Ben Clayton

I have confirmed that this is not reproducible on Linux, using gcc 9.3.0, building with: `cmake .. -GNinja -DMARL_ASAN=1 -DCMAKE_CXX_FLAGS=-m32 -DCMAKE_C_FLAGS=-m32 -DCMAKE_ASM_FLAGS=-m32` I suspect this is a bug with MSVC's...

Reported to Microsoft: https://developercommunity.visualstudio.com/content/problem/1225394/possible-x86-asan-false-positives-when-using-win32.html

This could be broken down into two issues: 1. We're performing 'system' heap allocations outside of the Allocator interface. 2. unordered_map and deque aren't great for memory locality and heap...

Hi @waelhalbawi, The documentation isn't quite right, in that `will be executed immediately` should be `will be scheduled for execution immediately`. > It seems that the code schedules the callback...

[It's used in SwiftShader, but the code is pretty complex](https://github.com/google/swiftshader/blob/f1223ce969c78b15d536fa6a4a917b0deaedeb22/src/Device/Renderer.cpp#L566) The basic idea is that you want to avoid starting a fiber that immediately blocks on another sync primitive, as...

Microsoft closed the issue with: > We’re not able to [prioritize this issue](https://docs.microsoft.com/visualstudio/ide/developer-community-guidelines#prioritizing-which-issues-to-fix) over the other higher-impact issues we receive every week, based on the votes and comments from others...

A counter to `require`: `enable` being an ordinary verb makes sense. It is a directive to the compiler to do something - enable a feature. `require` is suggestive of something...

We've been through the bikeshedding of this before, [but to restate](https://github.com/gpuweb/gpuweb/issues/3431#issuecomment-1299976476) - my preferred syntax would be: ```c f32.highest // most positive, finite f32 value (0x1.fffffep+127f) f32.lowest // most negative,...

Answer: Yes you can. Updating Alan's code from above to latest WGSL, the following compiles with Tint: ```ts // Each call to foo, adds a pair for (t, s). fn...

Hi @Kojox, Sorry for the slow reply. I've been busy with [other things](https://developer.chrome.com/blog/webgpu-release). I don't know if this would work for you, but maybe a `marl::WaitGroup` could be used? *...