4 ms·

There are already packages that let people run CUDA programs unmodified on other GPUs: see https://news.ycombinator.com/item?id=40970560 https://news.ycombinato

There are already packages that let people run CUDA programs unmodified on other GPUs: see https://news.ycombinator.com/item?id=40970560 https://news.ycombinator.com/item?id=40970560 For whatever reason, people just delete these tools from their minds, then claim Nvidia still has a monopoly on CUDA.

jcranmer 2y ago
There's already a panoply of CUDA alternatives, and even several CUDA-to-non-Nvidia-GPU alternatives (which aren't supported by the hardware vendors and are in some sense riskier). To my knowledge (this isn't really my space), many of the higher-level frameworks already support these CUDA alternatives. And yet still the popcorn gallery says "there no [realistic] alternative to CUDA." Methinks the real issue is that CUDA is the best software solution for Nvidia GPUs, and the alternative hardware vendors aren't seen as viable competitor for hardware reasons, and people attribute the failure to software failures.
- quotemstr 2y ago
  > alternative hardware vendors aren't seen as viable competitor for hardware reasons, and people attribute the failure to software failures. It certainly seems like there's a "nobody ever got fired for buying nvidia" dynamic going on. We've seen this mentality repeatedly in other areas of the industry: that's why the phrase is a snowclone. Eventually, someone is going to use non-nvidia GPU accelerators and get a big enough cost or performance win that industry attitudes will change.
- jdewerd 2y ago
  > There's already a panoply of CUDA alternatives Is there? 10 years ago, I burned about 6 months of project time slogging through AMD / OpenCL bugs before realizing that I was being an absolute idiot and that the green tax was far cheaper than the time I was wasting. If you asked AMD, they would tell you that OpenCL was ready for new applications and support was right around the corner for old applications. This was incorrect on both counts. Disastrously so, if you trusted them. I learned not to trust them. Over the years, they kept making the same false promises and failing to deliver, year after year, generation after generation of grad students and HPC experts, filling the industry with once-burned-twice-shy received wisdom. When NVDA pumped and AMD didn't, presumably AMD could no longer deny the inadequacy of their offerings and launched an effort to fix their shit. Eventually I am sure it will bear fruit. But is their shit actually fixed? Keeping in mind that they have proven time and time and time and time again that they cannot be trusted to answer this question themselves? 80% margins won't last forever, but the trust deficit that needs to be crossed first shouldn't be understated.
  - arresin 2y ago
    This is absolutely it. You pay the premium not to have to deal with the BS.
- lmm 2y ago
  > There's already a panoply of CUDA alternatives, and even several CUDA-to-non-Nvidia-GPU alternatives (which aren't supported by the hardware vendors and are in some sense riskier). To my knowledge (this isn't really my space), many of the higher-level frameworks already support these CUDA alternatives. On paper, yes. But how many of them actually work? Every couple of years AMD puts out a press release saying they're getting serious this time and will fully support their thing, and then a couple of people try it and it doesn't work (or maybe the basic hello world test works, but anything else is too buggy), and they give up.
dzdt 2y ago
And which of these have the level of support that would let a company put a multi-million dollar project on top of?
- equestria 2y ago
  We have trillions of dollars riding on one-person open-source projects. This is not the barrier for "serious businesses" that it used to be.
  - error_logic 2y ago
    Resilience is not as valued as it should be... Average bus factor is how small these days? :/
  - p1esk 2y ago
    What are you talking about?
stonogo 2y ago
Those packages only really perform with low-precision work. For scientific computing, using anything but CUDA is a painful workflow. DOE has been deploying AMD and Intel alternatives in their leadership class machines and it's been a pretty bad speedbump.
- smcin 2y ago
  ('DOE' = US Department of Energy)

jcranmer 2y ago

There's already a panoply of CUDA alternatives, and even several CUDA-to-non-Nvidia-GPU alternatives (which aren't supported by the hardware vendors and are in some sense riskier). To my knowledge (this isn't really my space), many of the higher-level frameworks already support these CUDA alternatives. And yet still the popcorn gallery says "there no [realistic] alternative to CUDA." Methinks the real issue is that CUDA is the best software solution for Nvidia GPUs, and the alternative hardware vendors aren't seen as viable competitor for hardware reasons, and people attribute the failure to software failures.

quotemstr 2y ago
> alternative hardware vendors aren't seen as viable competitor for hardware reasons, and people attribute the failure to software failures. It certainly seems like there's a "nobody ever got fired for buying nvidia" dynamic going on. We've seen this mentality repeatedly in other areas of the industry: that's why the phrase is a snowclone. Eventually, someone is going to use non-nvidia GPU accelerators and get a big enough cost or performance win that industry attitudes will change.
jdewerd 2y ago
> There's already a panoply of CUDA alternatives Is there? 10 years ago, I burned about 6 months of project time slogging through AMD / OpenCL bugs before realizing that I was being an absolute idiot and that the green tax was far cheaper than the time I was wasting. If you asked AMD, they would tell you that OpenCL was ready for new applications and support was right around the corner for old applications. This was incorrect on both counts. Disastrously so, if you trusted them. I learned not to trust them. Over the years, they kept making the same false promises and failing to deliver, year after year, generation after generation of grad students and HPC experts, filling the industry with once-burned-twice-shy received wisdom. When NVDA pumped and AMD didn't, presumably AMD could no longer deny the inadequacy of their offerings and launched an effort to fix their shit. Eventually I am sure it will bear fruit. But is their shit actually fixed? Keeping in mind that they have proven time and time and time and time again that they cannot be trusted to answer this question themselves? 80% margins won't last forever, but the trust deficit that needs to be crossed first shouldn't be understated.
- arresin 2y ago
  This is absolutely it. You pay the premium not to have to deal with the BS.
lmm 2y ago
> There's already a panoply of CUDA alternatives, and even several CUDA-to-non-Nvidia-GPU alternatives (which aren't supported by the hardware vendors and are in some sense riskier). To my knowledge (this isn't really my space), many of the higher-level frameworks already support these CUDA alternatives. On paper, yes. But how many of them actually work? Every couple of years AMD puts out a press release saying they're getting serious this time and will fully support their thing, and then a couple of people try it and it doesn't work (or maybe the basic hello world test works, but anything else is too buggy), and they give up.

dzdt 2y ago

And which of these have the level of support that would let a company put a multi-million dollar project on top of?

equestria 2y ago
We have trillions of dollars riding on one-person open-source projects. This is not the barrier for "serious businesses" that it used to be.
- error_logic 2y ago
  Resilience is not as valued as it should be... Average bus factor is how small these days? :/
- p1esk 2y ago
  What are you talking about?

stonogo 2y ago

Those packages only really perform with low-precision work. For scientific computing, using anything but CUDA is a painful workflow. DOE has been deploying AMD and Intel alternatives in their leadership class machines and it's been a pretty bad speedbump.

smcin 2y ago
('DOE' = US Department of Energy)