juj
http://clbri.com:8010/api/v2/logs/107593/raw_inline
Peculiarly I find that in Chrome, the max ArrayBuffer limit is not 2GB (2048MB), but 2046MB: Firefox supports up to 16GB ArrayBuffers: In node.js I get up to 4GB ArrayBuffers:...
Several developers have reported this showing up as a problem. Talking with them, it is interesting how this seems to be one of those features that trapping instructions seem to...
> @juj, wouldn't it be trivial for producers to generate such debugging code themselves, on top of non-trapping instructions? Yeah, that's true, if one had nontrapping instructions, adding trapping variants...
I'll try to find a moment to look into this further - right now looking at some of the more immediate errors.
CC @jdashg , @kenrussell , @kainino0x
https://github.com/juj/wasm_webgpu/blob/cb2f20154aba141c3e80d485a59d4eb7f41f8946/samples/buffer_map_sync/buffer_map_sync.c#L19-L31 shows a demo of doing synchronous buffer mapping from C code with Emscripten and WebGPU by using the upcoming JSPI API. (i.e. using `-sJSPI` link flag in Emscripten) https://github.com/juj/wasm_webgpu/blob/cb2f20154aba141c3e80d485a59d4eb7f41f8946/samples/clear_screen/clear_screen_sync.c#L27-L58...
I agree, given that we do have a --hostname=0.0.0.0 option, we would now be able to default to listening to 127.0.0.1. I would recommend that when emrun listens to localhost/127.0.0.1,...
> Do we have a policy on when we update the default set of wasm features enabled? > How long do we want to wait before enabling these by default...
That looks great - yeah, with that kind of info accumulated, users will be able to make a decision on which features they'll care about.