They aren't going to win anything until they fix their laughably bad drivers and keep them fixed. I have several Ryzens, but there's no way I'd go anywhere near their graphics cards.
Agree, I dropped Linux due bad graphics drivers (AMD)
AMD reckons it can win the high-end GPU battle
11–20 of 160 posts
Re: AMD reckons it can win the high-end GPU battle
#12They aren't going to win anything until they fix their laughably bad drivers and keep them fixed. I have several Ryzens, but there's no way I'd go anywhere near their graphics cards.
I'm running an RX 570 on Linux and it works fine for my purposes (work, Steam games & Blizzard games under wine).
Same with a 580. Linux, many monitors, work, light gaming. Drivers are the first time I've ever not had issues with Linux graphics.
Three years in I'm still waiting for the other shoe to drop. Where are the driver issues? I'm certain it was like a law of physics in Linux, but somehow, not this time, not for me?
Re: AMD reckons it can win the high-end GPU battle
#13It needs the help from Nvidia just like Intel helped them winning the CPU battle.
Re: AMD reckons it can win the high-end GPU battle
#14Their software things (drivers, compilers, apis) are laughable.
I would add libraries to that list (cuda equivalent).
Re: AMD reckons it can win the high-end GPU battle
#15Latest Ampere datacenter rumors are 70% speedup compared to V100, meaning 25TFlops in FP32, whereas AMD in theory could to 15Tflops and in practice only 8Tflops due to their software. They would have to improve speed 3x just to match NVidia's next generation GPUs; I am not sure upcoming Big Navi would be sufficient.
It's not comparable in efficiency but AMD is already shipping Vega II modules in the Mac Pro that claim "up to 28.3 FP32 TFLOPS". And regardless, the article is about 4K gaming. I don't doubt Nvidia will have some 7nm monster but there might be a time period before it launches where AMD briefly regains the crown.
That's not equivalent. That's a card with two ASICs on it, with each being 14TFLOPS. It's identical to the V100, but 2-3 years later.
Re: AMD reckons it can win the high-end GPU battle
#16Latest Ampere datacenter rumors are 70% speedup compared to V100, meaning 25TFlops in FP32, whereas AMD in theory could to 15Tflops and in practice only 8Tflops due to their software. They would have to improve speed 3x just to match NVidia's next generation GPUs; I am not sure upcoming Big Navi would be sufficient.
I've seen those A100 rumors, but I think they should be taken with a grain of salt. They did not specify where the 70% speedup will be, and it's highly likely it will not be fp32 increasing by that much.
Re: AMD reckons it can win the high-end GPU battle
#17Latest Ampere datacenter rumors are 70% speedup compared to V100, meaning 25TFlops in FP32, whereas AMD in theory could to 15Tflops and in practice only 8Tflops due to their software. They would have to improve speed 3x just to match NVidia's next generation GPUs; I am not sure upcoming Big Navi would be sufficient.
You’re being downvoted but it’s true. People don’t realize that the data center is now driving the GPU market volume, and AMD is not yet seriously competing in this space.
Re: AMD reckons it can win the high-end GPU battle
#18They aren't going to win anything until they fix their laughably bad drivers and keep them fixed. I have several Ryzens, but there's no way I'd go anywhere near their graphics cards.
Agree, I dropped Linux due bad graphics drivers (AMD)
How long ago did you try it? It was definitely bad in the fglrx days!
Re: AMD reckons it can win the high-end GPU battle
#19And when that day comes, I will buy it. But for now I will go with nVidia. Also I badly need same CUDA codebase/documentation on AMD part as well to make such a move. ML is such a breeze when you throw CUDA at it these days.
Re: AMD reckons it can win the high-end GPU battle
#20AMD's rather old Radeon VII can really smash a 2080 Ti if you're doing relatively simple streaming computation, like say video processing. That 1TB/s HBM is no joke, and there's 16GB of it, at ~half the price of the 2080 Ti with 11GB.
Nvidia's memory compression is very impressive and allows them to do better with a narrower memory bus in many situations (such as games), however if your data is not so compressible (often the case in GPGPU) this advantage disappears.
Really all I care about is a good OpenCL driver implementation, especially OpenCL 2.2 support would be extremely attractive.