# Unable to test or use HIP libraries on Windows 10

**URL:** <https://discourse.julialang.org/t/unable-to-test-or-use-hip-libraries-on-windows-10/109639>\
**Category:** GPU\
**Tags:** amdgpu\
**Created:** [February 2, 2024, 6:16pm UTC](https://discourse.julialang.org/t/unable-to-test-or-use-hip-libraries-on-windows-10/109639 "2024-02-02T18:16:18Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![mp59](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/mp59/32/206717_2.png) [@mp59](https://discourse.julialang.org/u/mp59)\
**Post date:** [February 2, 2024, 6:16pm UTC](https://discourse.julialang.org/t/unable-to-test-or-use-hip-libraries-on-windows-10/109639/1 "2024-02-02T18:16:18Z")

</div>

Hello,  
I am relatively new to Julia and very new to GPU programming. Recently I decided to use AMDGPU.jl, however, it does not seem like Julia can AMD’s HIP libraries, even though it recognizes that they are there. Here’s a minimum working example:

> **AMDGPU HIP Test**
>
> ```julia
> using Pkg
> Pkg.test("AMDGPU", test_args=["hip"])
> Testing Running tests...
> ┌ Warning: MIOpen is unavailable, functionality will be disabled. # I am not doing any ML so this is fine
> └ @ AMDGPU C:\Users\*\.julia\packages\AMDGPU\BhNdC\src\AMDGPU.jl:213
> [Info: Running following tests: ["hip"].
> [ Info: Running tests with 2 workers.
> [ Info: Testing using device HIPDevice(name="AMD Radeon(TM) Graphics", id=1, gcn_arch=gfx90c:xnack-).
> Julia Version 1.10.0
> Commit 3120989f39 (2023-12-25 18:01 UTC)
> Build Info:
> Official https://julialang.org/ release
> Platform Info:
> OS: Windows (x86_64-w64-mingw32)
> CPU: 12 × AMD Ryzen 5 5500U with Radeon Graphics
> WORD_SIZE: 64
> LIBM: libopenlibm
> LLVM: libLLVM-15.0.7 (ORCJIT, znver2)
> Threads: 17 on 12 virtual cores
> Environment:
> JULIA_LOAD_PATH = @;C:\Users\*\AppData\Local\Temp\jl_sLCpo0
> JULIA_EDITOR = code
> ROCm provided by: system
> [+] ld.lld
> @ C:\Program Files\AMD\ROCm\5.5\bin\ld.lld.exe # I am using version 5.5 to hopefully smooth over compatibility issues between ROCm and Julia
> [+] ROCm-Device-Libs
> @ C:\Users\*\.julia\artifacts\5ad5ecb46e3c334821f54c1feecc6c152b7b6a45\amdgcn/bitcode
> [+] HIP Runtime v5.7.31541
> @ C:\Windows\SYSTEM32\amdhip64.DLL
> [+] rocBLAS v2.47.0
> @ C:\Program Files\AMD\ROCm\5.5\bin\rocblas.dll
> [+] rocSOLVER v3.21.0
> @ C:\Program Files\AMD\ROCm\5.5\bin\rocsolver.dll
> [+] rocALUTION
> @ C:\Program Files\AMD\ROCm\5.5\bin\rocalution.dll
> [+] rocSPARSE
> @ C:\Program Files\AMD\ROCm\5.5\bin\rocsparse.dll
> [+] rocRAND v2.10.5
> @ C:\Program Files\AMD\ROCm\5.5\bin\rocrand.dll
> [+] rocFFT v1.0.21
> @ C:\Program Files\AMD\ROCm\5.5\bin\rocfft.dll
> [-] MIOpen
> 
> HIP Devices [1]
> 1. HIPDevice(name="AMD Radeon(TM) Graphics", id=1, gcn_arch=gfx90c:xnack-)
> [ Info: Scanning for test items in project `AMDGPU` at paths: C:\Users\*\.julia\packages\AMDGPU\BhNdC
> [ Info: Finished scanning for test items in 2.05 seconds. Scheduling 2 tests on pid 6032 with 2 worker processes and 1 threads per worker.
> [ Info: Starting test workers
> Worker 4376: [ Info: Starting test worker 1 on pid = 4376, with 1 threads
> Worker 9808: [ Info: Starting test worker 2 on pid = 9808, with 1 threads[ Info: Starting running test items
> 
> Worker 9808: 11:53:27 | maxrss 4.4% | mem 83.4% | START (2/2) test item "hip - extra" at test\hip_extra_tests.jl:1
> Worker 4376: 11:53:27 | maxrss 4.4% | mem 83.4% | START (1/2) test item "hip - core" at test\hip_core_tests.jl:1
> 
> Captured Logs for test item "hip - core" at test\hip_core_tests.jl:1 on worker 4376
> 
> rocBLAS error: Cannot read C:\Program Files\AMD\ROCm\5.5\bin\/rocblas/library/TensileLibrary.dat: No such file or directory
> ┌ Error: Worker(pid=4376, terminated=true, termsignal=0) died running test item "hip - core". Recording test error.
> └ @ ReTestItems C:\Users\*\.julia\packages\ReTestItems\ELM7z\src\ReTestItems.jl:561
> 
> Captured Logs for test item "hip - extra" at test\hip_extra_tests.jl:1 on worker 9808
> 
> rocBLAS error: Cannot read C:\Program Files\AMD\ROCm\5.5\bin\/rocblas/library/TensileLibrary.dat: No such file or directory
> ┌ Error: Worker(pid=9808, terminated=true, termsignal=-2147483648) died running test item "hip - extra". Recording test error.
> └ @ ReTestItems C:\Users\*\.julia\packages\ReTestItems\ELM7z\src\ReTestItems.jl:561
> Test Summary: | Error Total Time
> AMDGPU | 2 2 1m11.3s
> test | 2 2  
> test\hip_core_tests.jl | 1 1  
> hip - core | 1 1 0.0s
> test\hip_extra_tests.jl | 1 1  
> hip - extra | 1 1 0.0s
> ERROR: LoadError: Some tests did not pass: 0 passed, 0 failed, 2 errored, 0 broken.
> in expression starting at C:\Users\*\.julia\packages\AMDGPU\BhNdC\test\runtests.jl:105
> ERROR: Package AMDGPU errored during testing
> Stacktrace:
> [1] pkgerror(msg::String)
> @ Pkg.Types C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\Types.jl:70
> [2] test(ctx::Pkg.Types.Context, pkgs::Vector{…}; coverage::Bool, julia_args::Cmd, test_args::Cmd, test_fn::Nothing, force_latest_compatible_version::Bool, allow_earlier_backwards_compatible_versions::Bool, allow_reresolve::Bool)
> @ Pkg.Operations C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\Operations.jl:2018       
> [3] test
> @ C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\Operations.jl:1899 [inlined]
> [4] test(ctx::Pkg.Types.Context, pkgs::Vector{…}; coverage::Bool, test_fn::Nothing, julia_args::Cmd, test_args::Vector{…}, force_latest_compatible_version::Bool, allow_earlier_backwards_compatible_versions::Bool, allow_reresolve::Bool, kwargs::@Kwargs{…})
> @ Pkg.API C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\API.jl:444
> [5] test(pkgs::Vector{Pkg.Types.PackageSpec}; io::Base.TTY, kwargs::@Kwargs{test_args::Vector{String}})
> @ Pkg.API C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\API.jl:159
> [6] test
> @ Pkg.API C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\API.jl:147 [inlined]
> [7] #test#74
> @ Pkg.API C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\API.jl:146 [inlined]
> [8] top-level scope
> @ c:\Users\*\Documents\GoL\CGoL.jl\scripts\AMDGPUtest.jl:4
> Some type information was truncated. Use `show(err)` to see complete types.
> 
> julia> show(err)
> 1-element ExceptionStack:
> LoadError: Package AMDGPU errored during testing
> Stacktrace:
> [1] pkgerror(msg::String)
> @ Pkg.Types C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\Types.jl:70
> [2] test(ctx::Pkg.Types.Context, pkgs::Vector{Pkg.Types.PackageSpec}; coverage::Bool, julia_args::Cmd, test_args::Cmd, test_fn::Nothing, 
> force_latest_compatible_version::Bool, allow_earlier_backwards_compatible_versions::Bool, allow_reresolve::Bool)
> @ Pkg.Operations C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\Operations.jl:2018      
> [3] test
> @ C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\Operations.jl:1899 [inlined]
> [4] test(ctx::Pkg.Types.Context, pkgs::Vector{Pkg.Types.PackageSpec}; coverage::Bool, test_fn::Nothing, julia_args::Cmd, test_args::Vector{String}, force_latest_compatible_version::Bool, allow_earlier_backwards_compatible_versions::Bool, allow_reresolve::Bool, kwargs::@Kwargs{io::Base.TTY})
> @ Pkg.API C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\API.jl:444
> [5] test(pkgs::Vector{Pkg.Types.PackageSpec}; io::Base.TTY, kwargs::@Kwargs{test_args::Vector{String}})
> @ Pkg.API C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\API.jl:159
> [6] test
> @ Pkg.API C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\API.jl:147 [inlined]
> [7] #test#74
> @ Pkg.API C:\Users\*\.julia\juliaup\julia-1.10.0+0.x64.w64.mingw32\share\julia\stdlib\v1.10\Pkg\src\API.jl:146 [inlined]
> [8] top-level scope
> @ c:\Users\*\Documents\GoL\CGoL.jl\scripts\AMDGPUtest.jl:4
> [9] eval
> @ Core .\boot.jl:385 [inlined]
> [10] include_string(mapexpr::typeof(identity), mod::Module, code::String, filename::String)
> @ Base .\loading.jl:2070
> [11] include_string(m::Module, txt::String, fname::String)
> @ Base .\loading.jl:2080
> [12] invokelatest(::Any, ::Any, ::Vararg{Any}; kwargs::@Kwargs{})
> @ Base .\essentials.jl:887
> [13] invokelatest(::Any, ::Any, ::Vararg{Any})
> @ Base .\essentials.jl:884
> [14] inlineeval(m::Module, code::String, code_line::Int64, code_column::Int64, file::String; softscope::Bool)
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:263
> [15] (::VSCodeServer.var"#67#72"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams})()
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:181
> [16] withpath(f::VSCodeServer.var"#67#72"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams}, 
> path::String)
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\repl.jl:274
> [17] (::VSCodeServer.var"#66#71"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams})()        
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:179
> [18] hideprompt(f::VSCodeServer.var"#66#71"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams})
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\repl.jl:38
> [19] (::VSCodeServer.var"#65#70"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams})()        
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:150
> [20] with_logstate(f::Function, logstate::Any)
> @ Base.CoreLogging .\logging.jl:515
> [21] with_logger
> @ .\logging.jl:627 [inlined]
> [22] (::VSCodeServer.var"#64#69"{VSCodeServer.ReplRunCodeRequestParams})()
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:255
> [23] #invokelatest#2
> @ Base .\essentials.jl:887 [inlined]
> [24] invokelatest(::Any)
> @ Base .\essentials.jl:884
> [25] (::VSCodeServer.var"#62#63")()
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:34
> in expression starting at c:\Users\*\Documents\GoL\CGoL.jl\scripts\AMDGPUtest.jl:4
> 
> ```

However, the core of the package still works, so I can generate ROCArrays, but I can’t use rocRAND or rocFFT, which I need.

> **Manual testing**
>
> ```julia
> using AMDGPU
> using AMDGPU.rocFFT
> AMDGPU.allowscalar(false)
> A = rand(3,3)
> # A_d = AMDGPU.rand(3,3) # errors
> A_d = ROCArray(A) # works just fine
> # fft(A_d) # gives following error
> 1-element ExceptionStack:
> LoadError: could not load library "C:\Program Files\AMD\ROCm\5.5\bin\rocfft.dll" # this file does exist on my hardware
> The specified module could not be found.
> Stacktrace:
> [1] rocfft_plan_create(plan::Base.RefValue{Ptr{AMDGPU.rocFFT.rocfft_plan_t}}, placement::AMDGPU.rocFFT.rocfft_result_placement_e, transform_type::AMDGPU.rocFFT.rocfft_transform_type_e, precision::AMDGPU.rocFFT.rocfft_precision_e, dimensions::Int64, lengths::Vector{Int64}, number_of_transforms::Int64, description::Ptr{Nothing})
> @ AMDGPU.rocFFT C:\Users\*\.julia\packages\AMDGPU\BhNdC\src\fft\librocfft.jl:77
> [2] create_plan(xtype::AMDGPU.rocFFT.rocfft_transform_type_e, xdims::Tuple{Int64, Int64}, T::Type, inplace::Bool, region::UnitRange{Int64})
> @ AMDGPU.rocFFT C:\Users\*\.julia\packages\AMDGPU\BhNdC\src\fft\util.jl:59
> [3] #3
> @ C:\Users\*\.julia\packages\AMDGPU\BhNdC\src\fft\util.jl:38 [inlined]
> [4] (::AMDGPU.var"#6#9"{AMDGPU.rocFFT.var"#3#4"{Tuple{AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Int64, Int64}, DataType, Bool, UnitRange{Int64}}}, HandleCache{Tuple{HIPContext, AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Vararg{Int64, N}} where N, Type, Bool, Any}, Tuple{Ptr{AMDGPU.rocFFT.rocfft_plan_t}, Int64}}, Tuple{HIPContext, AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Int64, Int64}, DataType, Bool, UnitRange{Int64}}})()
> @ AMDGPU C:\Users\*\.julia\packages\AMDGPU\BhNdC\src\cache.jl:29
> [5] lock(f::AMDGPU.var"#6#9"{AMDGPU.rocFFT.var"#3#4"{Tuple{AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Int64, Int64}, DataType, Bool, UnitRange{Int64}}}, HandleCache{Tuple{HIPContext, AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Vararg{Int64, N}} where N, Type, Bool, Any}, Tuple{Ptr{AMDGPU.rocFFT.rocfft_plan_t}, Int64}}, Tuple{HIPContext, AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Int64, Int64}, DataType, Bool, UnitRange{Int64}}}, l::ReentrantLock)
> @ Base .\lock.jl:229
> [6] (::AMDGPU.var"#check_cache#7"{HandleCache{Tuple{HIPContext, AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Vararg{Int64, N}} where N, Type, Bool, Any}, Tuple{Ptr{AMDGPU.rocFFT.rocfft_plan_t}, Int64}}, Tuple{HIPContext, AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Int64, Int64}, DataType, Bool, UnitRange{Int64}}})(f::AMDGPU.rocFFT.var"#3#4"{Tuple{AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Int64, Int64}, DataType, Bool, UnitRange{Int64}}})
> @ AMDGPU C:\Users\*\.julia\packages\AMDGPU\BhNdC\src\cache.jl:27
> [7] pop!(f::Function, cache::HandleCache{Tuple{HIPContext, AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Vararg{Int64, N}} where N, Type, 
> Bool, Any}, Tuple{Ptr{AMDGPU.rocFFT.rocfft_plan_t}, Int64}}, key::Tuple{HIPContext, AMDGPU.rocFFT.rocfft_transform_type_e, Tuple{Int64, Int64}, DataType, Bool, UnitRange{Int64}})
> @ AMDGPU C:\Users\*\.julia\packages\AMDGPU\BhNdC\src\cache.jl:48
> [8] get_plan(::AMDGPU.rocFFT.rocfft_transform_type_e, ::Vararg{Any})
> @ AMDGPU.rocFFT C:\Users\*\.julia\packages\AMDGPU\BhNdC\src\fft\util.jl:37
> [9] plan_fft
> @ AMDGPU.rocFFT C:\Users\*\.julia\packages\AMDGPU\BhNdC\src\fft\fft.jl:110 [inlined]
> [10] fft
> @ AbstractFFTs C:\Users\*\.julia\packages\AbstractFFTs\4iQz5\src\definitions.jl:67 [inlined]
> [11] fft
> @ AbstractFFTs C:\Users\*\.julia\packages\AbstractFFTs\4iQz5\src\definitions.jl:214 [inlined]
> [12] fft(x::ROCArray{Float64, 2, AMDGPU.Runtime.Mem.HIPBuffer})
> @ AbstractFFTs C:\Users\*\.julia\packages\AbstractFFTs\4iQz5\src\definitions.jl:66
> [13] top-level scope
> @ c:\Users\*\Documents\GoL\CGoL.jl\scripts\AMDGPUtest.jl:7
> [14] eval
> @ Core .\boot.jl:385 [inlined]
> [15] include_string(mapexpr::typeof(identity), mod::Module, code::String, filename::String)
> @ Base .\loading.jl:2070
> [16] include_string(m::Module, txt::String, fname::String)
> @ Base .\loading.jl:2080
> [17] invokelatest(::Any, ::Any, ::Vararg{Any}; kwargs::@Kwargs{})
> @ Base .\essentials.jl:887
> [18] invokelatest(::Any, ::Any, ::Vararg{Any})
> @ Base .\essentials.jl:884
> [19] inlineeval(m::Module, code::String, code_line::Int64, code_column::Int64, file::String; softscope::Bool)
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:263
> [20] (::VSCodeServer.var"#67#72"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams})()        
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:181
> [21] withpath(f::VSCodeServer.var"#67#72"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams}, 
> path::String)
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\repl.jl:274
> [22] (::VSCodeServer.var"#66#71"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams})()        
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:179
> [23] hideprompt(f::VSCodeServer.var"#66#71"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams})
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\repl.jl:38
> [24] (::VSCodeServer.var"#65#70"{Bool, Bool, Bool, Module, String, Int64, Int64, String, VSCodeServer.ReplRunCodeRequestParams})()        
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:150
> [25] with_logstate(f::Function, logstate::Any)
> @ Base.CoreLogging .\logging.jl:515
> [26] with_logger
> @ .\logging.jl:627 [inlined]
> [27] (::VSCodeServer.var"#64#69"{VSCodeServer.ReplRunCodeRequestParams})()
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:255
> [28] #invokelatest#2
> @ Base .\essentials.jl:887 [inlined]
> [29] invokelatest(::Any)
> @ Base .\essentials.jl:884
> [30] (::VSCodeServer.var"#62#63")()
> @ VSCodeServer c:\Users\*\.vscode\extensions\julialang.language-julia-1.66.2\scripts\packages\VSCodeServer\src\eval.jl:34
> in expression starting at c:\Users\*\Documents\GoL\CGoL.jl\scripts\AMDGPUtest.jl:7
> 
> ```

I’m aware that some packages like rocFFT may still be broken with Windows 10 ([Some AMDGPU tests failing on Radeon VII under Windows](https://discourse.julialang.org/t/some-amdgpu-tests-failing-on-radeon-vii-under-windows/108074)) but it seems like here Julia can not even begin to access them. Please let me know if I need to provide any more information. Any help with this would be greatly appreciated!

---

<div class="post-metadata">

**Author:** ![Palli](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/palli/32/3380_2.png) [@Palli](https://discourse.julialang.org/u/Palli)\
**Post date:** [February 2, 2024, 6:37pm UTC](https://discourse.julialang.org/t/unable-to-test-or-use-hip-libraries-on-windows-10/109639/2 "2024-02-02T18:37:35Z")

</div>

> [@mp59](#):
>
> … `library/TensileLibrary.dat: No such file or directory`

It’s not Julia-specific at least:

> <https://github.com/ROCm/rocBLAS/issues/1339>
>
> \### Describe the bug
> Basically getting some form of this error, either \`\`\`rocBL…AS error: Cannot read /opt/rocm-5.4.0/lib/rocblas/library/TensileLibrary.dat: Illegal seek\`\`\` or \`\`\`Cannot read TensileLibrary.dat: No such file or directory\`\`\`
> 
> \### To Reproduce
> rocblas-dev (= 2.46.0.50400-72~22.04)
> 
> Steps to reproduce the behavior:
> 1. Basically followed the guide on AMD Docs to install ROCm along with almost all usecases because I kept having errors with packages missing.
> 2. Built wheel for onnxruntime using docker
> 3. Trying to run the python application (roop) and getting this error
> 
> \### Expected behavior
> No error?
> 
> \### Log-files
> 
> \`\`\`rocBLAS error: Cannot read /opt/rocm-5.4.0/lib/rocblas/library/TensileLibrary.dat: Illegal seek
> Aborted (core dumped)
> (roop) hobi@hobi:~/roop$ ~python run.py --execution-provider rocm --execution-threads
> Command '~python' not found, did you mean:
> command 'bpython' from deb bpython (0.22.1-2)
> command 'xpython' from deb xpython (0.12.5-1build1)
> Try: sudo apt install \<deb name\>
> (roop) hobi@hobi:~/roop$ python run.py --execution-provider rocm --execution-threads
> usage: run.py \[-h\] \[-s SOURCE\_PATHS\] \[-t TARGET\_PATHS\] \[-o OUTPUT\_PATH\] \[--frame-processor {face\_swapper,face\_enhancer} \[{face\_swapper,face\_enhancer} ...\]\] \[--keep-fps\] \[--keep-audio\] \[--keep-frames\] \[--keep-filenames\] \[--many-faces\]
> \[--video-encoder {libx264,libx265,libvpx-vp9}\] \[--video-quality \[0-51\]\] \[--max-memory MAX\_MEMORY\] \[--execution-provider {rocm,cpu} \[{rocm,cpu} ...\]\] \[--execution-threads EXECUTION\_THREADS\] \[-v\]
> run.py: error: argument --execution-threads: expected one argument
> (roop) hobi@hobi:~/roop$ python run.py --execution-provider rocm --execution-threads 2
> \[ROOP.CORE\] Creating temp resources...
> \[ROOP.CORE\] Extracting frames...
> Applied providers: \['ROCMExecutionProvider', 'CPUExecutionProvider'\], with options: {'CPUExecutionProvider': {}, 'ROCMExecutionProvider': {'tunable\_op\_tuning\_enable': '0', 'do\_copy\_in\_default\_stream': '1', 'miopen\_conv\_exhaustive\_search': '0', 'arena\_extend\_strategy': 'kNextPowerOfTwo', 'gpu\_external\_empty\_cache': '0', 'gpu\_external\_free': '0', 'miopen\_conv\_use\_max\_workspace': '1', 'gpu\_mem\_limit': '18446744073709551615', 'tunable\_op\_enable': '0', 'gpu\_external\_alloc': '0', 'device\_id': '0'}}
> find model: /home/hobi/.insightface/models/buffalo\_l/1k3d68.onnx landmark\_3d\_68 \['None', 3, 192, 192\] 0.0 1.0
> Applied providers: \['ROCMExecutionProvider', 'CPUExecutionProvider'\], with options: {'CPUExecutionProvider': {}, 'ROCMExecutionProvider': {'tunable\_op\_tuning\_enable': '0', 'do\_copy\_in\_default\_stream': '1', 'miopen\_conv\_exhaustive\_search': '0', 'arena\_extend\_strategy': 'kNextPowerOfTwo', 'gpu\_external\_empty\_cache': '0', 'gpu\_external\_free': '0', 'miopen\_conv\_use\_max\_workspace': '1', 'gpu\_mem\_limit': '18446744073709551615', 'tunable\_op\_enable': '0', 'gpu\_external\_alloc': '0', 'device\_id': '0'}}
> find model: /home/hobi/.insightface/models/buffalo\_l/2d106det.onnx landmark\_2d\_106 \['None', 3, 192, 192\] 0.0 1.0
> Applied providers: \['ROCMExecutionProvider', 'CPUExecutionProvider'\], with options: {'CPUExecutionProvider': {}, 'ROCMExecutionProvider': {'tunable\_op\_tuning\_enable': '0', 'do\_copy\_in\_default\_stream': '1', 'miopen\_conv\_exhaustive\_search': '0', 'arena\_extend\_strategy': 'kNextPowerOfTwo', 'gpu\_external\_empty\_cache': '0', 'gpu\_external\_free': '0', 'miopen\_conv\_use\_max\_workspace': '1', 'gpu\_mem\_limit': '18446744073709551615', 'tunable\_op\_enable': '0', 'gpu\_external\_alloc': '0', 'device\_id': '0'}}
> find model: /home/hobi/.insightface/models/buffalo\_l/det\_10g.onnx detection \[1, 3, '?', '?'\] 127.5 128.0
> Applied providers: \['ROCMExecutionProvider', 'CPUExecutionProvider'\], with options: {'CPUExecutionProvider': {}, 'ROCMExecutionProvider': {'tunable\_op\_tuning\_enable': '0', 'do\_copy\_in\_default\_stream': '1', 'miopen\_conv\_exhaustive\_search': '0', 'arena\_extend\_strategy': 'kNextPowerOfTwo', 'gpu\_external\_empty\_cache': '0', 'gpu\_external\_free': '0', 'miopen\_conv\_use\_max\_workspace': '1', 'gpu\_mem\_limit': '18446744073709551615', 'tunable\_op\_enable': '0', 'gpu\_external\_alloc': '0', 'device\_id': '0'}}
> find model: /home/hobi/.insightface/models/buffalo\_l/genderage.onnx genderage \['None', 3, 96, 96\] 0.0 1.0
> Applied providers: \['ROCMExecutionProvider', 'CPUExecutionProvider'\], with options: {'CPUExecutionProvider': {}, 'ROCMExecutionProvider': {'tunable\_op\_tuning\_enable': '0', 'do\_copy\_in\_default\_stream': '1', 'miopen\_conv\_exhaustive\_search': '0', 'arena\_extend\_strategy': 'kNextPowerOfTwo', 'gpu\_external\_empty\_cache': '0', 'gpu\_external\_free': '0', 'miopen\_conv\_use\_max\_workspace': '1', 'gpu\_mem\_limit': '18446744073709551615', 'tunable\_op\_enable': '0', 'gpu\_external\_alloc': '0', 'device\_id': '0'}}
> find model: /home/hobi/.insightface/models/buffalo\_l/w600k\_r50.onnx recognition \['None', 3, 112, 112\] 127.5 127.5
> set det-size: (640, 640)
> 
> rocBLAS error: Cannot read /opt/rocm-5.4.0/lib/rocblas/library/TensileLibrary.dat: No such file or directory
> Aborted (core dumped)
> \`\`\`
> 
> \### Environment
> 
> | Hardware | description |
> |-----|-----|
> | CPU | Rzyen 5 5600 |
> | GPU | Radeon 6700XT |
> 
> | Software | version |
> |-----|-----|
> | rocm-core | v5.4.0.50400-72~22.04 |
> | rocblas | v2.46.0.50400-72~22.04 |
> 
> Make sure that ROCm is correctly installed and to capture detailed environment information run the following command:
> \`\`\`
> printf '=== environment\\n' \> environment.txt &&
> printf '\\n\\n=== date\\n' \>\> environment.txt && date \>\> environment.txt &&
> printf '\\n\\n=== Linux Kernel\\n' \>\> environment.txt && uname -a \>\> environment.txt &&
> printf '\\n\\n=== rocm-smi' \>\> environment.txt && rocm-smi \>\> environment.txt &&
> printf '\\n\\n' \>\> environment.txt && hipconfig \>\> environment.txt &&
> printf '\\n\\n=== rocminfo\\n' \>\> environment.txt && rocminfo \>\> environment.txt &&
> printf '\\n\\n=== lspci VGA\\n' \>\> environment.txt && lspci | grep -i vga \>\> environment.txt
> \`\`\`
> 
> Getting this error: \`\`\`No LSB modules are available.
> \`\`\` 
> 
> 
> \### Additional context
> I am super new to machine learning and I am having a nightmare time of making things work with ROCm. Pretty much at the end of the rope here guys. Any help would be appreciated. Thank you.

I didn’t read carefully, maybe there’s a workaround, or unsupported hardware, or driver not installed? CUDA.jl is well-supported, Julia for AMD GPUS (HIP/ROCm a bit newer, but at least confirm it works/correctly installed first somehow without Julia).

[I don’t have your hardware, nor Windows, to test anything, just trying to explain/read the log.]

---

<div class="post-metadata">

**Author:** ![pxl-th](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/pxl-th/32/31939_2.png) [@pxl-th](https://discourse.julialang.org/u/pxl-th)\
**Post date:** [February 2, 2024, 7:05pm UTC](https://discourse.julialang.org/t/unable-to-test-or-use-hip-libraries-on-windows-10/109639/3 "2024-02-02T19:05:00Z")

</div>

I suspect that your GPU is not supported on Windows.  
See the list of supported devices:

> **[System requirements (Windows) — HIP SDK installation Windows](https://rocm.docs.amd.com/projects/install-on-windows/en/latest/reference/system-requirements.html)**
>
> Windows GPU and OS support

On Linux you could define env variable `HSA_OVERRIDE_GFX_VERSION=9.0.0` before launching Julia and that should make it work.  
On Windows however, there’s no such option…

---

<div class="post-metadata">

**Author:** ![mp59](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/mp59/32/206717_2.png) [@mp59](https://discourse.julialang.org/u/mp59)\
**Post date:** [February 2, 2024, 7:43pm UTC](https://discourse.julialang.org/t/unable-to-test-or-use-hip-libraries-on-windows-10/109639/4 "2024-02-02T19:43:13Z")

</div>

Well, that’s unfortunate. I even tried changing the environment variable globally and there was no luck.  
Would it be worth trying to use Julia on WSL instead? I’m doing this for a personal project and I don’t have the option of switching hardware.  
Thank you for your replies.

---

<div class="post-metadata">

**Author:** ![pxl-th](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/pxl-th/32/31939_2.png) [@pxl-th](https://discourse.julialang.org/u/pxl-th)\
**Post date:** [February 2, 2024, 7:52pm UTC](https://discourse.julialang.org/t/unable-to-test-or-use-hip-libraries-on-windows-10/109639/5 "2024-02-02T19:52:39Z")

</div>

I’m not sure ROCm works under WSL… So I’d suggest trying with a proper Linux installation (dual-boot or not)