141–150 of 154 posts
Earlier quoted context omitted.
did you learn anything?
I learned that AMD GPUs have layers upon layers of drivers, in user space, kernel space, and drivers running on the device itself.
They only open sourced the API. It's still closed firmware just like Nvidia.
Earlier quoted context omitted.
> After all, your CPU almost certainly runs firmware with all the same issues. CPUs don't really have firmware in the same sense. CPUs expose basic primitives (machine instructions). Modern CPUs will translate them to micro-ops in microcode instead of executing them directly, but the mapping is relatively simple and bugs are correspondingly uncommon. The GPU firmware is doing something complicated, and therefore bugg…
> CPUs don't really have firmware in the same sense. How would you classify e.g. the Intel Management Engine and the firmware that runs on it?
Earlier quoted context omitted.
> More concretely, what they open sourced calls closed source functions that run on the GPU which is where all the bugs are, and causes all the crashes. In particular, what they did was open source the drivers but not the firmware , and then when the firmware has bugs nobody outside the company can fix it because it's not open source.
Worth noting that is a step up from NVidia who moved their driver to the GPU and open sourced a configuration file.
and intel has a solution for this too, in fact…
One thing everyone should keep in mind about this NVIDIA/AMD battle right now: CUDA has been published for 16 years, it's been a huge push by NVIDIA to do GPGPU computation. I remember seeing it as a new thing in university back then, after the advanced shaders that were only available on NVIDIA. NVIDIA pretty rightful has the lead there, because they worked and invested into it for something like 20 years (you could…
(The Michelin star model of hardware sales - did you know that Michelin actually makes money from that book!? They’re not doing it because they’re financially disinterested, can you believe that!?!?)
Anyway net profit is more like 15% in a normal year. Recently it is actually lower, Ada is already lower margin than pascal for example.
It is only this high because nvidia finally struck good - and they spent a lot of effort and money that might never have pqidnoffZ
Earlier quoted context omitted.
Why do I have the feeling that CUDA will outlive whatever Khronos has proposed (do you refer to Sycl)?
How may standards Khronos endorced over the years/decades on compute? From an uneducated and external view, it seems every 2-5years there is a new standard.
Sycl's job isn't that, its meant to abstract implementations of common components across different kinds of hardware, and doesn't force you into any particular style of impl. As in, I could write a component for Sycl for my GPU in OpenCL and what Sycl would abstract away from the consumer of my component would be the entire usage of OpenCL itself; but I could write a component for a DSP, and it'd use an entirely closed source SDK for that hardware and is entirely opaque, and a Sycl user could use that impl for that function of they owned that DSP (instead of a CPU-based or GPU-based impl).
Also, Vulkan's compute doesn't replace OpenCL (not even in the sense that Vulkan, as a graphics API, replaces OpenGL). They're different levels of abstraction. Most Vulkan games are written almost entirely in compute shaders (ex: the powerhouse that is the Doom 2016 and Doom Eternal engines; and why they perform so fucking amazingly on paltry hardware like the original revision Xbox One, or hell, even the Switch).
In addition, I almost consider DX12 a flavor of Vulkan. Same job, written largely by the same people from the same companies, but instead of being OpenGL C-dialect flavored, its D3D C++-dialect flavored, but they both have entirely equivalent APIs that often call the same driver internals and produce nearly identical MIR. Microsoft did this on purpose to reflect the nature of how modern GPUs are almost entirely software renderers, sans certain parts of the texture units.
Earlier quoted context omitted.
NVIDIA’s consumer GPUs are highly reliable and support the same CUDA capabilities as the enterprise products. This is something NVIDIA gets right. Meanwhile, ROCm isn’t even supported on certain lower end AMD consumer chips.
Rocm is only supported on the last generation of flagship graphics cards. Which is a shame because the Radeon VII is still better than anything else when it comes to 64bit float computation.
I watched geohot streams and if he can go for an hour with a need to restart the computer due to crashed kernel driver/locked up card, it's a miracle (various computers, one or multi-card setup).
And he is not trying to crash the driver, just making the most basic stuff work.
I seriously think there is a deep HW bug that caused the difference in performance between RDNA3 presentation and reality and they had to do some horrible hacks to make it work.
Either that or the card never seen a tester for compute. Which is strange, considering they added it as a "supported" option.
Earlier quoted context omitted.
No thanks. We do need people to wreck the bullshit machines, but they don't have to be total asshats. That's not inherent to being that smart.
What if the "wants to wreck bullshit machines" trait often co-occurrs with the "grandstanding attitude" trait? You'd be unnecessarily restricting the applicant pool because you don't like how they express themselves, which seems pretty dumb given that we're being flooded by bullshit machines.
Earlier quoted context omitted.
Why is that, that AMD seemingly can't act like a sane individual would do?
AMD has simply never invested into software. Their code has been atrocious since even before AMD bought ATI. ATI "Catalyst Control Center" was their consumer driver code before Vista and into Windows 7 IIRC, and that was utter trash. Granted, nVidia's drivers were ALSO trash back then, accounting for literally 65% of ALL Vista BSODs. nVidia decided to redouble their efforts, and now they might still crash occasionall…
Earlier quoted context omitted.
Worth noting that is a step up from NVidia who moved their driver to the GPU and open sourced a configuration file.
That is an interesting way to frame the situation when AMD’s solution doesn’t even run hdmi 2.1 under Linux. and intel has a solution for this too, in fact…
Earlier quoted context omitted.
> This is where the melodrama kicks in. The complaint was that they couldn't debug the GPU. Now they can, so they're soldiering on. Seems reasonable to me.
Surely if they had done a good job communicating with AMD they could have figured that out beforehand. The repo was already public.