Intel® SSF)

Solve Your Complex Challenges Faster

The Intel® Xeon Phi™ processor is a true evolution in design and architecture that delivers the performance of an accelerator with the benefits of a server-class processor for your most demanding tasks. Unlock deeper insights to solve your complex data challenges faster.

Designed from the ground up to eliminate bottlenecks, the Intel Xeon Phi processor is the Intel’s first bootable host processor specifically designed for highly parallel workloads, and the first to integrate both memory and fabric technologies.

Intel® Xeon Phi™ processor momentum guide ›

Competitive Performance Summary

Featuring up to 72 powerful and efficient cores with ultra-wide vector capabilities (Intel® Advanced Vector Extensions or AVX-512), the Intel® Xeon Phi™ processor raises the bar for highly parallel computing. With no dependency on the PCIe* bus, the versatile processor is capable of handling a wide variety of workloads and configurations that are not supported by accelerators.

  • Intel® Xeon Phi™ processor 7250 outperforms a hosted NVIDIA* GPU solution by up to 5x in performance—get to results faster.

Performance measurements as of June 14, 2016.

View configuration details ›

Competitive Performance Per Watt Summary

The higher degree of parallelism in the Intel® Xeon Phi™ processor results in greater compute per unit of energy consumed for highly-parallel applications. Integration of 16 GB of high-bandwidth memory delivers up to 490 GB/s of sustained memory bandwidth for memory-bound workloads, and the available dual-port Intel® Omni-Path Fabric further reduces solution cost, power and space utilization.

  • Intel® Xeon Phi™ processor 7250 outdoes a hosted NVIDIA* GPU solution by up to 8x in performance delivered per watt – lower energy costs.

Performance measurements as of June 14, 2016.

View configuration details ›

Competitive Performance Per Dollar Summary

The processor is binary-compatible with Intel® Xeon® processors, which allows it to run any x86 workload to optimize asset utilization across the data center. By contrast, special-purpose accelerators, like GPUs, often remain underutilized since applications are not always suitable or optimized for them. Standardizing on Intel® architecture means you can use a single programming model for your code, helping increase efficiency through a shared developer base and code reuse.

  • Intel® Xeon Phi™ processor 7250 exceeds a hosted NVIDIA* GPU solution by up to 9x in performance delivered per dollar – realize compelling value.

Performance measurements as of June 14, 2016.

View configuration details ›

Configuration Details as of June 14, 2016

System Overview

1-Node, 1P Intel® Xeon Phi™ Processor 7210

1-Node, 1P Intel® Xeon Phi™ Processor 7250

1-Node, 2P Intel® Xeon® Processor E5-2697 v4
(45M Cache, 2.30 GHz, 18 Cores) + NVIDIA Tesla* K80 (All other comparisons)

1-Node, 2P Intel® Xeon® Processor E5-2699 v4
(55M Cache, 2.20 GHz, 22 Cores) + NVIDIA Tesla K80 (Financial Services benchmarks)

System Description

Intel® Server System LADMP2312KXXX41

Intel Server System LADMP2312KXXX41

Intel® Server Board S2600WT2R, BMC 1.33.9832, FRU/SDR Package 1.09 or Colfax* CX1350s-XK6

Supermicro* SYS-1028GR-TR server

Processor Information

Intel® Xeon Phi™ Processor 7210 (16 GB MCDRAM, 1.3 GHz, 64 Cores)

Intel® Xeon Phi™ Processor 7250 (16 GB MCDRAM, 1.4 GHz, 68 Cores)

Intel® Xeon® Processor E5-2697 v4 (45M Cache, 2.30 GHz, 18 Cores)

Intel® Xeon® Processor E5-2699 v4 (55M Cache, 2.20 GHz, 22 Cores)

Number of Sockets

1

1

2

2

Number of Cores / Threads

64 / 256

68 / 272

18 / 36

22 / 44

Memory Configuration

96 GB DDR4-2133 Quadrant Cluster/ MCDRAM Flat mode

96 GB DDR4-2400 Quadrant Cluster/ MCDRAM Flat mode

128 GB DDR4-2400

256 GB DDR4-2133

Operating System

Red Hat Enterprise Linux* 6.7 (Santiago)

Red Hat Enterprise Linux 6.7 (Santiago) or CentOS 7.2 (Financial Services benchmarks)

Red Hat Enterprise Linux 6.7 (Santiago)

Red Hat Enterprise Linux 7.1 (Maipo)

Other Hardware/ Software

Intel® Compiler 16.0.2, Intel® MPI 5.1.2.150 One NVIDIA Tesla* K80 GPUs, NVIDIA CUDA* 7.5.17 (Driver: 352.39), ECC enabled, persistence mode enabled

Intel® Hyper-Threading Technology

Yes

Yes

Yes

Yes

Intel® Turbo Boost Technology 2.0

Yes

Yes

Yes

Yes

Data Source

Measured by Intel – please contact your Intel sales representative for more information.

Additional information: 1 2 3 4 5 6 7 8

產品與效能資訊

1

效能測試中使用的軟體與工作負載,可能只有針對 Intel® 微處理器的執行效能最佳化。效能測試 (例如 SYSmark 與 MobileMark) 使用特定的電腦系統、元件、軟體、作業及功能進行評量。這些因素若有任何變更,可能會導致不同的結果。考慮購買時,為了協助您充分評估,您應該參考其他資訊及效能測試,包括該產品結合其他產品使用時的效能表現。如需詳細資訊,請參閱 www.intel.com/performance

2

請參閱本公司的 Optimization Notice (http://software.intel.com/en-us/articles/optimization-notice),以取得有關 Intel® 軟體產品的效能和最佳化選項的詳細資訊。

3

此處提供的所有資訊可能變更,恕不另行通知。若要取得最新的 Intel 產品規格與發展藍圖,請聯絡您的 Intel 代表。

4

Intel® 技術的功能與優勢取決於系統配置,而且可能需要支援的硬體、軟體或服務啟動。任何電腦系統都不可能保證絕對安全。進一步瞭解,請前往 Intel.com,或是向 OEM 或零售商查詢。

5

Intel® processors of the same SKU may vary in frequency or power as a result of natural variability in the production process.

6

所述之成本降低情境,用意是要提供範例,指出搭載特定 Intel® 處理器的產品,在特定情況與配置,可能會如何影響未來各項成本以及提供成本節省。實際情況可能有所差異。對於各項成本,或是成本降低幅度,Intel 不提供任何保證。

7

對於本文件提及的各項第三方效能標竿或網站,Intel 並不會控制或稽核其設計或執行。Intel 鼓勵所有的客戶參閱提及的網站,或是其他發佈類似評量效能標竿報告的網站,確認提及的效能標竿是否正確,並且反映出可購買系統的效能表現。

8

功能與優勢可能需要支援的系統,搭配第三方廠商硬體、軟體或服務。效能會有所不同,取決於您的具體配置。請向您的系統供應商查詢。