Leveling up CUDA Performance on WSL2 with New Enhancements

Originally published at: Leveling up CUDA Performance on WSL2 with New Enhancements | NVIDIA Technical Blog

In June 2020, we released the first NVIDIA Display Driver that enabled GPU acceleration in the Windows Subsystem for Linux (WSL) 2 for Windows Insider Program (WIP) Preview users. At that time, it was still an early preview with a limited set of features. A year later, as we have steadily added new capabilities, we…

Hi,

First, I want to thank all of you for the work you’re doing. It has made so much possible for so many people. I’ve been using RAPIDS with WSL to develop a GPU based ray tracer for astronomical simulations. I’ve had some trouble with share memory, linked below. Do you have any insight into if this is a limitation of WLS, or perhaps the existing memory space can be resized?

WSL2 CUDA/CUDF Unable to establish a shared memory space between system and Vram · Issue #7198 · microsoft/WSL (github.com)

Any further enhancements in WSL2 cuda performance done (or planned) since this update a year ago. Very much looking forward to using it when it gets closer to native Linux. Thx!