Anton Smirnov

Results 52 issues of Anton Smirnov

I'd argue that we also need to check for inferrability of a function. Since it is an important factor when considering performance, see for example: https://github.com/FluxML/NNlib.jl/pull/370.

Hi. I'm currently implementing self-supervised depth estimation network, similar to Monodepth2 and would like to point out a possible performance concern. Taking gradient for the first time takes very long...

performance

Following kernel produces segfault on AMDGPU. On CUDA it works though. **MWE:** ```julia using AMDGPU using ROCKernels using KernelAbstractions Base.zeros(::ROCDevice, T, shape) = AMDGPU.zeros(T, shape) linear_threads(::ROCDevice) = 512 @kernel function...

Hi! I get the following error when trying to run LanguageServer in Vim with YCM (as described in their official [examples](https://github.com/ycm-core/lsp-examples#julia)): ``` [ Info: Received new data from Julia Symbol...

bug
non-vscode
vim

When `device = CPU()` and Julia is started with more than one thread (e.g. `-t16`), the program stalls. **MWE:** ```julia using Enzyme using KernelAbstractions using KernelGradients linear_threads(::CPU) = Threads.nthreads() Base.zeros(::CPU,...

When building rocBLAS (ROCm 5.2.3) with `AMDGPU_TARGETS="all"` on `musl` it errors with the following error (see below). Tried reducing number of threads, but that didn't help (although I haven't tried...

I encounter this bug on Julia 1.7.3. Julia 1.8.0-rc1 does not have this. Here's MWE: [R.tar.gz](https://github.com/JuliaPackaging/Preferences.jl/files/9127521/R.tar.gz) The package is very minimal, it only loads preference `const BACKEND = @load_preference("backend")` from...

```julia function ker(x) x[1] = pointer_from_objref(Int64) return end ``` Because `pointer_from_objref` is no longer const-prop'd. - 1.10: ```julia julia> @code_llvm pointer_from_objref(Int64) ; @ pointer.jl:267 within `pointer_from_objref` define i64 @julia_pointer_from_objref_146() #0...