Milos wrote:It actually is, instead of 4TPUs required to run Alpha0 so far, on x64 hardware one would need around 2000 Haswell cores to achieve the same speed of NN
For NN inference, 1 TPU is around 15-30 times faster than an Haswell (with multiple cores, so your figure of 2000x faster for 4 TPUs vs 1 core is right), but other comparable properties favor the CPU:
TPU vs CPU
Die size: TPU is < half size of an Haswell
Frequency: 700 Mhz vs 2300
Power: 40W vs 145W
Memory Bandwitdth: 51 vs 34
The TPU is much more efficient than a CPU (for NN inference), but it does not eat more power, occupies more size, etc.
So using 4 TPU vs 64 Haswell cores, is fair I think.
Apart from frequency, how would the other parameters affect speed?
Is not frequency already factored in the '20 times faster assessment'?
2000/1, with added SMP inefficiencies, will make it 3000/1, 3000/64,
in what way could that be fair?
I agree with you.
Also what people who claim the HW was much faster..... what they don’t understand is that the thing learned from itself in a very short time!
What if we now give it 1 Year to further learn?
Side note, I looked at the games and they are really impressive!
Lyudmil Tsvetkov wrote:With what is this different from a self-tuning software, as widely used in autotuning engines, applied on a very large scale/involving tremendous hardware?
Most of current chess eval can be modelised as a single neuron (or perceptron). AlphaGo uses many neurons. That's a big difference.
Many neurons, but of what quality?
It should be very bad, in case with 1000 times more neurons than SF, it will achieve 400 elo lower strength on single core.
jhellis3 wrote:Or 4 V100s which will be available in 2018, or you can order in a workstation now for 70k.
You seem to think GPU progress and process node shrinks are just going to stop. V100 is on 12nm and 7nm chips will have mass availability in 2018.
It is over. Deal with it. Or not.
7nm chips in 2018, give me a break. You clearly have no clue what you are talking about beside quoting some marketing BS.
First you have to understand basic staff about NN inference, difference between floating point single or double precision operation and int8 multiplication done in TPU.
Performance-wise 2nd gen TPU is still 5-10x compared to V100 that we first have to see come out and that comparing marketing data from the advertisement, not real numbers.
Lyudmil Tsvetkov wrote:Alpha had considerable hardware advantage
That comparison is not straightforward, but this claim does not seem to be true. SF had 64 threads. I'm not up on the latest scaling behavior of the engine but that has got to be near saturation.
-Carl
From what I gleaned from hardware comparisons, the advantage is 16/1.
Why would one want to run a similar very unfair match?
Only one thing comes to mind: that the company will want to advertise its colossal breakthrough with TPUs and artificial intelligence and then sell its products.
But then, the achievement is not there.
The fact that Google has created a chess playing entity that crushes SF is notable (and fascinating).
TPUs are not for sale, and (at the moment) are applied only to Googles deep learning and research projects,
except when Google donates them to research for free.
What would be the score between SF on 64 cores and SF on 1024 cores out of 100 games?
You think the bigger-hardware SF would score less than 64 points?
I guess at least 80.
So what is so new?
They applied some big hardware, that is all.
The real strength of Alpha is 2850, so around spot 97 or so among engines.
97 is not such a bad achievement, after all.
I doubt if SF on 1024 cores is going to score even 50%
Maybe after some point more cores are counter productive for stockfish.
I also doubt if it is possible to get at least 80 points against stockfish with 64 cores at 1 minute per move.
Lyudmil Tsvetkov wrote:
So, that, in reality, the hardware difference is not 16/1, as I thought initially, but more like 30/1.
Add to this the early opening advantage Alpha gets due to the simulated book, and conducting the test has been fully meaningless.
Alpha would play not stronger than 1850 on a single core.
Why would I care for such an engine?
Meaningless?
What simulated book?
The point of this was to develop and demonstrate a new level of machine learning and AI using these processors...
AlphaZero obtained the knowledge it needed to beat SF in 4 hours!
No one at Google believes this was a super TCEC World Championship event,
it's simply science, development, and research.
Last edited by kranium on Wed Dec 06, 2017 9:44 pm, edited 1 time in total.
