cancel
Showing results for 
Search instead for 
Did you mean: 

AI Discussions

Science Project

Hello! I have my own scientific project on integrating a special neural network into a GPU, which can make a major breakthrough in performance. How can I get in touch with AMD to present this project to them?

1 Solution

Hello! If you are an AMD employee, please contact me via a private message or on Telegram.

View solution in original post

9 Replies
BigAl01
Volunteer Moderator

Are you looking for technical or financial support?  I would suggest contacting @Sam_AMD to see if he has a contact at AMD that can chat with you.  Being a retired MSEE, I am curious about your project at a high level.  Are you going to develop software to run on a single GPU or are you looking into assembling multiple GPUs into this neural network?  If this is for a school science project, what is the time frame for completing it?


As Albert Einstein said, "I could have done so much more with a Big Al's Computer!".

Thank you for your response! My project involves creating a neural network that will perform video memory caching, predict video memory usage, and other functions aimed at optimizing video memory performance.

Sam_AMD
Community Manager

Hello, @viktor_vizner 

I'm going to move your post to the new AI Community Discussions. We have a Community Manager that keeps an eye on discussion and should be able to help you. 

Sam
Your biggest fan!
CPU: AMD Ryzen 7 5800X3D GPU: AMD Radeon RX 6800XT

I'm sorry. Could you please provide a direct contact?

Hello @viktor_vizner, we do not provide our employee's contact information, I'm sure you understand. But I'm mentioning her now @Lindsey_Brown

Please give her a day or two to respond. 

 

Sam
Your biggest fan!
CPU: AMD Ryzen 7 5800X3D GPU: AMD Radeon RX 6800XT

Hello! Thanks you for response.

Hi @viktor_vizner !

Nice to meet you! I would love to learn more. May I ask what AMD GPU your project is based on so I know the correct contact for you? 

thanks!

Hello! Thank you for the response! My project involves two integration options:

1. Through remote AMD servers, in this case, it is necessary to develop a low-level driver, and this neural network will be able to work with any GPU.

2. Direct integration into the GPU, in Compute Units cores, then a sufficiently large number of CU cores is required, which are available in the RX 5000 and RX 6000 lineups.

Hello! If you are an AMD employee, please contact me via a private message or on Telegram.