Search found 4155 matches

by syzygy
Tue Jul 14, 2020 12:32 pm
Forum: Computer Chess Club: General Topics
Topic: Which CPUs was used against Kramnik in 2002 ?
Replies: 12
Views: 506

Re: Which CPUs was used against Kramnik in 2002 ?

Not really, the Pentium 4 is known to do less thing per clock cycle. It was a will from Intel to produce a processor with very high frequency but sacrifice optimization. This changed completely with Pentium-M and Core-2 : - a 1.6 GHz Pentium M can typically attain or even surpass the performance of...
by syzygy
Mon Jul 13, 2020 9:38 pm
Forum: Computer Chess Club: General Topics
Topic: Which CPUs was used against Kramnik in 2002 ?
Replies: 12
Views: 506

Re: Which CPUs was used against Kramnik in 2002 ?

It seems Fritz 9 reached about 480knps on a 1Ghz Pentium III:
http://www.jens-hartmann.at/Fritzmarks/

I have no idea when Fritz 9 was released.

This article mentions 4Mnps:
http://news.bbc.co.uk/2/hi/europe/1466810.stm
by syzygy
Mon Jul 13, 2020 6:59 pm
Forum: Computer Chess Club: General Topics
Topic: Cfish is back!
Replies: 44
Views: 2071

Re: Cfish is back!

Cfish 11a- 8 cores 2048 MB Hash: 44/64 13:45 6.094.447.028 7.384.049 +1,59 Dd2-e3 Ta8-b8 This is the official popcntn windows version from the github page. i doubled the cores and the result is clearly worse against the old version with only 4 cores. Not so nice. Same version Cfish 11a with 2048 MB...
by syzygy
Mon Jul 13, 2020 6:43 pm
Forum: Computer Chess Club: General Topics
Topic: Cfish is back!
Replies: 44
Views: 2071

Re: Cfish is back!

Cfish 11a- 8 cores 2048 MB Hash: 44/64 13:45 6.094.447.028 7.384.049 +1,59 Dd2-e3 Ta8-b8 This is the official popcntn windows version from the github page. i doubled the cores and the result is clearly worse against the old version with only 4 cores. Not so nice. Your "old" version of 6 July is a l...
by syzygy
Mon Jul 13, 2020 10:19 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Data structure choice for TT
Replies: 16
Views: 511

Re: Data structure choice for TT

As a corollary of von Mises's birthday theorem (e.g. cited as Theorem 2.1 in this preprint ), the number of j-collisions is ~m(k/m)^j/j! (m->inf, k/m->0) if the distribution of hashes is close enough to uniform, so at low nps in engine games, the ratio of 3-collisions is small enough (considering a...
by syzygy
Sun Jul 12, 2020 3:27 pm
Forum: Computer Chess Club: General Topics
Topic: Cfish is back!
Replies: 44
Views: 2071

Re: Cfish is back!

Unless you've used one thread only, what your seeing in these two searches might just be "random noise". And even with 1 thread finding the winning move or not can easily depend on the size of the hash table and the content of the hash table at the start of the search. If Thomas was indeed using th...
by syzygy
Sun Jul 12, 2020 2:25 pm
Forum: Computer Chess Club: General Topics
Topic: Cfish is back!
Replies: 44
Views: 2071

Re: Cfish is back!

Thomas Lagershausen wrote:
Sun Jul 12, 2020 12:57 pm
I am not so happy with latest update because Stockfish getting dumber with more selectivity.
But the aim of Cfish is to track Stockfish.
by syzygy
Sun Jul 12, 2020 10:25 am
Forum: Computer Chess Club: General Topics
Topic: Ryzen 9 3950x with 37 HT cores??
Replies: 8
Views: 510

Re: Ryzen 9 3950x with 37 HT cores??

I just build a PC with Ryzen 9 3950x When I run the Ipman-test (special bench of asmFishW 17.05.22) It showed 37 SMT/HT threads! What is this? The asmFishW obviously calculate bench for 37 threads what yield bad Bench-results. Maybe this is the cause why among the bench-value there are lot of "tele...
by syzygy
Sat Jul 11, 2020 11:36 pm
Forum: Computer Chess Club: General Topics
Topic: Cfish is back!
Replies: 44
Views: 2071

Re: Cfish is back!

I also noticed that, on my laptop, Cfish and SF are about equal in speed when I set the hash size to 1024 or higher. I haven't tried on my desktop yet, but this surprised me a bit. It seems support for transparent large pages on Linux was accidentally disabled quite some time ago in this patch: htt...
by syzygy
Sat Jul 11, 2020 9:10 pm
Forum: Computer Chess Club: General Topics
Topic: Cfish is back!
Replies: 44
Views: 2071

Re: Cfish is back!

I also noticed that, on my laptop, Cfish and SF are about equal in speed when I set the hash size to 1024 or higher. I haven't tried on my desktop yet, but this surprised me a bit. It seems support for transparent large pages on Linux was accidentally disabled quite some time ago in this patch: htt...