←back to thread

The AMD Radeon Instinct MI300A's Giant Memory Subsystem

(chipsandcheese.com)
212 points pella | 2 comments | 18 Jan 25 12:28 UTC | HN request time: 0.449s | source
Show context
behnamoh ◴[18 Jan 25 18:21 UTC] No.42750215[source]▶
AMD is done, no one uses their GPUs for AI because AMD were too dumb to understand the value of software lock-in like Nvidia did with CUDA.
replies(3): >>42750458 #>>42753721 #>>42756635 #
1. guywhocodes ◴[18 Jan 25 18:54 UTC] No.42750458[source]▶
More like the value of drivers that doesn't require one in-house team per customer to "fix" driver crashes in the customers' particular workloads.
replies(1): >>42751175 #
2. numpy-thagoras ◴[18 Jan 25 20:29 UTC] No.42751175[source]▶
Yeah, the labour involved in running non Nvidia equipment is the elephant in the room.

Nvidia GPU: spin up OS, run your sims or load your LLM, gather results.

AMD GPU: spin up OS, grok driver fixes, try and run your sims, grok more driver fixes, can't even gather results until you can verify software correctness of your fixes. Yeah, sometimes you need someone with specialized knowledge of numerical methods to help tune your fixes.

... What kind of maddening workflows are these? It's literally negative work: you are busy, you barely get anywhere, and you end up having to do more.

In light of that, the Nvidia tax doesn't look so bad.