Facebook Is Forming a Team to Design Its Own Chips

Re:Facebook Is Forming a Team to Design Its Own Chips by tsomctl on Thursday April 19, 2018 @10:41AM (#16876274)

This is something that I've wondered about. There is nothing about Tensorflow (for example) that is specific to CUDA. Why doesn't AMD dedicate a couple of engineers to port it? They don't even have to do a complete job. Only about half of the Tensorflow operations are used for most models.

[ Parent | on HN ]

Re:Facebook Is Forming a Team to Design Its Own Chips by sanxiyn on Thursday April 19, 2018 @11:31AM (#16876723)

AMD does. https://github.com/ROCmSoftwarePlatform/tensorflow

[ Parent | on HN ]

Re:Facebook Is Forming a Team to Design Its Own Chips by patrickg_zill on Thursday April 19, 2018 @11:01AM (#16876467)

If they make CUDA the industry standard by porting it to AMD, and then Nvidia is still able to control the standard, over the long term AMD will always be in the position of having to catch up.

[ Parent | on HN ]

Re:Facebook Is Forming a Team to Design Its Own Chips by jandrese on Thursday April 19, 2018 @11:27AM (#16876694)

I think the GP is suggesting AMD port TensorFlow to OpenCL to break up the CUDA monopoly.

[ Parent | on HN ]

Re:Facebook Is Forming a Team to Design Its Own Chips by shmerl on Thursday April 19, 2018 @12:08PM (#16877045)

No need, there is already an industry standard that works across all hardware.

[ Parent | on HN ]