Questions (and suggestions) to CCRL, CEGT and GUI authors

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

Milos
Posts: 4190
Joined: Wed Nov 25, 2009 1:47 am

Re: Questions (and suggestions) to CCRL, CEGT and GUI author

Post by Milos »

Adam Hair wrote:You are right!

Also, Nathan submitted a benchmark for that computer to Sedat. His benchmark @ 4 cores is 15% slower than the i5-6600k @ 4 cores. Assuming that also holds for 1 core, the i7-6900k would probably use a time control of 40/13.
All assuming prehistoric 32bit Crafty not using any modern speed-up instruction.
Just SSE4.2 that AMD doesn't have brings at least 15% speed-up, plus much more efficient 64bit architecture of newest Haswells compared to old AMD K8 core. On top of that add much faster hash access and you'll get at least 25% speed-up overall using the latest SF.
Modern Times
Posts: 3894
Joined: Thu Jun 07, 2012 11:02 pm

Re: Questions (and suggestions) to CCRL, CEGT and GUI author

Post by Modern Times »

It is an incredibly old 32-bit exe, yes.
Adam Hair
Posts: 3226
Joined: Wed May 06, 2009 10:31 pm
Location: Fuquay-Varina, North Carolina

Re: Questions (and suggestions) to CCRL, CEGT and GUI author

Post by Adam Hair »

Milos wrote:
Adam Hair wrote:You are right!

Also, Nathan submitted a benchmark for that computer to Sedat. His benchmark @ 4 cores is 15% slower than the i5-6600k @ 4 cores. Assuming that also holds for 1 core, the i7-6900k would probably use a time control of 40/13.
All assuming prehistoric 32bit Crafty not using any modern speed-up instruction.
Just SSE4.2 that AMD doesn't have brings at least 15% speed-up, plus much more efficient 64bit architecture of newest Haswells compared to old AMD K8 core. On top of that add much faster hash access and you'll get at least 25% speed-up overall using the latest SF.
That is true. And I won't disagree that changing the benchmark engine to a recent SF would bring the time control for an i7-6900k at stock speed closer to 40/10. But that would be a lower bound according to the Sedat SF benchmarks.
Modern Times
Posts: 3894
Joined: Thu Jun 07, 2012 11:02 pm

Re: Questions (and suggestions) to CCRL, CEGT and GUI author

Post by Modern Times »

Doesn't worry me what we call it, it is what it is. Another few years and it will be the new 40/4 !!