151–160 of 160 posts
Earlier quoted context omitted.
I recently purchased a laptop with an Nvidia/Intel composite setup. As a Linux user, I’ll never buy a laptop with an Nvidia GPU again (if I can help it). What a pain. A nice piece of hardware with relatively pathetic software to drive it.
I think using pop-os can solve the problem
I’ll believe it when I see it. ATI has never made a good product and nothing has changed recently to turn them around.
Earlier quoted context omitted.
AMD'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. The Radeon VII is also pretty fast in training transformer networks. In some models that I trained, it was only ~25% slower than our Quadro RTX 5000s.
On Tensorflow ROC (or whatever the name is?)
Earlier quoted context omitted.
My condolences. On Nvidia there is a Full Composition Pipeline option in Nvidia driver settings which gets rid of vsync issues. It's odd it's not on by default.
Wow, that's good to know! How do you turn it on? (I don't see an option for it in the Nvidia X Server Settings app.)
Note, it is not saved during a reboot unless you click 'Save to X Configuration File' and then save to /etc/X11/xorg.conf for most distros.
Also, many distros have partial v-sync of some sort on. For most it isn't noticeable, but for many desktops it causes the screen to run at an effective 59 instead of 60fps (or that is what it looks like to me, non-scientific here), so it can be beneficial to look up how to turn off vsync on your distro. Eg, I'm on Mint so for me it's General Settings -> VSync method -> None. For most people, this is not a noticeable change.
The big problem right now is that ROCm, AMD's equivalent to Cuda for deep learning, is not supported for the consumer Radeon Navi cards. I had hoped it would be supported in ROCm 3.0 when that was released - but there has been no official word, which is worrying. If AMD are to compete in the GPU/DeepLearning space, it will be because they have hardware that is competitive (not even as good as Nvidia), low cost, and d…
Earlier quoted context omitted.
Nvidia doesn't make any cards with two ASICs on it. There is no equivalent.
Yes, they make the K40, which has two ASICs. They do not make a current generation card with two ASICs.
Earlier quoted context omitted.
Yes, they make the K40, which has two ASICs. They do not make a current generation card with two ASICs.
You probably meant K80. K80 = 2x K40 basically, i.e. 2x 12GB 4TFlops GPUs on a single board.
Earlier quoted context omitted.
Wow, that's good to know! How do you turn it on? (I don't see an option for it in the Nvidia X Server Settings app.)
Nvidia X Server Settings app -> X Server Display Configuration -> Advanced... -> Force Full Composition Pipeline Note, it is not saved during a reboot unless you click 'Save to X Configuration File' and then save to /etc/X11/xorg.conf for most distros. Also, many distros have partial v-sync of some sort on. For most it isn't noticeable, but for many desktops it causes the screen to run at an effective 59 instead of 6…
Earlier quoted context omitted.
I chose an AMD card specifically because of their great Linux drivers. Building a system on top of proprietary drivers is a non starter for me. Being heavily dependent on a vendor to support your platform has bitten me multiple times and it's especially frustrating when a proprietary software company uses something like electron but never ships a Linux release. Running these applications in wine is a complete pain in…
I’ve found amdgpu to be a terrible driver for high end displays; 8K doesn’t work, for example.
Earlier quoted context omitted.
I’ve found amdgpu to be a terrible driver for high end displays; 8K doesn’t work, for example.
FYI, 8K/4K120 support was fixed with Linux 5.5 release.