How to install intel-compute-runtime

how to install intel-compute-runtime, it doesnt seem like there is a package for it, and in the this github repo Release 24.17.29377.6 · intel/compute-runtime · GitHub there are no packages for clear linux and clear linux doesnt have apt or dpkg
how can i do this

Dont expect an answer…

I opened a Github request Nov 15, 2023, closed it mars 2024 due to no response.

Then i opened a new ticket

Not a single response.

Sad but ClearLinux does not care about IntelOneApi fans.

In my humble view… Intel ClearLinux management people should take cues from a smart guy and put some more resources into making ClearLinux the ultimate OneApi Desktop developer workstation.

I’m no developer, but I spent a few afternoons trying to figure this out
Here is the repo for the graphics compiler GitHub - clearlinux-pkgs/intel-graphics-compiler
I tried building it through CLs repo packages and ended up failing due to llvm14 and weirdness with how the dependencies are being called.

If this can be solved, can try and resurrect this GitHub - clearlinux-pkgs/intel-compute-runtime since the requirements for it are intel-gmmlib (available) and the graphics compiler from earlier (currently broken)

This is unofficial.

I installed the Intel Compute Runtime using the Python pip command. The files are placed under ~/.local/, in your home folder.

$ pip install --user intel-cmplr-lib-rt --upgrade
$ pip install --user intel-opencl-rt --upgrade

sudo mkdir -p /etc/OpenCL/vendors

echo "/home/$USER/.local/lib/libintelocl.so" |\
     sudo tee /etc/OpenCL/vendors/intel-cpu.icd

Miniconda works too. Afterwards, I copy the nvidia.icd file to the MiniConda mandel environment.

conda activate mandel
conda install intel-cmplr-lib-rt
conda install intel-opencl-rt

cp /etc/OpenCL/vendors/nvidia.icd ~/miniconda3/envs/mandel/etc/OpenCL/vendors/

Using OpenCL, the Python Mandelbrot demo runs on the GPU or CPU.

$ ./mandel_ocl.py 
pygame 2.5.2 (SDL 2.28.2, Python 3.12.4)
Hello from the pygame community. https://www.pygame.org/contribute.html
NOTE: Choose a device with double-precision capabilities.
Choose platform:
[0] <pyopencl.Platform 'NVIDIA CUDA' at 0x342fa900>
[1] <pyopencl.Platform 'Intel(R) OpenCL' at 0x342cba68>
[2] <pyopencl.Platform 'Intel(R) FPGA Emulation Platform for OpenCL(TM)' at 0x342d2038>
Choice [0]: