AMD's New RX Radeon VII - NVIDIA RTX 2080 Killer?

[Damir]

Here is a nice video: https://www.youtube.com/watch?v=uhxMUDOf6I0

[henk2]

Video opened with 'for AMD fans' and that was enough for me to end it.
Don't need that silly trabilism.

[Jhoravi]

Does it have A.I. hardware? If not then it's not a 2080 killer as far as chess is concerned.

[smatovic]

Does it have A.I. hardware? If not then it's not a 2080 killer as far as chess is concerned.

Afaik, AMD still has no TensorCores on board,
but it remains open how the price per Knps for LC0 will be...

--
Srdja

[Dann Corbit]

Does it have A.I. hardware? If not then it's not a 2080 killer as far as chess is concerned.

Afaik, AMD still has no TensorCores on board,
but it remains open how the price per Knps for LC0 will be...

--
Srdja

I think the LC0 code is strictly written for Nvidia cards, so a complete rewrite would be needed for OpenCL.

[CMCanavessi]

Does it have A.I. hardware? If not then it's not a 2080 killer as far as chess is concerned.

Afaik, AMD still has no TensorCores on board,
but it remains open how the price per Knps for LC0 will be...

--
Srdja

I think the LC0 code is strictly written for Nvidia cards, so a complete rewrite would be needed for OpenCL.

There are 3 versions of lc0, CUDA, OpenCL and CPU so it's already there. Thing is it's way slower than CUDA.

[Dann Corbit]

Does it have A.I. hardware? If not then it's not a 2080 killer as far as chess is concerned.

Afaik, AMD still has no TensorCores on board,
but it remains open how the price per Knps for LC0 will be...

--
Srdja

I think the LC0 code is strictly written for Nvidia cards, so a complete rewrite would be needed for OpenCL.

There are 3 versions of lc0, CUDA, OpenCL and CPU so it's already there. Thing is it's way slower than CUDA.

Srdja is an OpenCL expert, so maybe he can fix it.

[smatovic]

Does it have A.I. hardware? If not then it's not a 2080 killer as far as chess is concerned.

Afaik, AMD still has no TensorCores on board,
but it remains open how the price per Knps for LC0 will be...

--
Srdja

I think the LC0 code is strictly written for Nvidia cards, so a complete rewrite would be needed for OpenCL.

There are 3 versions of lc0, CUDA, OpenCL and CPU so it's already there. Thing is it's way slower than CUDA.

Srdja is an OpenCL expert, so maybe he can fix it.

Thanks for the Kudos, but i am really a layman in neural networks.

Afaik, GCP wrote Leela Zero initially with an OpenCL backend,
and this was ported to Leela Chess Zero,
then Ankan added an high performant cuDNN backend,
and with the new Nivida RTX series this backend profits from TensorCores.

So i think it is AMDs turn to release some optimized OpenCL NN libraries
and add TensorCores, or a similar technique, to their gpus.

--
Srdja

[Dann Corbit]

So i think it is AMDs turn to release some optimized OpenCL NN libraries
and add TensorCores, or a similar technique, to their gpus.

--
Srdja

Here is a NN lib for OpenCL:
https://github.com/hughperkins/DeepCL
I've no idea how good it is.
Or maybe:
https://github.com/plaidml/plaidml
Or maybe:
https://github.com/keras-team/keras

[Dann Corbit]

So i think it is AMDs turn to release some optimized OpenCL NN libraries
and add TensorCores, or a similar technique, to their gpus.

--
Srdja

Here is a NN lib for OpenCL:
https://github.com/hughperkins/DeepCL
I've no idea how good it is.
Or maybe:
https://github.com/plaidml/plaidml
Or maybe:
https://github.com/keras-team/keras

As to your second point:
https://hothardware.com/news/amd-7nm-na ... p-learning