CCC has serious hardware update!

[Dann Corbit]

The point is that on both CCC and TCEC the fast GPUs are being vastly underutilized by the extremely slow MHz CPUs.

I am not sure if it works the same with the Epyc CPUs, but if you are using Threadripper CPUS and only use a few, they run at a much higher clock rate.
The scheduler seems to know which CPUs are able to take the most clock rate.

[brianr]

Look at the specs for the max extra clock speeds...they are very low for the CPUs with a huge number of cores.

[Dann Corbit]

Look at the specs for the max extra clock speeds...they are very low for the CPUs with a huge number of cores.

Seems like the optimal LC0 server might have 8 cores at 7 or 8 GHz cooled with liquid nitrogen for the CPU part.
It would probably still cost less than what they are using.

The 3970x has been overclocked to 5.5GHz:
https://www.pcgamesn.com/amd/cpu-overcl ... ld-records

[brianr]

No, N GPUs plus 1 is optimal number of CPUs, so a fast OC's six way would handle just about anything.

[Dann Corbit]

Despite the big advantage for SF over LC0, I predict LC0 will leapfrog SF within a year or two due to the new AMD architecture.
The CPU and GPU will both have transparent, synchronized access to both the video RAM and the regular RAM, so all of the copy back and forth nonsense will be a thing of the past. The CPUs can read directly from the video RAM and the GPUs can read directly from the system RAM.
Do I hear angels singing?

[Modern Times]

CPUs: 2 x AMD EPYC 7H12
GPU: 2x A100 (40 GB GPU memory)
Cores: 256 cores (128 physical)
RAM: 512GB DIMM DDR4 2933 MHz (0.3 ns)
SSD: 2x Micron 5210 MTFD (2TB) in RAID1
OS: CentOS 8

What's TCEC answer ?

Why do they need to answer. Both have massive hardware. Will it change any results.

If you notice even on our much smaller hardware. The ranking of the engines does not really change.

And this is also what I am seeing in my testing.

Yes, I don't see the point in this huge hardware. As you say, the results are probably the same on 32 cores, and I'd guess none of the authors have tested their engines on this level of hardware. Better quality chess however you define that, maybe, but you also probably increase the draw rate as well. Which is when they have to move towards unbalanced openings as an artificial way to spice things up.

[Modern Times]

Despite the big advantage for SF over LC0, I predict LC0 will leapfrog SF within a year or two due to the new AMD architecture.

Yes interesting times ahead, Lc0 is not out for the count just yet.

[brianr]

Do I hear angels singing?

The angels were singing Alpha Zero four years ago.
The show went on the road and tickets were free.
And, now you want the angels to do a new show?

[Dann Corbit]

Do I hear angels singing?

The angels were singing Alpha Zero four years ago.
The show went on the road and tickets were free.
And, now you want the angels to do a new show?

Ecclesiastes 5:10

[shinkarom]

256 cores is more than many companies had in the 1990s, maybe the whole universities.