https://www.mosaicml.com/blog/coreweave-nvidia-h100-part-1Benchmarking Large Language Models on NVIDIA H100 GPUs with CoreWeave (Part 1)The research and engineering teams here at MosaicML collaborated with CoreWeave, one of the leading cloud providers for NVIDIA GPU-accelerated server platforms, to provide a preview of the performance that can be achieved when training large language models (LLMs) with NVIDIA H100 GPUs on the MosaicML platform.www.mosaicml.com
대충 H100 사용시 A100보다 최소 2.2배 ~ 최대 3.3배 빠른 학습 시간으로 돌릴 수 있다는 뜻
현재 클라우드 GPU 시간당 가격
H100 SXM GPU 시간당 $4.76(/hrGPU)
A100 80GB GPU 가격$2.21(/hrGPU)
오 유찬좌 돌아왔누