argmax(::CuArray) returns nothing with NaN-values

[simeonschaub]

For regular arrays, argmax returns a valid index if NaNs are present:

julia> argmax([NaN32])
1

while for CuArray, this returns nothing:

julia> argmax(gpu([NaN32])) === nothing
true

Julia and CUDA versions:

julia> versioninfo()
Julia Version 1.5.3
Commit 788b2c77c1 (2020-11-09 13:37 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
  CPU: AMD Ryzen 9 3900X 12-Core Processor
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-9.0.1 (ORCJIT, znver2)

julia> CUDA.versioninfo()
CUDA toolkit 11.0.3, artifact installation
CUDA driver 11.0.0
/usr/bin/nvidia-modprobe: unrecognized option: "-f"

ERROR: Invalid commandline, please run `/usr/bin/nvidia-modprobe --help` for usage information.

/usr/bin/nvidia-modprobe: unrecognized option: "-f"

ERROR: Invalid commandline, please run `/usr/bin/nvidia-modprobe --help` for usage information.

NVIDIA driver 450.80.2

Libraries: 
- CUBLAS: 11.2.0
- CURAND: 10.2.1
- CUFFT: 10.2.1
- CUSOLVER: 10.6.0
- CUSPARSE: 11.1.1
- CUPTI: 13.0.0
- NVML: 11.0.0+450.80.2
- CUDNN: 8.0.4 (for CUDA 11.0.0)
- CUTENSOR: 1.2.1 (for CUDA 11.0.0)

Toolchain:
- Julia: 1.5.3
- LLVM: 9.0.1
- PTX ISA support: 3.2, 4.0, 4.1, 4.2, 4.3, 5.0, 6.0, 6.1, 6.3, 6.4
- Device support: sm_35, sm_37, sm_50, sm_52, sm_53, sm_60, sm_61, sm_62, sm_70, sm_72, sm_75

1 device:
  0: GeForce GTX 1080 Ti (sm_61, 10.355 GiB / 10.913 GiB available)

[maleadt]

The reason:

julia> findmax([NaN32] |> cu)
(NaN32, nothing)

julia> findmax([NaN32])
(NaN32, 1)

Fixed by the alternative implementation in #320 (comment).