Julian Samaroo

If Cthulhu is crashing on any code, you should file an issue on its repo. It's really important we have a working tool to easily inspect the typed IR, otherwise...

Sorry for the slow reply - this is probably a known memory leak, also reported offline by @mofeing in a similar case. I'll investigate and see if I can resolve...

Using the example above, I've found the initial source of retained memory, and am fixing it in https://github.com/JuliaParallel/Dagger.jl/pull/558 (branch is very WIP, expect it to not work right now). I'll...

`AMDGPU.versioninfo()` is reporting HIP is missing, so no real code will work. Please install the ROCm SDK and verify that it does in fact work. We can address wrapping the...

> https://github.com/ROCm/TheRock/blob/main/ROADMAP.md shows that no RDNA2 hardware is supported. I wouldn't trust what AMD officially supports, I would try it for yourself first. Sometimes what is listed as not supported...

Just want to say that I really like this approach! ❤️ I especially like the design of making `Base.@cancel_check` a point that the task can longjmp back to from effect-free...

Regarding Threads: I would never expect using `Threads.@spawn` to be *slower* than Dagger - we use `Threads.@spawn` for our own Dagger tasks. The example in the docs was about `Threads.@threads`,...

Another counter-point to https://github.com/JuliaLang/julia/pull/55047 as an alternative viable solution: not all Julia programs can be made free from dynamic dispatch, yet https://github.com/JuliaLang/julia/pull/55047 requires that for the programs it generates. In...

Seems like the `LocalFuture` changes are having some concurrency issues, presumably due to the shenanigans we perform around `ThunkFuture` serialization/deserialization. The other commit was already merged, so this PR is...

This might be happening because of our distributed refcounting logic in MemPool.jl. Dagger's `DArray` is implemented to work like a normal `Array`, such that it cleans itself up when no...