←back to thread

Ask HN: Why hasn’t AMD made a viable CUDA alternative?

183 points spacebanana7 | 8 comments | 01 Apr 25 14:37 UTC | HN request time: 0.217s | source | bottom

I appreciate developing ROCm into something competitive with CUDA would require a lot of work, both internally within AMD and with external contributions to the relevant open source libraries.

However the amount of resources at stake is incredible. The delta between NVIDIA's value and AMD's is bigger than the annual GDP of Spain. Even if they needed to hire a few thousand engineers at a few million in comp each, it'd still be a good investment.

1. echelon ◴[01 Apr 25 14:41 UTC] No.43547340[source]▶
Leadership. At the end of the day, the buck stops with leadership.

If they wanted to prioritize this, they would. They're simply not taking it seriously.

replies(1): >>43547363 #
2. spacebanana7 ◴[01 Apr 25 14:43 UTC] No.43547363[source]▶
I always thought of Lisa Su as an effective leader, but this does make me question it.
replies(3): >>43547469 #>>43547604 #>>43552617 #
3. setgree ◴[01 Apr 25 14:50 UTC] No.43547469[source]▶
I thought so too until I read a Stratechery interview that painted a pretty bad portrait of her time at Sony [0].

I tend to think that AMD looks well run when compared with Intel [1], but when you consider Nvidia as the relevant counterfactual [2], things don't look so good.

I wrote about this here [3].

[0] https://news.ycombinator.com/item?id=40697341

[1] https://news.ycombinator.com/item?id=41446766

[2] https://news.ycombinator.com/item?id=39344815

[3] https://setharielgreen.com/blog/amd-also-seems-to-be-flounde...

4. adra ◴[01 Apr 25 15:00 UTC] No.43547604[source]▶
Why, because 90% of her job is talking to and appeasing shareholders, grand standing with fat whales, and what else.. what do you think a CEO at these companies actually does? They aren't in the trenches of each subdivision nurturing and cracking whips. She likely attends a 2 hour briefing with a line item: CUDA parity project: on schedule release date not set
replies(1): >>43549531 #
5. fulladder ◴[01 Apr 25 17:39 UTC] No.43549531{3}[source]▶
Fair points. I think the difference is that AMD has other businesses that are much larger than GPUs, whereas NVIDIA was always just about GPUs.
6. grg0 ◴[02 Apr 25 00:21 UTC] No.43552617[source]▶
And why do you think that? Have you worked with her personally, or is it maybe because Fortune spent a decade preaching "women in tech" regardless of what those women may actually be doing?
replies(1): >>43552659 #
7. acdha ◴[02 Apr 25 00:31 UTC] No.43552659{3}[source]▶
Do you think AMD out-performing Intel for years might have something to do with it? I know that doesn’t fit whatever axe you have to grind against “women in tech” but it seems more in keeping with the business world’s widely recognized tendency to credit CEOs for the decisions made by large numbers of people who work for them but don’t get up on stage.
replies(1): >>43552911 #
8. grg0 ◴[02 Apr 25 01:25 UTC] No.43552911{4}[source]▶
Intel mostly just shot itself in the foot; easy to "outperform" a handicapped opponent, and there is no merit in doing so. x86 is also a piece of shit in the grander scheme of things, and will be eaten away by alternative hardware soon. So no, I don't think that has anything to do with it, which is why I was asking a legitimate question. Do people just regurgitate the crap they read online, or do they actually put thought into the beliefs that they form?

Agree with your last statement, though. For-hire CEOs--as opposed to founder-CEOs, who often have an actual vision--take way more credit for their "work" than they deserve.

Also rofl at that "axe to grind". That one took me offline for a few seconds.