Intel oneAPI DPCPP vs open-source oneAPI and Nvidia support

[wsa]

In Goutham's response to a separate question, he said that there is an Intel oneAPI and an opensource oneAPI.

1) Can you elaborate further on the difference? I bought the Data Parallel C++ by Reinders et al. Is that book for Intel oneAPI or opensource oneAPI? It tells me to install  dpc++ using https://github.com/intel/llvm  and recommends using clang++, which makes me think it's the opensource oneAPI. 

2) If that getting started page is supported by intel, can you help me understand the getting started guide and it's relationship with Nvidia GPUs?  Do i need to install the experimental CUDA support toolchain to utilize my nvidia gpu? What if I just have nvidia driver installed in my /etc/openCL directory? 

[PriyanshuK_Intel]

Hi,

Thank you for posting on Intel Communities.

>> Can you elaborate further on the difference?

The difference between Intel oneAPI and Opensource oneAPI is that Intel oneAPI is more refined and optimized version of oneAPI including additional features. 

The performance of the Intel C/C++ compilers can be expected to give higher performance than the base clang+ LLVM compilers for Intel architecture. 

For more information you can refer this link: https://www.intel.com/content/www/us/en/developer/articles/technical/adoption-of-llvm-complete-icx.html

>> If that getting started page is supported by Intel, can you help me understand the getting started guide and it's relationship with Nvidia GPUs?

This forum is intended to support Intel oneAPI Products and other Intel Products. As we can see that you are using an open-source version of oneAPI for Nvidia devices support, we request you to raise an issue in GitHub for faster response: https://github.com/intel/llvm/issues

Thanks ,

Priyanshu.

[wsa]

Okay thank you. if  intel doesn't support the opensource oneAPI, why does the opensource oneAPI have "intel" in the url for it's github page? it's a little confusing. 

also 

A moderatore wrote in the below page that as of 2020, nvidia gpu's weren't supported by Intel OneAPI. Has that support changed in the past 2 years? Any plans to add Nvidia GPU support? 

https://community.intel.com/t5/Intel-oneAPI-Data-Parallel-C/Running-DPC-on-NVIDIA-GPU-via-OpenCL-instead-of-CUDA/m-p/1193882/thread-id/597

thanks for taking the time to answer my questions. really helpful.  

[PriyanshuK_Intel]

Hi,

Thank you for your patience.

>>why does the opensource oneAPI have "intel" in the url for it's github page?

Since the repository is still mainly contributed by Intel, that's why there is Intel in the url. 

>> Any plans to add Nvidia GPU support? 

Currently NVIDIA GPU is not supported by the dpcpp compiler released in the oneAPI toolkit but 

there is a plan to add such support in oneAPI future release. 

Thanks,

Priyanshu.

[PriyanshuK_Intel]

Hi,

We assume that your issue is resolved. If you need any additional information,

please post a new question as this thread will no longer be monitored by Intel.

Thanks,

Priyanshu.