The Potential Disruptiveness of AMD’s Open Source Deep Learning Strategy

[Bacon1]

Theano

The backend was designed to support OpenCL, however current support is incomplete. A lot of very useful ops still do not support it because they were ported from the old backend with minimal change.

Plus before this new backend, it only worked with CUDA.

TensorFlow

Which is from Google which for sure woudl ahev the money and capacity to add OpenCL if they wanted to. But given above link it's very unlikely to happen. There are some efforts to make this work but if you want to use it now you need CUDA eg NV GPU.

Deeplearning4j

Roadmap:
Medium priority:

• OpenCL for ND4J

Molecular Modeling

Go through the list. Most of them only support CUDA eg NV. Some mention OpenCL but if you read the docs it only really works with CUDA. Yeah some work with OpenCL but it's far the minority.

Vega will support Tensorflow, Cafe2, Cafe, Torch7 and MxNet via MIOpen.

 

[krumme]

Vulkan isn't generating any more Android profits for AMD than Mantle did, which is precisely $0.00. AMD is still losing money. That makes Mantle, Vulkan and DX12 a collective financial failure for AMD. They spent money on it and got nothing in return.


AMD isn't a not-for-profit charity. They're a corporate entity with a legal obligation to make a profit for shareholders. You don't do that by giving away IP for free.


Nonsense, they're just lead by people with a non-existent business acumen. Mantle/Freesync/Crossfire/HBM all should've been licensed, just as nVidia collects fees from every single gsync monitor sold, and every single SLI motherboard with "certification".

They also should be locking out nVidia GPUs from their chipsets unless nVidia pay "certification" fees, exactly how nVidia locks out paying customers from PhysX when they detect a non-nVidia GPU in the system. Clearly nVidia's customers appreciate such tactics given they continue to make a profit. So now it's time for AMD to start cashing in.

Instead AMD expends resources and then hands the result to competitors for free. Now they're doing the same with this deep learning.

The definition of insanity is to keep repeating the same thing over and over, yet expect a different result.

While i agree to a certain degree here about the lack of business talent its pretty obvious looking at history Amd couldn't compete on cpu st perf and monster size gpu.

Why. Because its a Capex game and they got no funds. Dx11 was evidently a glass roof that needed to be removed. Mantle did that. It paved the way for ryzen. Without mantle and dx12/vulcan ryzen would fare far worse. In pc as well as future consoles. Its financially a huge success.
If we take gsynch vs freesynch the problem with gsynch is it locks a lot of users out. Its stupid expensive. Ofcource it locks a lot of consumers in. Thats nice but i guess most of those users would stay there anyway. Gsynch certainly didnt progress as planned and freesynch is the reason for it. Freesynch is a success.
Neither freesynch or mantle could give those crucial important results for amd. Lifethreatening important results - if they would have been licenced.
So its not so simple as you describe it.
As a matket leader you can act as nv. If you are the small guy you cant do the same moves.

The big companies tries to monopolize - not to be creative. They learn how to do that in school so most are pretty good at it.