Error: pgnvd-Error-CUDADIR value is not a directory:

Dear community,
I just installed PGI on my Windows workstation. My workstation has a Quadro P2000 card.
When I type: pgaccelinfo

I get the following output:

CUDA Driver Version: 9010

Device Number: 0
Device Name: Quadro P2000
Device Revision Number: 6.1
Global Memory Size: 5368709120
Number of Multiprocessors: 8
Concurrent Copy and Execution: Yes
Total Constant Memory: 65536
Total Shared Memory per Block: 49152
Registers per Block: 65536
Warp Size: 32
Maximum Threads per Block: 1024
Maximum Block Dimensions: 1024, 1024, 64
Maximum Grid Dimensions: 2147483647 x 65535 x 65535
Maximum Memory Pitch: 2147483647B
Texture Alignment: 512B
Clock Rate: 1480 MHz
Execution Timeout: Yes
Integrated Device: No
Can Map Host Memory: Yes
Compute Mode: default
Concurrent Kernels: Yes
ECC Enabled: No
Memory Clock Rate: 3504 MHz
Memory Bus Width: 160 bits
L2 Cache Size: 1310720 bytes
Max Threads Per SMP: 2048
Async Engines: 2
Unified Addressing: Yes
Managed Memory: Yes
Concurrent Managed Memory: No
PGI Default Target: -ta=tesla:cc60

So, when I try to compile the Fortran example found on the OpenACC Getting Started Guide (page 24)
I got the following error:
$ pgfortran -acc f1.f90
pgnvd-Error-CUDADIR value is not a directory: C:\PROGRA~1\PGI\win64/2019/cuda/9.1/bin
PGF90-F-0155-Compiler failed to translate accelerator region (see -Minfo messages): Device compiler exited with error status code (f1.f90: 1)
PGF90/x86-64 Windows 19.4-0: compilation aborted


Any suggestion?

Hi jmendez,

During compilation, the compiler driver will look to see what CUDA driver you have installed, in your case CUDA 9.1, and then attempt to use this version. However, we only ship the last few versions of CUDA with the compilers which with 19.4 is CUDA 9.2, CUDA 10.0, and CUDA 10.1.

You have a few options:

  1. If you have the CUDA 9.1 SDK installed elsewhere on your system, you can set the environment flag “CUDA_HOME” to point to the root directory of this installation, and the compilers will use that installation.

  2. Update your CUDA driver.

Details can be found at: OpenACC Getting Started Guide Version 19.4 for x86 and NVIDIA Processors

Hope this helps,
Mat