GPUInspector.jl icon indicating copy to clipboard operation
GPUInspector.jl copied to clipboard

Bump versions of dependencies

Open blegat opened this issue 10 months ago • 2 comments

Maybe the CompatHelper workflow could be reenabled: https://github.com/pc2/GPUInspector.jl/actions/workflows/CompatHelper.yml

blegat avatar Mar 10 '25 08:03 blegat

Thanks. Have you checked that things work with these new versions?

carstenbauer avatar Mar 10 '25 09:03 carstenbauer

I get a few errors

toggle_tensorcoremath: Error During Test at /auto/home/users/b/l/blegat/.julia/dev/GPUInspector/test/utility_tests.jl:70
  Test threw exception
  Expression: isnothing(CUDAExt.toggle_tensorcoremath(true; verbose = false))
  type Nothing has no field toggle_tensorcoremath
  Stacktrace:
   [1] getproperty(x::Nothing, f::Symbol)
     @ Base ./Base.jl:49
   [2] macro expansion
     @ /auto/home/users/b/l/blegat/.julia/dev/GPUInspector/test/utility_tests.jl:70 [inlined]

and

peakflops_gpu_matmul / scaling: Error During Test at /auto/home/users/b/l/blegat/.julia/dev/GPUInspector/test/peakflops_tests.jl:18
  Test threw exception
  Expression: typeof(CUDAExt.peakflops_gpu_matmul(; verbose = false)) == Float64
  type Nothing has no field peakflops_gpu_matmul
  Stacktrace:
   [1] getproperty(x::Nothing, f::Symbol)
     @ Base ./Base.jl:49
   [2] macro expansion
     @ /auto/home/users/b/l/blegat/.julia/juliaup/julia-1.11.3+0.x64.linux.gnu/share/julia/stdlib/v1.11/Test/src/Test.jl:676 [inlined]
   [3] top-level scope
     @ /auto/home/users/b/l/blegat/.julia/dev/GPUInspector/test/peakflops_tests.jl:515

and

bidirectional: Error During Test at /auto/home/users/b/l/blegat/.julia/dev/GPUInspector/test/bandwidth_tests.jl:24
  Got exception outside of a @test
  ArgumentError: cannot take the GPU address of inaccessible device memory.
  
  You are trying to use memory from GPU 0 on GPU 1.
  P2P access between these devices is not possible; either switch to GPU 0
  by calling `CUDA.device!(0)`, or copy the data to an array allocated on device 1.
  Stacktrace:
    [1] convert(::Type{CuPtr{Float32}}, managed::CUDA.Managed{CUDA.DeviceMemory})
      @ CUDA ~/.julia/packages/CUDA/1kIOw/src/memory.jl:552
    [2] unsafe_convert(typ::Type{CuPtr{Float32}}, x::CuArray{Float32, 1, CUDA.DeviceMemory})
      @ CUDA ~/.julia/packages/CUDA/1kIOw/src/array.jl:432
    [3] #pointer#1125
      @ ~/.julia/packages/CUDA/1kIOw/src/array.jl:390 [inlined]

and

UndefVarError: `unsafe_nvmlDeviceGetTemperature` not defined in `CUDA.NVML`

and

save / load: Error During Test at /auto/home/users/b/l/blegat/.julia/dev/GPUInspector/test/stresstest_tests.jl:54
  Got exception outside of a @test
  type Nothing has no field _device2string
  Stacktrace:
    [1] getproperty(x::Nothing, f::Symbol)
      @ Base ./Base.jl:49
    [2] macro expansion
      @ /auto/home/users/b/l/blegat/.julia/dev/GPUInspector/test/stresstest_tests.jl:60 [inlined]

blegat avatar Mar 10 '25 14:03 blegat