HPC & Quantum
27 subscribers
11.4K photos
668 videos
3 files
30.6K links
Download Telegram
HPC Guru (Twitter)

Interesting that both @Azure AND @NVIDIAAI both announced fairly similar #MLPerf results based on an identical number (10,752) of #H100 #GPUs

Guess who got more news coverage?

@Microsoft needs to do better marketing, IMO

#AI #HPC #GPU
HPCwire (Twitter)

Gaudi2, 4th Gen Xeon Show Strength in MLPerf Training 3.1, but still Trail Nvidia #MLPerf

hpcwire.com/2023/11/10/gaudi…
HPCwire (Twitter)

Top university #HPC teams clashed at @Supercomputing in Denver for a global challenge. Find out about the contest rules, the benchmarks of #HPL, #HPCG, and #MLPerf Inference, and the results. #SC23 ow.ly/EQys50Qcwkn
HPCwire (Twitter)

The latest #MLPerf Inference benchmark (4.0) results are in, showcasing the evolving landscape of #AIaccelerators. Dive into the data to uncover insights and compare systems for specific use cases. ow.ly/u9Oc50R6IjJ
HPC Guru (Twitter)

#MLPerf Training 4.0 – @nvidia Still King; Power and #LLM Fine Tuning Added

Nvidia (currently king of accelerated computing) wins again, sweeping all nine “events” (workflows) as it were

https://www.hpcwire.com/2024/06/12/mlperf-training-4-0-nvidia-still-king-power-and-llm-fine-tuning-added/

#AI #HPC via @HPCwire
HPCwire (Twitter)

.@NVIDIAAI dominates again in #MLPerf Training 4.0, winning all nine workflows. MLPerf has evolved into a crucial tool for evaluating ML-centric systems, with new benchmarks and power metrics. Check out the full results and insights: ow.ly/50p450Si6Mk
HPC Guru (Twitter)

RT @nvidia: In the latest round of #MLPerf industry benchmarks, the first submission of NVIDIA Blackwell delivered up to 4X more performance on #Llama 2 70B, while the NVIDIA H200 Tensor Core GPU continued to deliver outstanding results across every data center test. nvda.ws/3T5GaVk
HPC Guru (Twitter)

RT @eetimes: First #MLPerf scores for AMD MI300X and Nvidia Blackwell #GPUs, plus startup Untether, show comparable results to market leader Nvidia. https://www.eetimes.com/amd-and-untether-take-on-nvidia-in-mlperf-benchmarks/
HPCwire (Twitter)

MLCommons Releases New MLPerf Inference v5.0 Benchmark Results
ow.ly/cqir50VtsXG #MLCommons #MLPerf #HPC
HPC Guru (Twitter)

RT @nvidia: The NVIDIA Blackwell platform set records in the latest #MLPerf Training v5.0 round and debuted the first training submission using the NVIDIA GB200 NVL72 system, which achieved up to 2.6x more training performance per GPU compared to NVIDIA Hopper.
nvda.ws/3T5Sq7X