News for hackers. Stuff that matters.

A comment on AMD's MI300X Outperforms Nvidia's H100 for LLM Inference. Parent comment

Re:AMD's MI300X Outperforms Nvidia's H100 for LLM Inference by triblemaster on Thursday June 13, 2024 @05:56AM (#40667729)
Blackwell won't be here till next year.
[ Parent | on HN ]
Re:AMD's MI300X Outperforms Nvidia's H100 for LLM Inference by DarkmSparks on Thursday June 13, 2024 @06:33AM (#40667938)
GB200 based on blackwell launched in March of this year.

https://www.theregister.com/2024/03/21/nvidia_dgx_gb200_nvk7...

MI300X launched 3 months earlier at the end of December.

H100 launched March 2023,

[ Parent | on HN ]
Re:AMD's MI300X Outperforms Nvidia's H100 for LLM Inference by muxr on Sunday June 16, 2024 @06:51PM (#40700767)
You can paper launch as much as you want but Blackwell isn't shipping until Q4.

Also Blackwell's lead will be short lived, because mi350x is coming out next year, and it will have a node and architecture advantage. So AMD will be ahead again.

[ Parent | on HN ]
Re:AMD's MI300X Outperforms Nvidia's H100 for LLM Inference by acchow on Thursday June 13, 2024 @06:02AM (#40667764)
Nvidia expects to ship 420k Blackwell chips this year.
[ Parent | on HN ]