Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AMD Radeon VII: High-End 7nm Vega Video Card

anandtech.com

Post reply on HN
Print view View original

71–80 of 91 posts

Re: AMD Radeon VII: High-End 7nm Vega Video Card

#71

Post by giobox »

shmerl wrote: post #38
dragontamer wrote: post #2

Its an interesting design for sure. 4x Stacks of HBM2 means 1TBps memory bandwidth at 16GB. Only 60-compute units enabled (maybe 4-CUs are expected to break during manufacturing? Its a weird number for sure...). Since it shares dies with the MI50, the Radeon VII will have 1/2 speed double-precision, making this the cheapest high-performance double-precision card in existance. ------ FP16 compute is supported at doubl…

Would it make sense for them to make a 8GB variant as well? 16GB make it very expensive and actually for gaming it's most likely not needed yet.

It looks like one of the card’s headline features, 1TB/sec of memory bandwidth, was only achievable with a 16gb design (basically a doubling of channels/memory over the previous Vega 64 design, with which this card shares many similarities).

Re: AMD Radeon VII: High-End 7nm Vega Video Card

#72

Post by shaklee3 »

nradov wrote: post #11
doctorpangloss wrote: post #10

Earlier quoted context omitted.

> Deep learning enthusiasts probably will stick to NVIDIA cards We've been building the wrong hardware for ML for a while now. A TPU doesn't delivery supremacy for problems over GPU hardware. It's a Google senior engineering retention and PR project. Exceedingly few problems resemble image recognition. You wouldn't be able to tell from the research, because due to the tools it's sort of the only affordable thing to d…

What is the right hardware for ML?

They absolutely need tensor/matrix multiply cores. Incremental fp32 improvements aren't impressive.

Re: AMD Radeon VII: High-End 7nm Vega Video Card

#73

Post by shaklee3 »

vl wrote: post #50
twtw wrote: post #39

Earlier quoted context omitted.

So that's why Xeon Phi has been so successful! /s Real talk, though. Why do you think larabee is the "right" hardware?

The truth is that for many types/sizes of models AVS2 or AVS512 in Xeons is as fast and GPUs.

This is not true. I've benchmarked it, and you can see yourself by looking at the MKL benchmarks versus cuBLAS. cuBLAS is always significantly faster.

Re: AMD Radeon VII: High-End 7nm Vega Video Card

#74

Post by shaklee3 »

Cybiote wrote: post #58
frankchn wrote: post #55

Earlier quoted context omitted.

I am interested in seeing benchmarks regarding this. I believe this is true for models with truly huge embeddings, but otherwise if your models are even a little compute dense then GPUs are faster.

My experience is that you need the additional caveat of a streaming fairly homogeneous and highly parallelizable approach or the memory transfer, branching and communication overheads will eat away nearly all the GPU gains. I've also noticed that a lot of the time, people are comparing GPU to naive implementations (and sometimes, implementations written in dynamic languages) instead of to highly tuned BLAS or MKL imp…

No, they're not. Straight front the horse's mouth:

https://software.intel.com/en-us/mkl/features/benchmarks

In no case will you see it get close to 10TFLOPS. GPUs easily do this, and can approach 100TFLOPS with tensor cores.

Re: AMD Radeon VII: High-End 7nm Vega Video Card

#75

Post by shmerl »

giobox wrote: post #71
shmerl wrote: post #38

Earlier quoted context omitted.

Would it make sense for them to make a 8GB variant as well? 16GB make it very expensive and actually for gaming it's most likely not needed yet.

It looks like one of the card’s headline features, 1TB/sec of memory bandwidth, was only achievable with a 16gb design (basically a doubling of channels/memory over the previous Vega 64 design, with which this card shares many similarities).

Why can't they use 4x2GB and provide the same bandwidth? Or it's some limitation of HBM2?

Re: AMD Radeon VII: High-End 7nm Vega Video Card

#76

Post by greatpatton »

m0zg wrote: post #57

It'd be super-nice if AMD also produced, at a minimum, fully supported versions of PyTorch and TensorFlow for these. NVIDIA would crap its pants if perf is comparable (and possibly drop the prices some).

It is just right here:

* https://rocm-documentation.readthedocs.io/en/latest/Deep_lea... * https://github.com/ROCmSoftwarePlatform/tensorflow-upstream

Re: AMD Radeon VII: High-End 7nm Vega Video Card

#78

Post by dnautics »

doctorpangloss wrote: post #10
dragontamer wrote: post #2

Its an interesting design for sure. 4x Stacks of HBM2 means 1TBps memory bandwidth at 16GB. Only 60-compute units enabled (maybe 4-CUs are expected to break during manufacturing? Its a weird number for sure...). Since it shares dies with the MI50, the Radeon VII will have 1/2 speed double-precision, making this the cheapest high-performance double-precision card in existance. ------ FP16 compute is supported at doubl…

> Deep learning enthusiasts probably will stick to NVIDIA cards We've been building the wrong hardware for ML for a while now. A TPU doesn't delivery supremacy for problems over GPU hardware. It's a Google senior engineering retention and PR project. Exceedingly few problems resemble image recognition. You wouldn't be able to tell from the research, because due to the tools it's sort of the only affordable thing to d…

I think we are at a point where the cool hardware will actually be low power inference (deployed on basically the class of models we have now).

Re: AMD Radeon VII: High-End 7nm Vega Video Card

#79

Post by mda »

zozbot123 wrote: post #52
doctorpangloss wrote: post #10

Earlier quoted context omitted.

> Deep learning enthusiasts probably will stick to NVIDIA cards We've been building the wrong hardware for ML for a while now. A TPU doesn't delivery supremacy for problems over GPU hardware. It's a Google senior engineering retention and PR project. Exceedingly few problems resemble image recognition. You wouldn't be able to tell from the research, because due to the tools it's sort of the only affordable thing to d…

Matrix multiplication is the bread-and-butter of ML (both training and inference), for fairly obvious and solid reasons. Luckily for us, it's also very important in all sorts of domains, including graphics, rendering and gaming. That's all there is to it. There's no real "subsidy from the gaming industry", only a cross-subsidy situation in which common problems that are shared across domains can take advantage of the…

I'm not sure what you mean by "ML features", but AMD/ATI's dismissive attitude towards the ML and GPU-compute communities is looking quite silly by now.

I disagree, ML computing has tons of competition and its margins will be quickly diminish. I think they are right to focus on gaming.

Re: AMD Radeon VII: High-End 7nm Vega Video Card

#80

Post by pas »

shmerl wrote: post #75
giobox wrote: post #71

Earlier quoted context omitted.

It looks like one of the card’s headline features, 1TB/sec of memory bandwidth, was only achievable with a 16gb design (basically a doubling of channels/memory over the previous Vega 64 design, with which this card shares many similarities).

Why can't they use 4x2GB and provide the same bandwidth? Or it's some limitation of HBM2?

yes, apparently. though thry could probably get it done eventually, probably not worth it.

https://news.ycombinator.com/item?id=18868617

Post reply on HN