Jesper Stemann Andersen
Jesper Stemann Andersen
Could be relevant wrt. GitLab CI as well, where GitLab CI forces the user to stick `.julia` (i.e. `JULIA_DEPOT_PATH`) in the project directory to enable caching between CI jobs, and...
Wondering why latency sometimes exceed 1000 us (and I've even seen latencies > 10,000 us): https://github.com/JuliaInterop/ZMQ.jl/runs/7505323278?check_suite_focus=true#step:7:124 when the latency tests in C never exceeds 200 us: https://github.com/JuliaInterop/ZMQ.jl/runs/7505323278?check_suite_focus=true#step:6:576 Latency test implementations...
I have run into something similar in relation to a CxxWrap-based-submodule, say `CppHello` in a project, say `Foo.jl`: ``` module Foo module CppHello using CxxWrap @wrapmodule(joinpath("path/to/built/lib","libhello")) function __init__() @initcxx end...
> What's the status on this? Can we include: https://github.com/LaurentMazare/tch-rs/tree/main/torch-sys/libtch to generate C bindings? > > Not having Windows support here doesn't seem like the end of the world. Will...
> Bump? Indeed. Just need to clean up the mess - CUDA hack(s) no longer necessary.
> > What's the status on this? Can we include: https://github.com/LaurentMazare/tch-rs/tree/main/torch-sys/libtch to generate C bindings? > > Not having Windows support here doesn't seem like the end of the world....
> Is this good to go now? Yes, LGTM :-) I plan to follow-up with a PR with a recipe for building an updated version of the C wrapper in...
Couldn't we just make MKL a dependency of MKL_Headers with compat = exact version ?
> Does this need to be a different package? Or can we turn this into a variant? If eventually both CPU, CUDA and ROCm builds are available, wouldn't the different...
> You still have the same issue that if Torch.jl includes `Torch_jll, TorchCUDA_jll,...` all dependencies must be fulfilled. Better to use a platform tag and platform specific dependencies. Right, I...