Google Cloud has celebrated yet another new launch as the CSP reveals new offerings powered by NVIDIA’s hardware.
Announced over at the Google Cloud Next conference, the new A3 instances come with higher processing capability thanks to Team Green’s H100 Tensor Core GPUs and the associated Transformer Engine dedicated to accelerating Large Language Models (LLM).
Not only the infrastructure or hardware makes the headline but also the extensive software enablement NVIDIA has been known for.
With the addition of NVIDIA H100 now greenlighted for A3, Google Cloud now offers A100, H100, and L4 GPU for all sorts of AI-enabled workloads alongside complex computational tasks.
The original idea of accelerating PaxML via Google TPU accelerator slices now offers an alternative method of utilizing Team Green’s A100 and H100 GPUs for full-scale experimentation and scaling with immediate availability in NVIDIA NGC.
In terms of upcoming services, the NVIDIA DGX GH200 supercomputer and DGX Cloud platform are just some of the most anticipated for customers to unleash the power of generative AI yet.
