Tuesday, August 5, 2025
  • About Web3Wire
  • Web3Wire NFTs
  • .w3w TLD
  • $W3W Token
  • Web3Wire DAO
  • Media Network
  • RSS Feed
  • Contact Us
Web3Wire
No Result
View All Result
  • Home
  • Web3
    • Latest
    • AI
    • Business
    • Blockchain
    • Cryptocurrencies
    • Decentralized Finance
    • Metaverse
    • Non-Fungible Token
    • Press Release
  • Technology
    • Consumer Tech
    • Digital Fashion
    • Editor’s Choice
    • Guides
    • Stories
  • Coins
    • Top 10 Coins
    • Top 50 Coins
    • Top 100 Coins
    • All Coins
  • Exchanges
    • Top 10 Crypto Exchanges
    • Top 50 Crypto Exchanges
    • Top 100 Crypto Exchanges
    • All Crypto Exchanges
  • Stocks
    • Blockchain Stocks
    • NFT Stocks
    • Metaverse Stocks
    • Artificial Intelligence Stocks
  • Events
  • News
    • Latest Crypto News
    • Latest DeFi News
    • Latest Web3 News
  • Home
  • Web3
    • Latest
    • AI
    • Business
    • Blockchain
    • Cryptocurrencies
    • Decentralized Finance
    • Metaverse
    • Non-Fungible Token
    • Press Release
  • Technology
    • Consumer Tech
    • Digital Fashion
    • Editor’s Choice
    • Guides
    • Stories
  • Coins
    • Top 10 Coins
    • Top 50 Coins
    • Top 100 Coins
    • All Coins
  • Exchanges
    • Top 10 Crypto Exchanges
    • Top 50 Crypto Exchanges
    • Top 100 Crypto Exchanges
    • All Crypto Exchanges
  • Stocks
    • Blockchain Stocks
    • NFT Stocks
    • Metaverse Stocks
    • Artificial Intelligence Stocks
  • Events
  • News
    • Latest Crypto News
    • Latest DeFi News
    • Latest Web3 News
No Result
View All Result
Web3Wire
No Result
View All Result
Home Artificial Intelligence

UPDATE – New MLPerf Storage v2.0 Benchmark Results Demonstrate the Critical Role of Storage Performance in AI Training Systems

August 5, 2025
in Artificial Intelligence, GlobeNewswire, Web3
Reading Time: 8 mins read
5
SHARES
243
VIEWS
Share on TwitterShare on LinkedInShare on Facebook

SAN FRANCISCO, Aug. 04, 2025 (GLOBE NEWSWIRE) — Today, MLCommons® announced results for its industry-standard MLPerf® Storage v2.0 benchmark suite, which is designed to measure the performance of storage systems for machine learning (ML) workloads in an architecture-neutral, representative, and reproducible manner. This round of the benchmark saw dramatically increased participation, more geographic representation from submitting organizations, and greater diversity of the systems submitted for testing.

The benchmark results show that storage systems performance continues to improve rapidly, with tested systems serving roughly twice the number of accelerators than in the v1.0 benchmark round.

Additionally, the v2.0 benchmark adds new tests that replicate real-world checkpointing for AI training systems. The benchmark results provide essential information for stakeholders who need to configure the frequency of checkpoints to optimize for high performance – particularly at scale.

Version 2.0 adds checkpointing tasks, delivers essential insights

As AI training systems have continued to scale up to billions and even trillions of parameters, and the largest clusters of processors have reached one hundred thousand accelerators or more, system failures have become a prominent technical challenge. Because data centers tend to run accelerators at near-maximum utilization for their entire lifecycle, both the accelerators themselves and the supporting hardware (power supplies, memory, cooling systems, etc.) are heavily burdened, minimizing their expected lifetime. This is a chronic issue, especially in large clusters: if the mean time to failure for an accelerator is 50,000 hours, then a 100,000-accelerator cluster running for extended periods at full utilization will likely experience a failure every half-hour. A cluster with one million accelerators would expect to see a failure every three minutes. Worse, because AI training usually involves massively parallel computation where all the accelerators are moving in lockstep on the same iteration of training, a failure of one processor can grind an entire cluster to a halt.

It is now broadly accepted that saving checkpoints of intermediate training results at regular intervals is essential to keep AI training systems running at high performance. The AI training community has developed mathematical models that can optimize cluster performance and utilization by trading off the overhead of regular checkpoints against the expected frequency and cost of failure recovery (rolling back the computation, restoring the most recent checkpoint, restarting the training from that point, and duplicating the lost work). Those models, however, require accurate data on the scale and performance of the storage systems that are used to implement the checkpointing system.

The MLPerf Storage v2.0 checkpoint benchmark tests provide precisely that data, and the results from this round suggest that stakeholders procuring AI training systems need to carefully consider the performance of the storage systems they buy, to ensure that they can store and retrieve a cluster’s checkpoints without slowing the system down to an unacceptable level. For a deeper understanding of the issues around storage systems and checkpointing, as well as of the design of the checkpointing benchmarks, we encourage you to read this post from Wes Vaske, a member of the MLPerf Storage working group.

“At the scale of computation being implemented for training large AI models, regular component failures are simply a fact of life,” said Curtis Anderson, MLPerf Storage working group co-chair. “Checkpointing is now a standard practice in these systems to mitigate failures, and we are proud to be providing critical benchmark data on storage systems to allow stakeholders to optimize their training performance. This initial round of checkpoint benchmark results shows us that current storage systems offer a wide range of performance specifications, and not all systems are well-matched to every checkpointing scenario. It also highlights the critical role of software frameworks such as PyTorch and TensorFlow in coordinating training, checkpointing, and failure recovery, as well as some opportunities for enhancing those frameworks to further improve overall system performance.”

Workload benchmarks show rapid innovation in support of larger-scale training systems

Continuing from the v1.0 benchmark suite, the v2.0 suite measures storage performance in a diverse set of ML training scenarios. It emulates the storage demands across several scenarios and system configurations covering a range of accelerators, models, and workloads. By simulating the accelerators’ “think time” the benchmark can generate accurate storage patterns without the need to run the actual training, making it more accessible to all. The benchmark focuses the test on a given storage system’s ability to keep pace, as it requires the simulated accelerators to maintain a required level of utilization.

The v2.0 results show that submitted storage systems have substantially increased the number of accelerators they can simultaneously support, roughly twice the number compared to the systems in the v1.0 benchmark.

“Everything is scaling up: models, parameters, training datasets, clusters, and accelerators. It’s no surprise to see that storage system providers are innovating to support ever larger scale systems,” said Oana Balmau, MLPerf Storage working group co-chair.

The v2.0 submissions also included a much more diverse set of technical approaches to delivering high-performance storage for AI training, including:

  • 6 local storage solutions;
  • 2 solutions using in-storage accelerators;
  • 13 software-defined solutions;
  • 12 block systems;
  • 16 on-prem shared storage solutions;
  • 2 object stores.

“Necessity continues to be the mother of invention: faced with the need to deliver storage solutions that are both high-performance and at unprecedented scale, the technical community has stepped up once again and is innovating at a furious pace,” said Balmau.

MLPerf Storage v2.0: skyrocketing participation and diversity of submitters

The MLPerf Storage benchmark was created through a collaborative engineering process by 35 leading storage solution providers and academic research groups across 3 years. The open-source and peer-reviewed benchmark suite offers a level playing field for competition that drives innovation, performance, and energy efficiency for the entire industry. It also provides critical technical information for customers who are procuring and tuning AI training systems.

The v2.0 benchmark results, from a broad set of technology providers, reflect the industry’s recognition of the importance of high-performance storage solutions. MLPerf Storage v2.0 includes >200 performance results from 26 submitting organizations: Alluxio, Argonne National Lab, DDN, ExponTech, FarmGPU, H3C, Hammerspace, HPE, JNIST/Huawei, Juicedata, Kingston, KIOXIA, Lightbits Labs, MangoBoost, Micron, Nutanix, Oracle, Quanta Computer, Samsung, Sandisk, Simplyblock, TTA, UBIX, IBM, WDC, and YanRong. The submitters represent seven different countries, demonstrating the value of the MLPerf Storage benchmark to the global community of stakeholders.

“The MLPerf Storage benchmark has set new records for an MLPerf benchmark, both for the number of organizations participating and the total number of submissions,” said David Kanter, Head of MLPerf at MLCommons. The AI community clearly sees the importance of our work in publishing accurate, reliable, unbiased performance data on storage systems, and it has stepped up globally to be a part of it. I would especially like to welcome first-time submitters Alluxio, ExponTech, FarmGPU, H3C, Kingston, KIOXIA, Oracle, Quanta Cloud Technology, Samsung, Sandisk, TTA, UBIX, IBM, and WDC.”

“This level of participation is a game-changer for benchmarking: it enables us to openly publish more accurate and more representative data on real-world systems,” Kanter continued. “That, in turn, gives the stakeholders on the front lines the information and tools they need to succeed at their jobs. The checkpoint benchmark results are an excellent case in point: now that we can measure checkpoint performance, we can think about optimizing it.”

We invite stakeholders to join the MLPerf Storage working group and help us continue to evolve the benchmark suite.

View the Results

To view the results for MLPerf Storage v2.0, please visit the Storage benchmark results.

About MLCommons

MLCommons is the world’s leader in AI benchmarking. An open engineering consortium supported by over 125 members and affiliates, MLCommons has a proven record of bringing together academia, industry, and civil society to measure and improve AI. The foundation for MLCommons began with the MLPerf benchmarks in 2018, which rapidly scaled as a set of industry metrics to measure machine learning performance and promote transparency of machine learning techniques. Since then, MLCommons has continued using collective engineering to build the benchmarks and metrics required for better AI – ultimately helping to evaluate and improve AI technologies’ accuracy, safety, speed, and efficiency.

For additional information on MLCommons and details on becoming a member, please visit MLCommons.org or email participation@mlcommons.org.

About Web3Wire
Web3Wire – Information, news, press releases, events and research articles about Web3, Metaverse, Blockchain, Artificial Intelligence, Cryptocurrencies, Decentralized Finance, NFTs and Gaming.
Visit Web3Wire for Web3 News and Events, Block3Wire for the latest Blockchain news and Meta3Wire to stay updated with Metaverse News.
ShareTweet1ShareSendShare2
Previous Post

CPAClub Named CalCPA Firm of the Year and Launches CPAClub Ally™

Next Post

Business Software and Services Market Insights 2025-2032: Global Expansion, Revenue Trends, and Strategic Growth Plans Featuring Acumatica, Inc. and Unit4

Related Posts

NexulonAI ($NEX): Transforming Crypto Payments and DeFi with AI Solutions

As blockchain technology and cryptocurrency continue to evolve, NexulonAI is emerging as a leader in DeFi (Decentralized Finance) and crypto payments. By combining artificial intelligence (AI) with blockchain technology, NexulonAI is providing users with innovative solutions for engaging with digital assets and DeFi protocols. The platform’s native token, $NEX, supports...

Read moreDetails

CXApp Inc. (Nasdaq: CXAI) Schedules Second Quarter 2025 Financial Results and Business Update Conference Call

CXAI is shaping the future of work by enabling enhanced employee experiences using artificial intelligence PALO ALTO, CALIFORNIA / ACCESS Newswire / August 4, 2025 / CXApp Inc. (Nasdaq:CXAI), the global technology leader in employee workplace experiences, announced it will host a conference call at 5:00 PM Eastern Time on...

Read moreDetails

Z.ai Launches GLM-4.5 with Bold Open-source Push: China’s Top LLM Contender Challenges Big Tech

Z.ai (formerly Zhipu) has released GLM-4.5, marking the company's entry into open-source Mixture-of-Experts (MoE) modeling with a series designed specifically for autonomous AI agents. The launch positions these models as competitive alternatives to proprietary offerings from OpenAI, Anthropic, and Google.Two-Model Release StrategyThe GLM-4.5 family includes two variants optimized for different...

Read moreDetails

Brokeret Named 2025’s Most Customizable Forex CRM Platform by Independent Reviewers

Image: https://www.abnewswire.com/upload/2025/08/b89305b456aaa496358bd108deaf5713.jpgAugust 4, 2025 - Dubai, UAE - In the fast-moving world of financial technology, Brokeret has earned the distinction of being recognized as 2025's most customizable Forex CRM platform , according to multiple independent reviewers and industry-focused analyst groups. The recognition comes on the heels of a broad comparative...

Read moreDetails

Adtelligent Launches Retail Media Platform for Full-Funnel, Cross-Channel Retail Advertising

LEWES, Del., Aug. 04, 2025 (GLOBE NEWSWIRE) -- Adtelligent®, a global advertising technology company, proudly announces the official launch of its Retail Media Platform — a powerful solution designed to help retailers and marketplaces open new monetization opportunities by turning their digital properties into scalable media platforms.  Our platform gives retailers...

Read moreDetails

Broadcom Ships Jericho4, Enabling Distributed AI Computing Across Data Centers

PALO ALTO, Calif., Aug. 04, 2025 (GLOBE NEWSWIRE) -- Broadcom Inc. (NASDAQ: AVGO), a global leader in semiconductor and infrastructure software solutions, today announced it is now shipping the Jericho4 ethernet fabric router — a purpose-built platform for the next generation of distributed AI infrastructure. Designed to interconnect over one...

Read moreDetails

Bit Origin Announces Completion of Share Capital Reduction

SINGAPORE, Aug. 04, 2025 (GLOBE NEWSWIRE) -- Bit Origin Ltd (NASDAQ: BTOG) ("Bit Origin" or the "Company") today announced that it has completed a reduction of its issued share capital, following the approval by its shareholders and subsequent confirmation by the Grand Court of the Cayman Islands. On June 5,...

Read moreDetails

Stride, Inc. Signs Multi-District Agreements to Operate K12-Powered Online School in New Mexico

RESTON, VA, Aug. 04, 2025 (GLOBE NEWSWIRE) -- Stride, Inc. (NYSE: LRN) has now partnered with Chama Valley Independent Schools and Santa Rosa Consolidated Schools, to operate Destinations Career Academy of New Mexico (NMDCA), one of its K12-powered online public schools, starting this fall. With more than 3,000 students already...

Read moreDetails

Vodafone and Kaltura Sign a Long-Term Partnership Agreement to Expand Cloud TV Services

New York, Aug. 04, 2025 (GLOBE NEWSWIRE) -- Vodafone Group PLC (Nasdaq: VOD) and Kaltura (Nasdaq: KLTR), the AI Video Experience Cloud, today announced that they have extended their partnership around the Kaltura-powered Vodafone TV solution (VTV).  Vodafone launched VTV in 2014, and it has since grown to serve several...

Read moreDetails

EverQuote Announces Second Quarter 2025 Financial Results

Second Quarter Revenue Growth of 34% Year-Over-Year to $156.6 million Second Quarter Variable Marketing Dollars Increase 25% Year-Over-Year to $45.5 million Delivers Second Quarter Record Net Income of $14.7 million and Adjusted EBITDA of $22.0 million Announces $50.0 million Inaugural Share Repurchase Program CAMBRIDGE, Mass., Aug. 04, 2025 (GLOBE NEWSWIRE)...

Read moreDetails
Web3Wire NFTs - The Web3 Collective

Web3Wire, $W3W Token and .w3w tld Whitepaper

Web3Wire, $W3W Token and .w3w tld Whitepaper

Claim your space in Web3 with .w3w Domain!

Web3Wire

Trending on Web3Wire

  • Unifying Blockchain Ecosystems: 2024 Guide to Cross-Chain Interoperability

    87 shares
    Share 35 Tweet 22
  • Top Cross-Chain DeFi Solutions to Watch by 2025

    48 shares
    Share 19 Tweet 12
  • Discover 2025’s Top 5 Promising Low-Cap Crypto Gems

    66 shares
    Share 26 Tweet 17
  • Top 5 Wallets for Seamless Multi-Chain Trading in 2025

    43 shares
    Share 17 Tweet 11
  • Discover the Best Metaverse Crypto Projects and Virtual Worlds 2025

    44 shares
    Share 18 Tweet 11
Join our Web3Wire Community!

Our newsletters are only twice a month, reaching around 10000+ Blockchain Companies, 800 Web3 VCs, 600 Blockchain Journalists and Media Houses.


* We wont pass your details on to anyone else and we hate spam as much as you do. By clicking the signup button you agree to our Terms of Use and Privacy Policy.

Web3Wire Podcasts

Upcoming Events

Web 3.0 and AI Summit 2025

2025-09-11
Frankfurt
Summit

Latest on Web3Wire

  • NexulonAI ($NEX): Transforming Crypto Payments and DeFi with AI Solutions
  • CXApp Inc. (Nasdaq: CXAI) Schedules Second Quarter 2025 Financial Results and Business Update Conference Call
  • Z.ai Launches GLM-4.5 with Bold Open-source Push: China’s Top LLM Contender Challenges Big Tech
  • Brokeret Named 2025’s Most Customizable Forex CRM Platform by Independent Reviewers
  • Adtelligent Launches Retail Media Platform for Full-Funnel, Cross-Channel Retail Advertising

RSS Latest on Block3Wire

  • Covo Finance: Revolutionary Crypto Leverage Trading Platform
  • WorldStrides and HEX Announce Partnership to Offer High School and University Students Innovative Courses Designed to Improve Their Outlook in the Digital Age
  • Cathedra Bitcoin Announces Leasing of 2.5-MW Bitcoin Mining Facility
  • Global Web3 Payments Leader, Banxa, Announces Integration With Metis to Usher In Next Wave of Cryptocurrency Users
  • Dexalot Launches First Hybrid DeFi Subnet on Avalanche

RSS Latest on Meta3Wire

  • Thumbtack Honored as a 2023 Transform Awards Winner
  • Accenture Invests in Looking Glass to Accelerate Shift from 2D to 3D
  • MetatronAI.com Unveils Revolutionary AI-Chat Features and Interface Upgrades
  • Purely.website – Disruptive new platform combats rising web hosting costs
  • WEMADE and Metagravity Sign Strategic Alliance MOU to Collaborate on Blockchain Games for the Metaverse
Web3Wire

Web3Wire is your go-to source for the latest insights and updates in Web3, Metaverse, Blockchain, AI, Cryptocurrencies, DeFi, NFTs, and Gaming. We provide comprehensive coverage through news, press releases, event updates, and research articles, keeping you informed about the rapidly evolving digital world.

  • About Web3Wire
  • Web3Wire NFTs – The Web3 Collective
  • .w3w TLD
  • $W3W Token
  • Web3Wire DAO
  • Event Partners
  • Community Partners
  • Our Media Network
  • Media Kit
  • RSS Feeds
  • Contact Us

Whitepaper | Tokenomics

Crypto Coins

  • Top 10 Coins
  • Top 50 Coins
  • Top 100 Coins
  • All Coins – Marketcap
  • Crypto Coins Heatmap

Crypto Exchanges

  • Top 10 Exchanges
  • Top 50 Exchanges
  • Top 100 Exchanges
  • All Crypto Exchanges

Crypto Stocks

  • Blockchain Stocks
  • NFT Stocks
  • Metaverse Stocks
  • Artificial Intelligence Stocks

Media Portfolio: Block3Wire | Meta3Wire

Web3 Resources

  • Top Web3 and Crypto Youtube Channels
  • Latest Crypto News
  • Latest DeFi News
  • Latest Web3 News

Blockchain Resources

  • Blockchain and Web3 Resources
  • Decentralized Finance (DeFi) – Research Reports
  • All Crypto Whitepapers

Metaverse Resources

  • AR VR and Metaverse Resources
  • Metaverse Courses
Claim your space in Web3 with .w3w!
Top 50 Web3 Blogs and Websites
Web3Wire Podcast on Spotify Web3Wire Podcast on Amazon Music 
Web3Wire - Web3 and Blockchain - News, Events and Press Releases | Product Hunt
Web3Wire on Google News
  • Privacy Policy
  • Terms of Use
  • Disclaimer
  • Sitemap
  • For Search Engines
  • Crypto Sitemap
  • Exchanges Sitemap

© 2024 Web3Wire. We strongly recommend our readers to DYOR, before investing in any cryptocurrencies, blockchain projects, or ICOs, particularly those that guarantee profits.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In

Add New Playlist

No Result
View All Result
  • Coins
    • Top 10 Cryptocurrencies
    • Top 50 Cryptocurrencies
    • Top 100 Cryptocurrencies
    • All Coins
  • Exchanges
    • Top 10 Cryptocurrency Exchanges
    • Top 50 Cryptocurrency Exchanges
    • Top 100 Cryptocurrency Exchanges
    • All Crypto Exchanges
  • Stocks
    • Blockchain Stocks
    • NFT Stocks
    • Metaverse Stocks
    • Artificial Intelligence Stocks

© 2024 Web3Wire. We strongly recommend our readers to DYOR, before investing in any cryptocurrencies, blockchain projects, or ICOs, particularly those that guarantee profits.

This website uses cookies. By continuing to use this website you are giving consent to cookies being used. Visit our Privacy and Cookie Policy.