6 ms·

I found his video pretty dumb. Everybody is assuming the driver is shitty, but he tried is an officially unsupported[1] GPU with ROCm and it failed. Big whoop.

I found his video pretty dumb. Everybody is assuming the driver is shitty, but he tried is an officially unsupported[1] GPU with ROCm and it failed. Big whoop. He is a serial shit stirrer and this is no exception. Did he try Windows? As much as I hate that AMD is not really supporting consumer GPUs for compute, presenting his problems as some variant of production drivers breaking is a stretch at best. [1]: https://rocm.docs.amd.com/en/latest/release/gpu_os_support.html#linux-supported-gpus https://rocm.docs.amd.com/en/latest/release/gpu_os_support.h...

jjoonathan 3y ago
Leaving consumer GPUs unsupported is part of the problem and AMD deserves to have its shit stirred for it. They need to be better than nvidia, not "you'll take what you can get." We can get nvidia. That's the bar. NVidia charges a premium but their shit works (comparatively speaking). AMD has been half-assing their compute offerings for 15 years, it finally became important, and now they need to play catch-up not drag their heels and toss out excuses as to why that's OK. My prediction: AMD won't be second place in this race for long, someone who actually wants the #2 spot will take it from them and AMD will sit around wondering how they blew a 15 year head start.
- Jasper_ 3y ago
  Remember, geohot was going to write the driver -- he came out of the gate saying "AMD has ROCm but you can't use it on consumer cards, so I'm going to write my own software stack that works with AMD's consumer grade GPUs". He raised $5.1m on that promise. So far, with all that money, he has compiled their driver, ran it in an unsupported configuration, and then had a complete public mental breakdown because it didn't work. Something he already knew going in. Should AMD support ROCm on its consumer grade GPUs? Probably. But that's really not geohot's choice to make, unless he wants to get his hands dirty and actually write the software he promised to write. Having worked on GPU drivers, it's not a couple-lines fix situation, it would be a pretty big investment to add stable ROCm support for consumer GPUs. AMD higher-ups responding with anything other than "lol what it's unsupported what did you expect" is extending a pretty long olive branch here.
  - AshamedCaptain 3y ago
    > "AMD has ROCm but you can't use it on consumer cards, so I'm going to write my own software stack that works with AMD's consumer grade GPUs". He raised $5.1m on that promise. Having not followed this, I don't understand what the promise was. ROCm works just fine on at least some consumer cards like the 6xxx, and by fine I mean "as bad as on AMD's pro cards", but at least it works out of the box. Certainly it's not supported, and so therefore they are not shipping precompiled binaries, but it does seem to work...
  - hedgehog 3y ago
    To give him credit I don't think he was saying he was going to write new drivers, just build a ML software stack that worked on whatever they ship for the consumer cards. Certainly a large project by itself with questionable value at this point, but not writing new drivers.
  - jjoonathan 3y ago
    The problem is mismatch between spec sheet and customer needs, not mismatch between spec sheet and card. I don't know whether this is a management problem at AMD or an engineering problem at AMD or both. I don't really care, either -- a "wans't me!" from engineering is completely uninteresting to me. The problem is that AMD's consumer cards don't run ROCm while NVidia's consumer cards all run CUDA. I am amazed that Geohot reached out to AMD to extend an olive branch so far as to force their failure of a product to work despite itself, because frankly I'd just have expected AMD to spin excuses like you did. I am encouraged to see that they took a higher road; hopefully that translates to actual execution. We'll see.
    - iforgotpassword 3y ago
      We tried ROCm on the MI50 years ago. Complete shit show, crashes, lackluster support by AMD. I guess it was because we would've just bought a few dozen, so maybe their competent staff was doing support for some startup with huge VC backing buying a few hundred dozen for a new game streaming service.
- latchkey 3y ago
  The AI race is so hot right now that AMD will sit in #2 and pick up all of the people who can't get access to NVIDIA cards. There is a huge centralization issue around only writing your code for NVIDIA... and that is a huge business risk. People are going to wake up to that fast as the supplies of NVIDIA go to zero. That said, I don't think people realize that there is literally no more large scale tier 3 (redundant) data center power in the US. Even if you are sitting on 1000 or 10k NVIDIA cards, you can't deploy them anywhere. They also need to be in the same data center for speed. You can't just colo in 2-3 data centers to get what you want. If you want to train a large model across 1000 gpus? You're screwed. Think you can just go to the cloud? Go try to sign up for coreweave. They are full and not taking any more customers. A lot of the other sites advertising nvidia gpus are just reselling coreweave under the covers. Forget the software problems. There are far bigger issues and they are not getting better any time soon.
  - imtringued 3y ago
    AMD's AI cookie cutter business model is getting million dollar contracts with a handful of companies. The peasants don't get access to or use AMDs data center cards.
    - latchkey 3y ago
      And why do you think that is?
  - Aeolos 3y ago
    The problem is AMD is not #2 in this field, not by a long shot. Google and Amazon offer their own hardware that is both better supported and actually available for customers. Apple has fast inference hardware in every computer and mobile device they offer. I raised this issue with AMD 8 years ago in a technology conference, and the answer I got back then was a shoulder shrug "we don't think this is an important market." 8 years later and they have all but lost the war.
    - latchkey 3y ago
      G/A can't be used for large scale training, nobody is going to give their data to them. Major trust issues there. Apple is Apple. Not public. Let's also not mix consumer needs with enterprise. You're correct, 8 years ago and up until recently, AMD only cared about gamers. They are waking up fast though. ROCm 5.6 is a visible first step in that regard. MI300 will blow A/H100's out of the water. But again, hardware/software isn't the problem here. The problem is much deeper than that... even if you have those things resolved, you can't put them anywhere.
mindcrime 3y ago
As much as I hate that AMD is not really supporting consumer GPUs for compute, Keep in mind this post is around two months old now, and since it was published AMD have already officially announced plans to support (at least some) consumer GPU's in ROCm.
imtringued 3y ago
Your comment reminds me of another comment by (SilverBirch I believe) that AMD can do this and will get away with it and that George Hotz is unimportant to AMD. Then a few weeks later Lisa Su tweeted this https://twitter.com/LisaSu/status/1669848494637735936 https://twitter.com/LisaSu/status/1669848494637735936.

jjoonathan 3y ago

Leaving consumer GPUs unsupported is part of the problem and AMD deserves to have its shit stirred for it. They need to be better than nvidia, not "you'll take what you can get." We can get nvidia. That's the bar. NVidia charges a premium but their shit works (comparatively speaking). AMD has been half-assing their compute offerings for 15 years, it finally became important, and now they need to play catch-up not drag their heels and toss out excuses as to why that's OK. My prediction: AMD won't be second place in this race for long, someone who actually wants the #2 spot will take it from them and AMD will sit around wondering how they blew a 15 year head start.

Jasper_ 3y ago
Remember, geohot was going to write the driver -- he came out of the gate saying "AMD has ROCm but you can't use it on consumer cards, so I'm going to write my own software stack that works with AMD's consumer grade GPUs". He raised $5.1m on that promise. So far, with all that money, he has compiled their driver, ran it in an unsupported configuration, and then had a complete public mental breakdown because it didn't work. Something he already knew going in. Should AMD support ROCm on its consumer grade GPUs? Probably. But that's really not geohot's choice to make, unless he wants to get his hands dirty and actually write the software he promised to write. Having worked on GPU drivers, it's not a couple-lines fix situation, it would be a pretty big investment to add stable ROCm support for consumer GPUs. AMD higher-ups responding with anything other than "lol what it's unsupported what did you expect" is extending a pretty long olive branch here.
- AshamedCaptain 3y ago
  > "AMD has ROCm but you can't use it on consumer cards, so I'm going to write my own software stack that works with AMD's consumer grade GPUs". He raised $5.1m on that promise. Having not followed this, I don't understand what the promise was. ROCm works just fine on at least some consumer cards like the 6xxx, and by fine I mean "as bad as on AMD's pro cards", but at least it works out of the box. Certainly it's not supported, and so therefore they are not shipping precompiled binaries, but it does seem to work...
- hedgehog 3y ago
  To give him credit I don't think he was saying he was going to write new drivers, just build a ML software stack that worked on whatever they ship for the consumer cards. Certainly a large project by itself with questionable value at this point, but not writing new drivers.
- jjoonathan 3y ago
  The problem is mismatch between spec sheet and customer needs, not mismatch between spec sheet and card. I don't know whether this is a management problem at AMD or an engineering problem at AMD or both. I don't really care, either -- a "wans't me!" from engineering is completely uninteresting to me. The problem is that AMD's consumer cards don't run ROCm while NVidia's consumer cards all run CUDA. I am amazed that Geohot reached out to AMD to extend an olive branch so far as to force their failure of a product to work despite itself, because frankly I'd just have expected AMD to spin excuses like you did. I am encouraged to see that they took a higher road; hopefully that translates to actual execution. We'll see.
  - iforgotpassword 3y ago
    We tried ROCm on the MI50 years ago. Complete shit show, crashes, lackluster support by AMD. I guess it was because we would've just bought a few dozen, so maybe their competent staff was doing support for some startup with huge VC backing buying a few hundred dozen for a new game streaming service.
latchkey 3y ago
The AI race is so hot right now that AMD will sit in #2 and pick up all of the people who can't get access to NVIDIA cards. There is a huge centralization issue around only writing your code for NVIDIA... and that is a huge business risk. People are going to wake up to that fast as the supplies of NVIDIA go to zero. That said, I don't think people realize that there is literally no more large scale tier 3 (redundant) data center power in the US. Even if you are sitting on 1000 or 10k NVIDIA cards, you can't deploy them anywhere. They also need to be in the same data center for speed. You can't just colo in 2-3 data centers to get what you want. If you want to train a large model across 1000 gpus? You're screwed. Think you can just go to the cloud? Go try to sign up for coreweave. They are full and not taking any more customers. A lot of the other sites advertising nvidia gpus are just reselling coreweave under the covers. Forget the software problems. There are far bigger issues and they are not getting better any time soon.
- imtringued 3y ago
  AMD's AI cookie cutter business model is getting million dollar contracts with a handful of companies. The peasants don't get access to or use AMDs data center cards.
  - latchkey 3y ago
    And why do you think that is?
- Aeolos 3y ago
  The problem is AMD is not #2 in this field, not by a long shot. Google and Amazon offer their own hardware that is both better supported and actually available for customers. Apple has fast inference hardware in every computer and mobile device they offer. I raised this issue with AMD 8 years ago in a technology conference, and the answer I got back then was a shoulder shrug "we don't think this is an important market." 8 years later and they have all but lost the war.
  - latchkey 3y ago
    G/A can't be used for large scale training, nobody is going to give their data to them. Major trust issues there. Apple is Apple. Not public. Let's also not mix consumer needs with enterprise. You're correct, 8 years ago and up until recently, AMD only cared about gamers. They are waking up fast though. ROCm 5.6 is a visible first step in that regard. MI300 will blow A/H100's out of the water. But again, hardware/software isn't the problem here. The problem is much deeper than that... even if you have those things resolved, you can't put them anywhere.

mindcrime 3y ago

As much as I hate that AMD is not really supporting consumer GPUs for compute, Keep in mind this post is around two months old now, and since it was published AMD have already officially announced plans to support (at least some) consumer GPU's in ROCm.

imtringued 3y ago

Your comment reminds me of another comment by (SilverBirch I believe) that AMD can do this and will get away with it and that George Hotz is unimportant to AMD. Then a few weeks later Lisa Su tweeted this https://twitter.com/LisaSu/status/1669848494637735936 https://twitter.com/LisaSu/status/1669848494637735936.