181–190 of 548 posts
I really wish Microsoft would just tell NVidia (and AMD) to support SR-IOV on all their GPUs. Then we wouldn't need any major software changes to enable CUDA acceleration within VMs, just a configuration change to pass through a virtual function of the GPU to make it available to existing drivers.
Earlier quoted context omitted.
Why would it? That's a functionality that can't be used without Microsoft proprietary parts. Unless I missed something, I don't think there is an open source implementation of DirectX somewhere? I doubt the linux devs ever see WLS as a target they have to maintain themselves.
Wine has an open source implementation of DirectX, FWIW.
Using that with wine would mean adding two emulation layers before reaching the actual driver. I fail to see any use case for that.
Earlier quoted context omitted.
If you read the replies by Dave Airlie and Daniel Vetter, it seems somewhat likely that upstream won't accept it. Perhaps that's just initial skepticism that will evaporate after more discussion, but perhaps not. Frankly this does just seem like MS wanting to reduce their maintenance burden on what they expect will be a very important part of their WSL offering on Windows. There's nothing inherently wrong with that d…
>>>There's also the question of whether or not you believe WSL as a whole is good or bad for Linux. IMO it's a good thing. Given that windows accounts for 90%+ of the desktop OS share, Windows might very well become the world's most used Linux distro.
Earlier quoted context omitted.
I've had trouble researching if WSL2 is in fact a hyper-v managed VM. I've seen some documentation referring to WSL2 as a tightly integrated Krypton (scaled down hyper-V) VM. It seems to imply the host overhead isn't as high as a guest on hyper-V
WSL uses a Hyper-V derived virtual machine that is * Sparse & light - they only allocate resources from the host when needed, and release them back to the host when freed * Fast - it can boot a WSL distro from cold in Full Hyper-V VMs aim to (generally) grab all the resources they can and keep hold of those resources as long as possible in case they're needed. Full VMs are designed to run for months-years at a time.…
I imagine this will be addressed, but claims of lightweight seem exaggerated?
But even more on my mind is the impact on the windows host. Is it running as a guest under hyper v? What's the overhead?
Earlier quoted context omitted.
They are also adding an wayland server. https://devblogs.microsoft.com/commandline/the-windows-subsy...
Aren't large parts of the wayland stack incompatible with NVIDIA drivers? It would be ironic if Microsoft was the one to bridge that gap.
> This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment. Dollars to donuts this is why Microsoft is implementing this. GPU acceleration is becoming a critical feature for many users (but especially developers) and this will continue. If WSL is to be a serious competitor, this is necessary and I'm glad to see it showing up. This is true of…
Do ML people actually care about DirectX? I thought everyone is using CUDA? Anyway in my university building I have not seen anyone that does machine learning on Windows.
Earlier quoted context omitted.
Because this benefits only WSL. I don’t think that the Linux maintainers should spend time and effort maintaining a shim that only benefits users on another operating system. It does not even lead to cross platform code if a dev uses this on WSL. Microsoft can ship it however they please though within bounds of licenses.
They've already upstreamed lots of other hyper-v paravirtualization infrastructure. And why do you assume it only helps WSL? Doesn't this help all hyper-v setups get access to GPUs, including azure?
FTR, I'm not arguing whether or not this should be upstreamed. I just see where the other poster could be coming from. I could be wrong on my take and if so, someone please correct me.
Earlier quoted context omitted.
They've already upstreamed lots of other hyper-v paravirtualization infrastructure. And why do you assume it only helps WSL? Doesn't this help all hyper-v setups get access to GPUs, including azure?
Probably because its a shim to use a proprietary (only distributed by MS) binary blob version of DirectX compiled for Linux. Outside of ML utilizing DirectX in WSL on Windows, the only group that this helps is MS. FTR, I'm not arguing whether or not this should be upstreamed. I just see where the other poster could be coming from. I could be wrong on my take and if so, someone please correct me.
Would y'all be so worked up about VMWare upstreaming their virtual GPU bridge?
> This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment. Dollars to donuts this is why Microsoft is implementing this. GPU acceleration is becoming a critical feature for many users (but especially developers) and this will continue. If WSL is to be a serious competitor, this is necessary and I'm glad to see it showing up. This is true of…