- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
I'm using Intel oneAPI BaseKit 2022.1.2 (w_BaseKit_p_2022.1.2.154_offline.exe), VisualStudio 2022 17.0.6 LTS, CMake 3.22.2, Ninja 1.10.2, VSCode 1.64.2, Windows 10 Pro 21H1 x64 English.
OneAPI & VisualStudio was fully reinstalled today.
I'm using a basic example loop unrolling.
Using VisualStudio sln works correctly & exe file starts to execute.
Using CMakeLists.txt (file attached), executable file doesn't work correctly & prints error -46 (CL_INVALID_KERNEL_NAME). CMakeLists file content is based on that link & default provided file. An example of error was screened below.
Build log was attached too:
- from VSCode with Intel oneAPI plugins pack
- from clear CMD with manually called setvars.bat file
I need to use CMakeLists.txt for cross-compile code building.
That error looks like a linking error. How cal I fix it?
With regards,
Alex
P.S. Project for reproducing an error was attached too.
コピーされたリンク
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hi,
Thanks for reaching out to us.
We were able to reproduce your issue at our end using the Intel oneAPI Base Toolkit 2022.1.0 on the windows command prompt as seen in the attached screenshot(error.png) below.
We are working on your issue and we will get back to you soon.
Thanks & Regards,
Santosh
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hello,
Thank you for contacting Intel Support forum.
We have looked into your case.
This looks like a bug in usage of cmake with Intel oneAPI DPC++ compiler.
I have now forwarded this case to Intel development team for investigation.
I will update this forum once there is a feedback from Intel development team.
Please note that that this project may be successfully used from Visual Studio.
Moreover, you may compile the sources manually from the command line,
but without usage of cmake (I saw you statement that you need cmake - but maybe this would be suitable for you work on some other project, etc.):
>icx -fsycl /EHsc ..\src\loop-unroll.cpp
>loop-unroll.exe
Input array size: 33554432
Running on device: Intel(R) Graphics [0x3ea0]
Unroll factor: 1 Kernel time: 6685.64 ms
Throughput for kernel with unroll factor 1: 0.005 GFlops
Unroll factor: 2 Kernel time: 4171.452 ms
Throughput for kernel with unroll factor 2: 0.008 GFlops
Unroll factor: 4 Kernel time: 2332.509 ms
Throughput for kernel with unroll factor 4: 0.014 GFlops
Unroll factor: 8 Kernel time: 1441.188 ms
Throughput for kernel with unroll factor 8: 0.023 GFlops
Unroll factor: 16 Kernel time: 1252.165 ms
Throughput for kernel with unroll factor 16: 0.027 GFlops
PASSED: The results are correct.
Hope that this helps.
Regards,
Khalik.
