Cfish is back!

[syzygy]

Unfortunately the node counts of longer benches are still off

For example bench 512 1 24: 207806902 for Cfish, 209367314 for SF.

512 is just a number I picked. For other hash sizes the depth from where node counts start to differ may be lower or higher than 24.

[syzygy]

Unfortunately the node counts of longer benches are still off

For example bench 512 1 24: 207806902 for Cfish, 209367314 for SF.

512 is just a number I picked. For other hash sizes the depth from where node counts start to differ may be lower or higher than 24.

It seems I missed the patch that decided that Stockfish should not evaluate this position as a draw:
[d]8/8/5k1p/8/7p/7K/8/8 b - - 1 11

[MikeB]

Unfortunately the node counts of longer benches are still off

For example bench 512 1 24: 207806902 for Cfish, 209367314 for SF.

512 is just a number I picked. For other hash sizes the depth from where node counts start to differ may be lower or higher than 24.

It seems I missed the patch that decided that Stockfish should not evaluate this position as a draw:
[d]8/8/5k1p/8/7p/7K/8/8 b - - 1 11

Good catch Ronald, I had just tracked it down the simplification patch made on March 14 and you had already had posted a comment about the same patch on Github.

[syzygy]

With this fixed in Stockfish, bench 512 1 30 returns identical node counts.

[MOBMAT]

is there a link to builds?

[syzygy]

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:
https://github.com/syzygy1/Cfish/commit ... 30521a6b81

This makes a difference if /sys/kernel/mm/transparent_hugepage/enabled is set to "madvise".

[MikeB]

is there a link to builds?

https://abrok.eu/stockfish/ - cfish was already correct

[zullil]

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:
https://github.com/syzygy1/Cfish/commit ... 30521a6b81

This makes a difference if /sys/kernel/mm/transparent_hugepage/enabled is set to "madvise".

This might explain why my system is set to "always". It was "madvise" and I recall changing it to "always". But I can't remember why. Let me revert it and check that Cfish still uses anonymous hugepages.

{EDIT} Yes, all is good.

$ cat /sys/kernel/mm/transparent_hugepage/enabled
always [madvise] never

$ cat proc/3579/smaps | grep "AnonHugePages"
AnonHugePages: 0 kB
AnonHugePages: 0 kB
AnonHugePages: 0 kB
AnonHugePages: 0 kB
AnonHugePages: 0 kB
AnonHugePages: 33554432 kB

That would be Cfish's 32 GB hash table, I assume.

[syzygy]

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:
https://github.com/syzygy1/Cfish/commit ... 30521a6b81

This makes a difference if /sys/kernel/mm/transparent_hugepage/enabled is set to "madvise".

This might explain why my system is set to "always". It was "madvise" and I recall changing it to "always".

That explains why you never complained

I have now released Cfish 11a:
https://github.com/syzygy1/Cfish/releases/tag/cfish_11a

[Geonerd]

I have now released Cfish 11a:
https://github.com/syzygy1/Cfish/releases/tag/cfish_11a

Thank you for the executables!