Skip to main content


My colleague Antonio Maiorano added some basic support for capturing Chrome #WebGPU content with RenderDoc recently! I wrote a brief article about how to use it. It's pretty primitive right now, but I have high hopes that we can improve the ergonomics in the future.

toji.dev/webgpu-profiling/rend…

in reply to mcc

@mcc Curious what "deployable" means for you. I'm very happy to hear how this could be made useful for devs. It's rough and limited right now, but I want to make it better.
@mcc
in reply to Brandon Jones

"deployable" = "if i sat down and tried to get it working over the course of an afternoon, would it

- get working;
- get working, but in a way i would never be able to get working again without sacrificing another afternoon each time i wanted to take a single capture;
- not wind up working before i got frustrated and gave up".

I need to clarify i didn't read the blog post, it is 10:30 and my hands are still cold (from the snow). I definitely want to later tho

in reply to mcc

@mcc Got it. Current state is "Mostly just get working, but you'll probably have too look up a couple of command line flags if it's been a while." Other challenges are it's currently Win/D3D12 only and you can't control which frames it captures, which are likely to be blockers to serious use ATM.
@mcc
in reply to Brandon Jones

heck as long as i don't have to build chromium myself i'm happy