Do More with Your Cloud Investment
Deliver the powerful performance per dollar you need on Intel® technology. Critical data-heavy workloads such as database, high performance computing (HPC), and web perform better at a lower total cost of ownership on Intel® architecture-based clouds.1 2 3 4
4.15x Higher Performance
Up to 4.15x higher performance/$ on High Performance LINPACK* and LAMMPS*.2
1.74x Better Performance
Better performance/$ on Server-Side Java and 1.74x better performance /$ on WordPress PHP/HHVM.3
2.25x Higher Performance
Up to 2.25x higher performance/$ for memory bandwidth applications.4
Features and Benefits
Leadership Performance
With built-in acceleration for AI, software optimizations and socket scalability, Customers are delivering leading edge solutions on the Intel® Xeon® Scalable platform for your next modern datacenter.
AI & Analytics
Seamlessly build and deploy AI applications at scale. It all starts on a foundation of Intel® Xeon® Scalable processors, accelerated with Intel-optimized AI software.
High Performance Computing
Designed to solve your biggest challenges faster and with greater efficiency. The Intel® Xeon® Scalable processors with built-in HPC and AI acceleration help improve application performance to power new breakthroughs and speed time to insight.
Graphics & Accelerators
Intel delivers a new generation of architecture benefits to service providers with the Intel® Server GPU. It is purpose-built for the needs of differentiating services in the visual cloud era that will help providers rapidly scale out new offerings to meet new and escalating customer demands.
Scalable Performance
Intel® Xeon® Scalable processors are designed for the performance demands of complex, data-demanding workloads such as in-memory databases, and real-time business analytics.
Network Communications
Intel® Ethernet Network Adapters, Controllers, and Accessories enable agility within the data center to effectively deliver services efficiently and cost-effectively.
Enabled Software
Optimized independent software vendor (ISV) applications for Intel® Xeon® Scalable processors—increasing the speed and performance of vital applications.
Verified & Optimized Solutions
Accelerate your deployment of workload-optimized data center infrastructure with rigorously benchmark tested and verified solutions optimized for real-world performance.
Benchmarks
Click on the specific Intel® products and technologies to get performance-related benchmarks, white papers, and videos.
Compute
Check out the performance benchmarks for Intel® Xeon® Scalable processor family-based servers.
Storage
Get the benchmarks that show how Intel® technologies can optimize data storage.
Network
Review Intel's high-density virtualization, low-latency, and high-throughput benchmarks for networking.
Applications and Software
Find out how software solutions are optimized to work better with Intel infrastructure.
- 2nd Gen Intel® Xeon® Scalable processors & Intel® Optane™ persistent memory
- Intel® AI Analytics Toolkit and popular deep learning frameworks optimized for Intel® architecture
- DAOS: High-Performance Software Defined Storage Over Persistent Memory
- Enterprise IT using Intel® Cache Acceleration Software (Intel® CAS)
- Intel® Xeon® Scalable processors and software: Better together
Explore Data Center Resources
IT insights, guidance, and advice—for data center pros.
IT@Intel: Insight for Business Growth
Get innovative ideas and groundbreaking strategies from IT industry experts at Intel.
Resources for IT Leaders
Get the insights from Intel and other companies, plus social media to keep you informed and moving your IT projects forward.
產品與效能資訊
效能因使用情形、配置和其他因素而異。請造訪 www.Intel.com.tw/PerformanceIndex 進一步瞭解。
效能結果係依配置中所示日期的測試為準,且可能無法反映所有公開可用的安全性更新。請查看組態公開資料以獲得詳細資訊。沒有產品或元件能提供絕對的安全性。
Intel® 技術可能需要搭配啟用的硬體、軟體或服務啟動。
您的成本和成果可能有所落差。
Intel 並不控制或審核第三方的資料。您應該參考其他來源以評估準確性。
結果是根據 Intel P2CA 使用截至 2019 年 1 月 12 日的 AWS 定價(每小時的金額,標準 1 年期,無預付費用)所計算而得。
每元效能測試採用 AWS* EC2 M5 和 M5a 實例 (https://aws.amazon.com/ec2/instance-types/),比較 96 vCPU Intel® Xeon® 可擴充處理器的每元效能與 AMD EPYC* 處理器的每元效能。
工作負載:LAMMPS*
結果:AMD EPYC 每元效能 = 基準線 1;Intel® Xeon® 可擴充處理器的每元效能 = 4.15 倍(越高越好)
HPC 材料科學 – LAMMPS(越高越好):
AWS M5.24xlarge (Intel) 實例,LAMMPS 版本:2018-08-22(代碼:https://lammps.sandia.gov/download.html),工作負載:水 – 512K 粒子,Intel ICC 18.0.3.20180410,適用於 Linux* OS 的 Intel® MPI Library,版本 2018 Update 3 Build 20180411,48 個 MPI 階層,RedHat* Enterprise Linux 7.5,核心 3.10.0-862.el7.x86_64,OMP_NUM_THREADS=2,每秒時步分數 137.5,由 Intel 於 2018 年 10 月 31 日測量。
AWS M5a.24xlarge (AMD) 實例,LAMMPS 版本:2018-08-22(代碼:https://lammps.sandia.gov/download.html),工作負載:水 – 512K 粒子,Intel ICC 18.0.3.20180410,適用於 Linux* OS 的 Intel® MPI Library,版本 2018 Update 3 Build 20180411,48 個 MPI 階層,RedHat* Enterprise Linux 7.5,核心 3.10.0-862.el7.x86_64,OMP_NUM_THREADS=2,每秒時步分數 55.8,由 Intel 於 2018 年 11 月 7 日測量。
AMD 支援 AVX2 的變更(AMD 僅支援 AVX2,因此需要這些變更):
sed -i 's/-xHost/-xCORE-AVX2/g' Makefile.intel_cpu_intelmpi
sed -i 's/-qopt-zmm-usage=high/-xCORE-AVX2/g' Makefile.intel_cpu_intelmpi
工作負載:高效能 Linpack*
結果:AMD EPYC 每元效能 = 基準線 1;Intel® Xeon® 可擴充處理器的每元效能 = 4.15 倍(越高越好)
HPC Linpack(越高越好):
AWS M5.24xlarge (Intel) 實例,HP Linpack 版本 2.2(https://software.intel.com/en-us/articles/intel-mkl-benchmarks-suite 目錄:benchmarks_2018.3.222/linux/mkl/benchmarks/mp_linpack/bin_intel/intel64),Intel ICC 18.0.3.20180410 搭配 AVX512,適用於 Linux* OS 的 Intel® MPI Library,版本 2018 Update 3 Build 20180411,RedHat* Enterprise Linux 7.5,核心 3.10.0-862.el7.x86_64,OMP_NUM_THREADS=24,2 個 MPI 程序,分數 3152 GB/s,由 Intel 於 2018 年 10 月 31 日測量。
AWS M5a.24xlarge (AMD) 實例,HP Linpack 版本 2.2,(HPL 來源:http://www.netlib.org/benchmark/hpl/hpl-2.2.tar.gz:版本 2.2;使用 icc (ICC) 18.0.2 20180210 編譯與連結至 BLIS 程式庫版本 0.4.0;https://github.com/flame/blis;Addt’l 編譯器旗標:-O3 -funroll-loops -W -Wall –qopenmp;make arch=zen OMP_NUM_THREADS=8;6 個 MPI 程序。), Intel ICC 18.0.3.20180410 搭配 AVX2,適用於 Linux* OS 的 Intel® MPI Library,版本 2018 Update 3 Build 20180411,RedHat* Enterprise Linux 7.5,核心 3.10.0-862.el7.x86_64,OMP_NUM_THREADS=8,6 個 MPI 程序,分數 677.7 GB/s,由 Intel 於 2018 年 11 月 7 日測量。
結果是根據 Intel P2CA 使用截至 2019 年 1 月 12 日的 AWS 定價(每小時的金額,標準 1 年期,無預付費用)所計算而得。
每元效能測試採用 AWS* EC2 M5 和 M5a 實例 (https://aws.amazon.com/ec2/instance-types/),比較 96 vCPU Intel® Xeon® 可擴充處理器的每元效能與 AMD EPYC* 處理器的每元效能。
工作負載:伺服器端 Java* 1 JVM
結果:AMD EPYC 每元效能 = 基準線 1;Intel® Xeon® 可擴充處理器的每元效能 = 1.74 倍(越高越好)
伺服器端 Java(越高越好):
AWS M5.24xlarge (Intel) 實例,Java 伺服器效能標竿無 NUMA 繫結,2JVM,OpenJDK 10.0.1,RedHat* Enterprise Linux 7.5,核心 3.10.0-862.el7.x86_64,每秒傳輸分數 101767,由 Intel 於 2018 年 11 月 16 日測量。
AWS M5a.24xlarge (AMD) 實例,Java 伺服器效能標竿無 NUMA 繫結,2JVM,OpenJDK 10.0.1,RedHat* Enterprise Linux 7.5,核心 3.10.0-862.el7.x86_64,每秒傳輸分數 52068,由 Intel 於 2018 年 11 月 16 日測量。
工作負載:Wordpress* PHP/HHVM*
結果:AMD EPYC 每元效能 = 基準線 1;Intel® Xeon® 可擴充處理器的每元效能 = 1.75 倍(越高越好)
網站前端 Wordpress(越高越好):
AWS M5.24xlarge (Intel) 實例,oss-performance/wordpress Ver 4.2.0;Ver 10.2.19-MariaDB-1:10.2.19+maria~bionic;工作負載版本':u'4.2.0;用戶端執行緒:200;PHP 7.2.12-1;perfkitbenchmarker_version="v1.12.0-944-g82392cc;Ubuntu 18.04,核心 Linux 4.15.0-1025-aws,TPS 分數 3626.11,由 Intel 於 2018 年 11 月 16 日測量。
AWS M5a.24xlarge (AMD) 實例,oss-performance/wordpress Ver 4.2.0;Ver 10.2.19-MariaDB-1:10.2.19+maria~bionic;工作負載版本':u'4.2.0;用戶端執行緒:200;PHP 7.2.12-1;perfkitbenchmarker_version="v1.12.0-944-g82392cc;Ubuntu 18.04,核心 Linux 4.15.0-1025-aws,TPS 分數 1838.48,由 Intel 於 2018 年 11 月 16 日測量。
AWS M5.4xlarge (Intel) 執行個體,McCalpin Stream(OMP 版本),(來源:https://www.cs.virginia.edu/stream/FTP/Code/stream.c);Intel ICC 18.0.3 20180410 搭配 AVX512,-qopt-zmm-usage=high,-DSTREAM_ARRAY_SIZE=134217728 -DNTIMES=100 -DOFFSET=0 –qopenmp,-qoptstreaming-stores 一律 -o $OUT stream.c,Red Hat* Enterprise Linux 7.5,核心 3.10.0-862.el7.x86_64,OMP_NUM_THREADS:8,KMP_AFFINITY:proclist=[0-7:1],granularity=thread, explicit,分數 81216.7 MB/s,由 Intel 於 2018 年 12 月 6 日測量。
AWS M5a.4xlarge (AMD) 執行個體,McCalpin Stream(OMP 版本),(來源:https://www.cs.virginia.edu/stream/FTP/Code/stream.c);Intel ICC 18.0.3 20180410 搭配 AVX2,-DSTREAM_ARRAY_SIZE=134217728,-DNTIMES=100 -DOFFSET=0 -qopenmp -qopt-streaming-stores 一律 -o $OUT stream.c,Red Hat* Enterprise Linux 7.5,核心 3.10.0-862.el7.x86_64,OMP_NUM_THREADS:8,KMP_AFFINITY:proclist=[0-7:1],granularity=thread,explicit,分數 32154.4 MB/s,由 Intel 於 2018 年 12 月 6 日測量。
OpenFOAM 免責聲明:本產品未經 OpenCFD Limited、透過 www.openfoam.com 的 OpenFOAM 軟體製造商和經銷商,以及 OpenFOAM* 與 OpenCFD* 商標的擁有者核准或背書。
自 2019 年 1 月 12 日起的 AWS 定價,標準 1 年期保留執行個體定價 (https://aws.amazon.com/ec2/pricing/reserved-instances/pricing/),每小時的隨選 Linux/Unix 使用量定價 (https://aws.amazon.com/ec2/pricing/on-demand/)。