Getting started Problems

[lkx]

Hello, I installed the Intel oneAPI Base Toolkit and Intel oneAPI HPC Toolkit on Ubuntu18.04 system. But there are some questions during the installation of GPU and getting started of the tool.

The first one is that when I installed the GPU by following the guidance on the website "https://dgpu-docs.intel.com/installation-guides/ubuntu/ubuntu-bionic.html", but when I installed developer packages, it reminded me that the package cannot be found like the figure below. I have no idea about why it occurs and how to solve it, anyone can help me?

The second is met when configuring CPU or GPU systems.  when I input the command line "sudo sh -c "echo N> /sys/module/i915/parameters/enable_hangcheck"", it reminded me "cannot create /sys/module/i915/parameters/enable_hangcheck: Directory nonexistent" and I found the package named i915 didn't exit in the directory module.

The last one is that I want to know how to find and use the Fortran compiler.

Thanks ahead!

[RahulV_intel]

Hi,

Please provide your GPU details, oneAPI toolkits version.

In order to check the devices detected under the openCL backend, you may run "clinfo" command after sourcing the script file (setvars.bat).

To source oneAPI environment, run the following command (if installed in default directory):

source /opt/intel/oneapi/setvars.sh

oneAPI toolkits come with a classic Fortran compiler(ifort) as well as Fortran beta compilers (ifx), based on LLVM technology.

https://software.intel.com/content/www/us/en/develop/articles/intel-oneapi-fortran-compiler-release-notes.html

To compile and run Fortran code, you may either invoke ifort or ifx:

ifort filename && ./a.out

Thanks,

Rahul

[lkx]

The specific information of the GPU is shown as the following picture. And the oneAPI version is 2021.1. But I haven't got why you check the devices under the openCL backend by "clinfo" and what do you want to know through the process.

[RahulV_intel · Accepted solution]

Hi,

With the help of the "clinfo" command, you can check the devices that are being detected under the OpenCL backend. You may run your workloads on the detected devices with the help of oneAPI.

Please check the hardware requirements for oneAPI in the link below:

https://software.intel.com/content/www/us/en/develop/articles/intel-oneapi-base-toolkit-system-requirements.html

However, there is an alternate version of DPC++ available on Github, which has limited support for non-Intel accelerators.

https://github.com/intel/llvm/blob/sycl/sycl/doc/GetStartedGuide.md

Coming to the packages, the same issue is reproducible at my end. I will let the know to the concerned team to get it fixed. Since those packages are optional, it shouldn't make any difference. Let me know if you face any issues.

Here's a link to try out vector-add sample using oneAPI:

https://github.com/oneapi-src/oneAPI-samples/tree/master/DirectProgramming/DPC%2B%2B/DenseLinearAlgebra/vector-add

Thanks,

Rahul

[RahulV_intel]

Hi,

Thanks for accepting the solution. Intel will no longer monitor this thread. Further responses on this thread will be considered community only.

Regards,

Rahul