Linking pages
- Faulty Nvidia H100 GPUs and HBM3 memory caused half of failures during LLama 3 training — one failure every three hours for Meta's 16,384 GPU training cluster | Tom's Hardware https://www.tomshardware.com/tech-industry/artificial-intelligence/faulty-nvidia-h100-gpus-and-hbm3-memory-caused-half-of-the-failures-during-llama-3-training-one-failure-every-three-hours-for-metas-16384-gpu-training-cluster 193 comments
- Elon Musk spent roughly $10 billion on AI training hardware in 2024 | Tom's Hardware https://www.tomshardware.com/tech-industry/artificial-intelligence/elon-musk-spent-roughly-usd10-billion-on-ai-training-hardware-in-2024 5 comments
- AMD President Victor Peng retires — SVP of AI group Vamsi Boppana will lead AMD's AI business | Tom's Hardware https://www.tomshardware.com/tech-industry/amd-president-victor-peng-retires 1 comment
- Meta is using more than 100,000 Nvidia H100 AI GPUs to train Llama-4 — Mark Zuckerberg says that Llama 4 is being trained on a cluster “bigger than anything that I’ve seen” | Tom's Hardware https://www.tomshardware.com/tech-industry/artificial-intelligence/meta-is-using-more-than-100-000-nvidia-h100-ai-gpus-to-train-llama-4-mark-zuckerberg-says-that-llama-4-is-being-trained-on-a-cluster-bigger-than-anything-that-ive-seen 1 comment
- Nvidia preparing a China-focused variant of its B200 Blackwell AI GPU to comply with US export regulations | Tom's Hardware https://www.tomshardware.com/pc-components/gpus/nvidia-preparing-a-china-focused-variant-of-its-b200-blackwell-ai-gpu-to-comply-with-us-export-regulations 0 comments
- Elon Musk reveals photos of Dojo D1 Supercomputer cluster — roughly equivalent to 8,000 Nvidia H100 GPUs for AI training | Tom's Hardware https://www.tomshardware.com/tech-industry/artificial-intelligence/elon-musk-reveals-photos-of-dojo-d1-supercomputer-cluster-roughly-equivalent-to-8000-nvidia-h100-gpus-for-ai-training 0 comments
- Ex-Twitter dev reminisces about finding 700 unused Nvidia GPUs after takeover — forgotten cluster was 'powered on and idle' | Tom's Hardware https://www.tomshardware.com/pc-components/gpus/ex-twitter-dev-reminisces-about-finding-700-unused-nvidia-gpus-after-takeover-forgotten-cluster-was-powered-on-and-idle 0 comments
Linked pages
- Leaked RDNA 4 features suggest AMD drive to catch up in Ray Tracing — doubled RT intersect engine could come to PS5 Pro | Tom's Hardware https://www.tomshardware.com/pc-components/gpus/leaked-rdna-4-features-suggest-amd-drive-to-catch-up-in-ray-tracing-doubled-rt-intersect-engine-could-come-to-ps5-pro 253 comments
- Intel cancels fab investment in Italy and R&D facility in France — chipmaker remains committed to other European expansions | Tom's Hardware https://www.tomshardware.com/tech-industry/intel-cancels-fab-investment-in-italy-and-randd-facility-in-france-chipmaker-remains-committed-to-other-european-expansions 194 comments
- Nvidia GeForce RTX 50-series launch pushed back to early 2025 according to prominent leaker | Tom's Hardware https://www.tomshardware.com/pc-components/gpus/nvidia-geforce-rtx50-launch-pushed-back-to-early-2025-according-to-prominent-leaker 122 comments
- Intel finally announces a solution for CPU crashing errors — claims elevated voltages are the root cause; patch coming by mid-August [Updated] | Tom's Hardware https://www.tomshardware.com/pc-components/cpus/intel-finally-announces-a-solution-for-cpu-crashing-errors-claims-elevated-voltages-are-the-root-cause-fix-coming-by-mid-august 41 comments
- Microsoft's EU agreement means it will be hard to avoid CrowdStrike-like calamities in the future | Tom's Hardware https://www.tomshardware.com/software/windows/microsofts-eu-agreement-means-it-will-be-hard-to-avoid-crowdstrike-like-calamities-in-the-future 9 comments
- Elon Musk's liquid-cooled 'Gigafactory' AI data centers get a plug from Supermicro CEO — Tesla and xAI's new supercomputers will have 350,000 Nvidia GPUs, both will be online within months | Tom's Hardware https://www.tomshardware.com/tech-industry/artificial-intelligence/elon-musks-liquid-cooled-gigafactory-data-centers-get-a-plug-from-supermicro-ceo-tesla-and-xais-new-supercomputers-will-have-350000-nvidia-gpus-both-will-be-online-within-months 5 comments
- Nvidia preparing a China-focused variant of its B200 Blackwell AI GPU to comply with US export regulations | Tom's Hardware https://www.tomshardware.com/pc-components/gpus/nvidia-preparing-a-china-focused-variant-of-its-b200-blackwell-ai-gpu-to-comply-with-us-export-regulations 0 comments