- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hey! I am interested in purchasing an Intel ARC A770 graphics adapter to do GPU offload of calculations (using oneAPI Toolkits, either C++ or Fortran).
Can I install the A770 alongside with my existing adapter (an NVIDIA RTX A2000 12GB) and have the A770 do only the GPU offload of calculations?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks. Yes, I will look for further (soon it seems) info on the Intel Data Center GPU Max series boards.
An answer over in the Fortran forum indicates that the A770 can "emulate" double precision floating point using the Intel Graphic Compiler by IGC_EnableDPEmulation=1 before compiling, but that emulation is very slow.
Another answer there indicates that the GPU Max may be available on the DevCloud by the end of March.
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello wtstephens,
Thank you for posting on the Intel️® communities.
To further support you with this information, please provide us with further details about the specific project you are looking for.
Best regards,
Jean O.
Intel Customer Support Technician
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Jean,
Thanks for the response!
I have custom built neural network code, which does a lot of double-precision floating point dot products . It is written in Fortran, but I could migrate the intensive parts of the code to C++ -- since Go or Scala are not oneAPI languages, ha ha!
Also, there may be some things in the oneAPI Toolkits which may provide acceleration which could replace parts of my custom code.
My application data does not take up very much memory, say, much less than 1GB.
Warren
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello wtstephens,
Let us look into this and we will get back to you at our earliest convenience.
Best regards,
Jean O.
Intel Customer Support Technician
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello wtstephens,
Thanks for waiting for a response.
After reviewing this matter, It sounds like you are making your software. Therefore, we like to know if you are the software developer of the app you're planning on using. Or do you have a specific app you want to use with the card?
Best regards,
Jean O.
Intel Customer Support Technician
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Jean,
Yes, I am the developer of the software -- so I would have to write the GPU offload code myself.
Warren
----- edit ----
I did have some time for research on this topic today.
I will look further into the Intel GPU Max 1100 -- announced in January 2023, but no pricing or availability yet.
The GPU Max 1100 does not support an attached display, as it is for computation offload only -- which is what I am looking for.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello wtstephens,
Thanks for the information provided.
We understand that you found information about the Intel® Data Center GPU Max 1100, and this card meets your requirements. Therefore we would like to confirm if you would like to continue seeking an answer regarding your original questions.
Best regards,
Jean O.
Intel Customer Support Technician
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks. Yes, I will look for further (soon it seems) info on the Intel Data Center GPU Max series boards.
An answer over in the Fortran forum indicates that the A770 can "emulate" double precision floating point using the Intel Graphic Compiler by IGC_EnableDPEmulation=1 before compiling, but that emulation is very slow.
Another answer there indicates that the GPU Max may be available on the DevCloud by the end of March.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello wtstephens,
We noticed that you have marked this thread as solved, we want to confirm if you would like to close this thread now.
Let us know if this was an error, or if you agree with closing this case
Best regards,
Jean O.
Intel Customer Support Technician
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello wtstephens,
We hope you are doing fine.
We have not heard back from you. So we will close this thread. If you need any additional information, submit a new question, as this thread will no longer be monitored.
Best regards.
Jean O.
Intel Customer Support Technician.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello wtstephens,
Before closing the case, if you need more recommendations or support about your software implementation, we advise you to create an account to submit your inquiry in our Intel®️ Developer Zone, as the document you refer to comes from this website. So, you receive further troubleshooting support and get access to documentation.
Intel®️ Developer Zone - Contact Page
https://software.intel.com/en-us/support
Best regards,
Jean O.
Intel Customer Support Technician

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page