From 40b704d0c346ffd00087e4c86210522c43535e89 Mon Sep 17 00:00:00 2001 From: Zach Koopmans Date: Thu, 12 Dec 2024 17:01:11 -0800 Subject: [PATCH] Update gvisor.dev w/ instructions to get supported drivers. PiperOrigin-RevId: 705676256 --- ...2023-06-20-gpu-pytorch-stable-diffusion.md | 23 +++++++++++++------ 1 file changed, 16 insertions(+), 7 deletions(-) diff --git a/website/blog/2023-06-20-gpu-pytorch-stable-diffusion.md b/website/blog/2023-06-20-gpu-pytorch-stable-diffusion.md index 63a230f2b..946fdd77e 100644 --- a/website/blog/2023-06-20-gpu-pytorch-stable-diffusion.md +++ b/website/blog/2023-06-20-gpu-pytorch-stable-diffusion.md @@ -19,9 +19,17 @@ of sand. As of this writing (2023-06), [gVisor's GPU support][gVisor GPU support] is not generalized. Only some PyTorch workloads have been tested on NVIDIA T4, L4, -A100, and H100 GPUs, using the specific driver versions `525.60.13` and -`525.105.17`. Contributions are welcome to expand this set to support other GPUs -and driver versions! +A100, and H100 GPUs, using the specific driver versions that your runsc version +supports using the command below. Contributions are welcome to expand this set +to support other GPUs and driver versions! + +``` +# From a cloned gVisor repository: +$ make run TARGETS=runsc ARGS="nvproxy list-supported-drivers" + +# From a runsc binary: +$ runsc nvproxy list-supported-drivers +``` Additionally, while gVisor does its best to sandbox the workload, interacting with the GPU inherently requires running code on GPU hardware, where isolation @@ -98,14 +106,15 @@ currently compatible with. ```shell $ sudo apt-get update && sudo apt-get -y upgrade $ sudo apt-get install -y build-essential linux-headers-$(uname -r) -$ DRIVER_VERSION=525.60.13 +$ runsc nvproxy list-supported-drivers +$ DRIVER_VERSION=some-driver-version # Get from your runsc binary. $ curl -fSsl -O "https://us.download.nvidia.com/tesla/$DRIVER_VERSION/NVIDIA-Linux-x86_64-$DRIVER_VERSION.run" $ sudo sh NVIDIA-Linux-x86_64-$DRIVER_VERSION.run ``` Next, we install Docker, per [its instructions][Docker installation on Debian]. @@ -157,7 +166,7 @@ more of what we just set up. $ sudo nvidia-smi -L GPU 0: Tesla T4 (UUID: GPU-6a96a2af-2271-5627-34c5-91dcb4f408aa) $ sudo cat /proc/driver/nvidia/version -NVRM version: NVIDIA UNIX x86_64 Kernel Module 525.60.13 Wed Nov 30 06:39:21 UTC 2022 +NVRM version: NVIDIA UNIX x86_64 Kernel Module DRIVER_VERSION Wed Nov 30 06:39:21 UTC 2022 ＃ Check that Docker works. $ sudo docker version