clang++: error: SYCL target is invalid: 'nvptx64-nvidia-cuda-sycldevice'

[mmmm1998]

Dear All,

I have a some trouble with compiling DPC++ with CUDA support and not sure, how it should be done, because commands from documentation don't work. When I run

$ dpcpp -fsycl -fsycl-targets=nvptx64-nvidia-cuda-sycldevice main.cpp vector-add.cpp -o vector-add
clang++: error: SYCL target is invalid: 'nvptx64-nvidia-cuda-sycldevice'

I have gotten this error. But, whel, I have CUDA installation.

$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2017 NVIDIA Corporation
Built on Fri_Nov__3_21:07:56_CDT_2017
Cuda compilation tools, release 9.1, V9.1.85
$ nvidia-smi
Wed Jul 29 01:22:38 2020
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 418.87.00    Driver Version: 418.87.00    CUDA Version: 10.1     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 1080    On   | 00000000:01:00.0  On |                  N/A |
| 27%   36C    P8     8W / 180W |    172MiB /  8116MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|    0      2197      G   /usr/lib/xorg/Xorg                           169MiB |
+-----------------------------------------------------------------------------+

Also I have latest stable dpcpp version from Intel repo.

$ dpcpp --version
Intel(R) oneAPI DPC++ Compiler 2021.1-beta07 (2020.5.0.0604)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /opt/intel/inteloneapi/compiler/latest/linux/bin

And, if I am not mistaken, CUDA support already in dpcpp master

Also, my system information:

$ lsb_release -a
LSB Version:    core-9.20170808ubuntu1-noarch:printing-9.20170808ubuntu1-noarch:security-9.20170808ubuntu1-noarch
Distributor ID: Ubuntu
Description:    Ubuntu 18.04.4 LTS
Release:        18.04
Codename:       bionic

Also, can I print all valid sycl devices from compiler? It would be helpful, I think.

[bader]

The compiler version distributed with Intel® oneAPI toolkit does support CUDA.
Please, follow Get Started Guide instructions to build the compiler toolchain with CUDA support from sources.
Particularly pay attention to this section: https://intel.github.io/llvm-docs/GetStartedGuide.html#build-dpc-toolchain-with-support-for-nvidia-cuda

[mmmm1998]

Thanks, bader, I have check the section, and see, that requested CUDA version is 10.2, but I have only 10.1.