CUBLAS initialization

[erathorn]

Thank you for creating this amazing library.

However, I encounter a problem with CUBLAS. Using Matrix multiplication does not work for me. I get the following problem:

using CuArrays

CuArrays.seed!(42)

A = CuArrays.rand(5,5)

A*A

ERROR: CUBLASError: the library was not initialized (code 1, CUBLAS_STATUS_NOT_INITIALIZED)
Stacktrace:
 [1] throw_api_error(::CuArrays.CUBLAS.cublasStatus_t) at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/error.jl:52
 [2] macro expansion at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/error.jl:65 [inlined]
 [3] cublasCreate_v2 at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/libcublas.jl:9 [inlined]
 [4] cublasCreate_v2 at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/wrappers.jl:8 [inlined]
 [5] (::CuArrays.CUBLAS.var"#354#356"{CUDAdrv.CuContext})() at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/CUBLAS.jl:40
 [6] get! at ./abstractdict.jl:661 [inlined]
 [7] handle() at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/CUBLAS.jl:39
 [8] top-level scope at REPL[3]:1

I tried to search for the error, and I ended up here:

julia> using CuArrays

julia> CuArrays.CUBLAS.handle()
ERROR: CUBLASError: the library was not initialized (code 1, CUBLAS_STATUS_NOT_INITIALIZED)
Stacktrace:
 [1] throw_api_error(::CuArrays.CUBLAS.cublasStatus_t) at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/error.jl:52
 [2] macro expansion at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/error.jl:65 [inlined]
 [3] cublasCreate_v2 at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/libcublas.jl:9 [inlined]
 [4] cublasCreate_v2 at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/wrappers.jl:8 [inlined]
 [5] (::CuArrays.CUBLAS.var"#354#356"{CUDAdrv.CuContext})() at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/CUBLAS.jl:40
 [6] get! at ./abstractdict.jl:661 [inlined]
 [7] handle() at /home/jo/.julia/packages/CuArrays/1njKF/src/blas/CUBLAS.jl:39
 [8] top-level scope at REPL[3]:1

Here is all the information about the julia version and the packages:

Julia Version 1.3.1
Commit 2d5741174c (2019-12-30 21:36 UTC)
Platform Info:
  OS: Linux (x86_64-linux-gnu)
  CPU: AMD Ryzen 9 3900X 12-Core Processor
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-6.0.1 (ORCJIT, znver1)
Environment:
  JULIA_NUM_THREADS = 4

nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2019 NVIDIA Corporation
Built on Sun_Jul_28_19:07:16_PDT_2019
Cuda compilation tools, release 10.1, V10.1.243

  [b19378d9] ArrayFire v1.0.6
  [c52e3926] Atom v0.12.8
  [6e4b80f9] BenchmarkTools v0.5.0
  [336ed68f] CSV v0.5.22
  [3895d2a7] CUDAapi v2.1.0
  [c5f51814] CUDAdrv v5.0.1
  [be33ccc6] CUDAnative v2.8.1
  [49dc2e85] Calculus v0.5.1
  [a81c6b42] Compose v0.8.0
  [3a865a2d] CuArrays v1.7.0
  [a93c6f00] DataFrames v0.20.0
  [31c24e10] Distributions v0.22.3
  [587475ba] Flux v0.10.1
  [f6369f11] ForwardDiff v0.10.9
  [0c68f7d7] GPUArrays v2.0.1
  [c91e804a] Gadfly v1.0.1
  [e5e0dc1b] Juno v0.7.2
  [093fc24a] LightGraphs v1.3.0
  [2774e3e8] NLsolve v4.3.0
  [90014a1f] PDMats v0.9.11
  [189a3867] Reexport v0.2.0
  [992d4aef] Showoff v0.3.1
  [276daf66] SpecialFunctions v0.10.0
  [2913bbd2] StatsBase v0.32.0
  [4c63d2b9] StatsFuns v0.9.4
  [e88e6eb3] Zygote v0.4.15
  [8ba89e20] Distributed
  [37e2e46d] LinearAlgebra
  [d6f4376e] Markdown
  [9a3f8284] Random
  [10745b16] Statistics

[erathorn]

The root of this issue was a broken CUDA installation. I updated reinstalled Cuda and solved the problem. Sorry for the noise.