In my testing, a looping compute shader was enough to crash the tab on its own, but it needed waiting render shaders to crash the WindowServer.

Interestingly though there was another way to make only the tab crash, even if I had all three shaders in the pipeline: If I placed the canvas far offscreen using position: absolute, only the tab would crash even if the render shaders were waiting! There's some weird interactions going on I don't yet fully understand.

Historically, for loops in shaders were limited in the number of iterations they could run for. Among other things this ensures that rasterizing a particular pixel completes in a known amount of time (and ideally that amount of time is fast enough to avoid triggering TDR on windows and making the machine bluescreen). You could of course nest loops so it's not a perfect measure. I'm not certain whether that limitation applies to WebGPU, but it should apply to WebGL.