惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

有赞技术团队
有赞技术团队
Security Archives - TechRepublic
Security Archives - TechRepublic
大猫的无限游戏
大猫的无限游戏
The GitHub Blog
The GitHub Blog
Google DeepMind News
Google DeepMind News
Vercel News
Vercel News
U
Unit 42
L
LangChain Blog
M
MIT News - Artificial intelligence
S
SegmentFault 最新的问题
月光博客
月光博客
MongoDB | Blog
MongoDB | Blog
Jina AI
Jina AI
博客园 - 聂微东
H
Hackread – Cybersecurity News, Data Breaches, AI and More
cs.AI updates on arXiv.org
cs.AI updates on arXiv.org
Recent Announcements
Recent Announcements
Engineering at Meta
Engineering at Meta
Hacker News - Newest:
Hacker News - Newest: "LLM"
V2EX - 技术
V2EX - 技术
K
KPMG report finds enterprise disconnect between AI and its ROI | CIO
阮一峰的网络日志
阮一峰的网络日志
F
Fortinet All Blogs
D
DataBreaches.Net
Hacker News: Ask HN
Hacker News: Ask HN
W
WeLiveSecurity
N
News | PayPal Newsroom
量子位
Help Net Security
Help Net Security
Exploit-DB.com RSS Feed
Exploit-DB.com RSS Feed
N
News and Events Feed by Topic
Webroot Blog
Webroot Blog
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
人人都是产品经理
人人都是产品经理
S
Security @ Cisco Blogs
Y
Y Combinator Blog
H
Heimdal Security Blog
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
Stack Overflow Blog
Stack Overflow Blog
Attack and Defense Labs
Attack and Defense Labs
S
Secure Thoughts
P
Privacy International News Feed
CTFtime.org: upcoming CTF events
CTFtime.org: upcoming CTF events
SecWiki News
SecWiki News
Last Week in AI
Last Week in AI
AI
AI
Recorded Future
Recorded Future
C
Cyber Attacks, Cyber Crime and Cyber Security
Microsoft Security Blog
Microsoft Security Blog
P
Privacy & Cybersecurity Law Blog

Supermicro Data Center Stories

Rethinking Retail Edge Infrastructure: Why Efficiency and Scalability Matter Supermicro NVIDIA Blackwell Systems Demonstrate Linear Scalability for MLPerf Training v6.0 Right-Sizing Edge AI: Choosing the Right Processor Type for Inferencing Supermicro Announces General Availability of the NVIDIA DGX GB300-Powered Super AI Station at COMPUTEX 2026 Building More Efficient, Reliable AI Infrastructure with NVIDIA's Photonics Switches and NVIDIA Vera Rubin A Closer Look: Building a Modern Data Center with Supermicro Networking and Switching Solutions Supermicro 5U PCIe GPU Servers Using AMD Instinct™ MI350P GPUs Provides Ready-to-Deploy Enterprise AI for Your Existing Infrastructure From Platforms to Production: How Supermicro Is Powering the Rise of AI Factories Supermicro Leads Whisper Benchmark in MLPerf v6.0 with NVIDIA Blackwell Ultra GPUs Powering the Next Wave of AI Infrastructure with Cloud-Native MegaDC Systems Secure AI: Supermicro’s HGX B300 & GB300 NVL72 with NVIDIA Confidential Computing Experience the AI Factory SuperCloud Director: Operationalizing NeoCloud Infrastructure with NVIDIA NCX Infra Controller Built to Accelerate: Supermicro Delivers Powerful AI Factory Clusters and Intelligent Data Platforms for Enterprises Supermicro Announces General Availability of NVIDIA GB300-Powered Super AI Station at GTC 2026
SPEC CPU 2026 Benchmark Suites Released: See Supermicro's Strong Results
Supermicro Experts · 2026-06-12 · via Supermicro Data Center Stories

The Standard Performance Evaluation Corporation (SPEC) officially released the SPEC CPU 2026 benchmark suites, representing the largest update to industry-standard CPU testing in almost a decade. Stepping up from the previous CPU 2017 suite, the 2026 version expands the benchmark count from 43 to 52. It more than doubles the number of lines of code and heavily incorporates modern open-source workloads. Crucially, it broadens testing domains to cover AI orchestration, electronic design automation (EDA), complex databases, and deep graph analytics.

template for blog

The SPEC benchmark suite for CPU results shows both the performance of the CPU for:

  • Integer Speed (INT Speed): Measures the time it takes for a system to execute a single, sequential task that relies heavily on integer math. It consists of 13 benchmark testing tasks, such as compilation, cryptography, and compression.
  • Integer Rate (INT Rate): Measures the total amount of integer workloads a system can process simultaneously. It contains 14 benchmarks designed to test multiprogramming capabilities and the performance of heavily parallelized, multithreaded integer operations.
  • FP Speed (Floating-Point Speed): Measures single-task performance for complex mathematical calculations, such as physics simulations and fluid dynamics. It features 13 benchmarks stressing computing throughput and DRAM/cache performance.
  • FP Rate (Floating-Point Throughput): Measures aggregate throughput for simultaneous, heavy mathematical calculations. It contains 12 benchmarks pushing multi-core and multi-threaded floating-point hardware.

Each of these benchmarks can be run with different numbers of CPUs, but here we will focus only on the 1-socket and 2-socket results.

Integer Speed:
  • 1-socket results – Supermicro holds the top two spots using the CloudDC Server with AMD EPYC™ 9755 (128 cores) and 9965 (192 cores) CPUs, respectively1.
  • 2-socket results – Supermicro holds the top spot and 4 of the top 6 positions. The highest-performing result is a Hyper system with AMD EPYC 9755 CPUs (256 cores)2.
Integer Rate:
  • 1-socket results – Supermicro holds the top 2 results using aSupermicro SuperBlade® using a single Intel® Xeon® 6990E+ (288 cores) CPU. The second-place score also uses the Intel Xeon 6990E+ processor in a Hyper server3.
  • 2-socket results – Supermicro holds the top spot using a Hyper server with the Intel Xeon 6990E+ (total of 576 cores) CPU. The Peak Result is 14+ % faster than the 2nd place result4.

Floating Point Speed:

  • 1-socket results – Supermicro holds two of the top three positions using a CloudDC server with the AMD EPYC 9965 (192 cores) and 9755 (128 cores) CPUs, respectively5.
  • 2-socket results – A Supermicro Hyper server with the AMD EPYC 9755 CPU (total of 256 cores) was within 5% of the fastest system6.

Floating Point Rate:

  • 1-socket results - Supermicro owns the top three positions for this benchmark. The fastest system is a Hyper system using the Intel Xeon 6990E+ (288 cores) CPU, and the 2nd-place result is the same processor on the SuperBlade system. The 3rd place result is also a SuperBlade system with the Intel Xeon 6980E+ (264 cores) CPU7.
  • 2-socket results – Supermicro has the highest performance among the top two server vendors, based on the Q4 CY25 IDC results*. Using a Hyper server with two Intel Xeon 6980P processors, totaling 576 cores, Supermicro attained the top position8.

Visit www.spec.org for more information.

*Source: IDC Worldwide Quarterly Server Tracker, 2025Q4

References:

  1. https://spec.org/cpu2026/results/res2026q2/cpu2026-20260210-00227.html and https://spec.org/cpu2026/results/res2026q2/cpu2026-20260210-00233.html
  2. https://spec.org/cpu2026/results/res2026q2/cpu2026-20260210-00231.html
  3. https://spec.org/cpu2026/results/res2026q2/cpu2026-20260519-00265.html and https://spec.org/cpu2026/results/res2026q2/cpu2026-20260519-00261.html
  4. https://spec.org/cpu2026/results/res2026q2/cpu2026-20260518-00252.html
  5. https://spec.org/cpu2026/results/res2026q2/cpu2026-20260210-00234.html and https://spec.org/cpu2026/results/res2026q2/cpu2026-20260210-00228.html
  6. https://spec.org/cpu2026/results/res2026q2/cpu2026-20260210-00232.html
  7. https://spec.org/cpu2026/results/res2026q2/cpu2026-20260519-00262.html and https://spec.org/cpu2026/results/res2026q2/cpu2026-20260519-00266.html and https://spec.org/cpu2026/results/res2026q2/cpu2026-20260519-00258.html
  8. https://spec.org/cpu2026/results/res2026q2/cpu2026-20260518-00253.html  

Subscribe to Data Center Stories

By clicking subscribe, you consent to allow Supermicro to store and process the personal information submitted above to provide you the content requested.

You can unsubscribe from these communications at any time. For more information on how to unsubscribe, our privacy practices, and how we are committed to protecting and respecting your privacy, please review our Privacy Policy.