<p>Foresight News reports that cloud computing platform CoreWeave (Nasdaq: CRWV) announced a partnership with NVIDIA and IBM to successfully deploy an ultra-large-scale computing cluster consisting of 2,496 NVIDIA Blackwell GPUs, setting the largest submission record in MLPerf benchmark testing history. This cluster, utilizing 2,496 NVIDIA Blackwell GPUs, operates on CoreWeave's AI-optimized cloud platform.</p><p>The scale of this cluster is 34 times larger than those of other cloud service providers. In performance testing, the cluster completed training of the large language model Llama 3.1 405 billion parameters in 27.3 minutes, over twice as fast as similarly scaled clusters.</p>

According to Foresight News, cloud computing platform CoreWeave (Nasdaq: CRWV) announced a collaboration with NVIDIA and IBM to successfully deploy an ultra-large-scale computing cluster consisting of 2,496 NVIDIA Blackwell GPUs. This set a record for the largest submission in the history of the MLPerf benchmark tests, utilizing 2,496 NVIDIA Blackwell GPUs running on CoreWeave's AI-optimized cloud platform. This cluster is 34 times larger than those of other cloud service providers. In performance testing, the cluster completed training of the large language model Llama 3.1 405 billion in 27.3 minutes, more than twice as fast as similarly sized clusters.

CoreWeave Collaborates with NVIDIA and IBM to Build Large-Scale NVIDIA Blackwell GPU Clusters