Git Product home page Git Product logo

Comments (4)

maleadt avatar maleadt commented on July 19, 2024

Could you run with JULIA_DEBUG=CUDAapi? Where exactly is cudnn located?

from cudaapi.jl.

IanButterworth avatar IanButterworth commented on July 19, 2024

cudnn is located here:

/home/ibutter/anaconda3/pkgs/cudnn-7.3.1-cuda10.0_0/lib/libcudnn.so
/home/ibutter/anaconda3/pkgs/cudnn-7.3.1-cuda10.0_0/lib/libcudnn.so.7.3.1
/home/ibutter/anaconda3/pkgs/cudnn-7.3.1-cuda10.0_0/lib/libcudnn.so.7
/home/ibutter/anaconda3/lib/libcudnn.so
/home/ibutter/anaconda3/lib/libcudnn.so.7
/home/ibutter/anaconda3/lib/libcudnn.so.7.3.1

Running the above with JULIA_DEBUG=CUDAapi

Knet.gpuCount() = 8
Knet.gpu() = 3
Knet.tk = ["/usr/local/cuda-10.0"]
Knet.libknet8 = "/home/ibutter/.julia/packages/Knet/T1oum/deps/libknet8"
Knet.cudartfound = true
Knet.cudaRuntimeVersion = 10000
Knet.cudaDriverVersion = 10000
Knet.cudaGetDeviceCount() = 8
Knet.cudaGetDevice() = 3
Knet.cudaMemGetInfo() = (11604393984, 12621381632)
Knet.cudaDeviceSynchronize() = nothing
Knet.nvmlfound = true
Knet.nvmlDriverVersion = "410.93"
Knet.nvmlVersion = "10.410.93"
Knet.nvmlDeviceGetMemoryInfo() = (12621381632, 11604393984, 1016987648)
Knet.cublashandle() = Ptr{Nothing} @0x000000000fc025d0
Knet.cublasVersion = 10000
┌ Debug: Request to look for library cudnn
│   locations =
│    1-element Array{String,1}:
│     "/usr/local/cuda-10.0"
└ @ CUDAapi ~/.julia/packages/CUDAapi/hv5e7/src/CUDAapi.jl:8
┌ Debug: Looking for library libcudnn.so, libcudnn.so.7, libcudnn.so.7.4, libcudnn.so.7.3, libcudnn.so.7.1, libcudnn.so.7.0, libcudnn.so.6, libcudnn.so.6.0, libcudnn.so.5, libcudnn.so.5.1, libcudnn.so.5.0, libcudnn.so.4, libcudnn.so.4.0, libcudnn.so.3, libcudnn.so.3.0, libcudnn.so.2, libcudnn.so.2.0, libcudnn.so.1, libcudnn.so.1.0
│   locations =
│    3-element Array{String,1}:
│     "/usr/local/cuda-10.0"      
│     "/usr/local/cuda-10.0/lib"  
│     "/usr/local/cuda-10.0/lib64"
└ @ CUDAapi ~/.julia/packages/CUDAapi/hv5e7/src/CUDAapi.jl:8
gpu: Error During Test at /home/ibutter/.julia/packages/Knet/T1oum/test/gpu.jl:3
  Got exception outside of a @test
  Cannot find cudnn
  Stacktrace:
   [1] error(::String) at ./error.jl:33
   [2] cudnnCreate() at /home/ibutter/.julia/packages/Knet/T1oum/src/gpu.jl:238
   [3] cudnnhandle(::Int64) at /home/ibutter/.julia/packages/Knet/T1oum/src/gpu.jl:188
   [4] cudnnhandle() at /home/ibutter/.julia/packages/Knet/T1oum/src/gpu.jl:182
   [5] macro expansion at ./show.jl:555 [inlined]
   [6] macro expansion at /home/ibutter/.julia/packages/Knet/T1oum/test/gpu.jl:28 [inlined]
   [7] macro expansion at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.0/Test/src/Test.jl:1083 [inlined]
   [8] top-level scope at /home/ibutter/.julia/packages/Knet/T1oum/test/gpu.jl:5
   [9] include at ./boot.jl:317 [inlined]
   [10] include_relative(::Module, ::String) at ./loading.jl:1044
   [11] include(::Module, ::String) at ./sysimg.jl:29
   [12] include(::String) at ./client.jl:392
   [13] top-level scope at none:0
   [14] eval(::Module, ::Any) at ./boot.jl:319
   [15] eval_user_input(::Any, ::REPL.REPLBackend) at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.0/REPL/src/REPL.jl:85
   [16] macro expansion at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.0/REPL/src/REPL.jl:117 [inlined]
   [17] (::getfield(REPL, Symbol("##28#29")){REPL.REPLBackend})() at ./task.jl:259
Test Summary: | Error  Total
gpu           |     1      1
ERROR: LoadError: Some tests did not pass: 0 passed, 0 failed, 1 errored, 0 broken.
in expression starting at /home/ibutter/.julia/packages/Knet/T1oum/test/gpu.jl:3

from cudaapi.jl.

maleadt avatar maleadt commented on July 19, 2024

Yeah, that's a very custom path. You shouldn't expect CUDAapi to pick that up just like that.
We could add support for env vars to customize each library's location, but for now just set LD_LIBRARY_PATH to /home/ibutter/anaconda3/lib.

from cudaapi.jl.

maleadt avatar maleadt commented on July 19, 2024

Going to close this, since LD_LIBRARY_PATH is a perfectly good way to influence find_cuda_library (which falls back on Libdl.find_library).

from cudaapi.jl.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.