Is this version of stockfish supposed to be stronger than stockfish 8 (or the latest version)?
If so then can someone please give me detailed instructions on how to install it and use it with fritz 15 (I know how to install regular uci engines in fritz 15).
Thanks!
Stockfish 8 SSE (Special Edition)
Moderator: Ras
-
- Posts: 4889
- Joined: Thu Mar 09, 2006 6:34 am
- Location: Pen Argyl, Pennsylvania
Re: Stockfish 8 SSE (Special Edition)
Somebody else will have to decide - but if you run it with the Cerebellum book, tbe odds are looking it may be.h1a8 wrote:Is this version of stockfish supposed to be stronger than stockfish 8 (or the latest version)?
If so then can someone please give me detailed instructions on how to install it and use it with fritz 15 (I know how to install regular uci engines in fritz 15).
Thanks!
Code: Select all
3000 game(s) loaded
Rank Name Rating Δ + - # Σ Σ% W L D W% =% OppR
---------------------------------------------------------------------------------------------------------
1 Stockfish SSE 112116 3120 0.0 6 6 1500 811.5 54.1 276 153 1071 18.4 71.4 3093
2 Stockfish SSE 111316 3109 11.5 6 6 1500 776.5 51.8 225 172 1103 15.0 73.5 3097
3 Stockfish-201116 64 POPC 3088 20.6 6 6 1500 714.0 47.6 173 245 1082 11.5 72.1 3104
4 Stockfish 211116 64 POPC 3083 5.1 6 6 1500 698.0 46.5 156 260 1084 10.4 72.3 3106
---------------------------------------------------------------------------------------------------------
http://www.zipproth.de
can't help you with Fritz 15 - macOS user here.
-
- Posts: 4889
- Joined: Thu Mar 09, 2006 6:34 am
- Location: Pen Argyl, Pennsylvania
Re: Stockfish 8 SSE (Special Edition)
CoolVinvin wrote:Amazing ! I just started to play against SF-zh on Lichess yesterday too !MikeB wrote:just really learning Crazy House - this is my best game thus far , using the Ladder rating of 1200 - generally less than 100 nodes ,but keeping skill at 20 ( highest) - also no take backs here ;>). For now using no opening book.
Was in control at move 20, lost my way around move 40, but then came back.
ugh! just did some quick analysis - missed a mate in 2 and a mate in 1 , but like said I'm just getting used to crazyhouse . SF also had several chances to mate me, but due to the weak level, failed to see them.Code: Select all
[Event "Computer Chess Game"] [Site "Mac-Pro.local"] [Date "2016.11.20"] [Round "-"] [White "michaelbyrne"] [Black "Stockfish 8 Variant 64 POPCNT (UCI2WB)"] [Result "1-0"] [TimeControl "120+5"] [Variant "crazyhouse"] [Annotator "18... +5.43"] 1. e4 e5 {-0.45/2 7} 2. Bc4 Nf6 {+0.30/2 7} 3. d3 d5 {+0.93/2 7} 4. exd5 Nxd5 {+0.23/2 7} 5. P@e4 Nf4 {+4.10/2 7} 6. Bxf4 exf4 {+0.21/2 7} 7. Bxf7+ Kxf7 {+1.11/2 7} 8. Nf3 Kg8 {+1.70/2 7} 9. N@e5 P@f7 {+2.14/2 8} 10. O-O Bd6 {+1.62/2 8} 11. Nc4 B@g4 {+1.62/2 8} 12. Nxd6 Qxd6 {+1.83/2 8} 13. P@e2 Nc6 {+3.20/2 8} 14. e5 Nxe5 {+5.35/2 8} 15. Nxe5 Qxe5 {+6.24/2 8} 16. B@c3 Qxe2 {+7.57/2 8} 17. Qxe2 Bxe2 {+0.58/2 8} 18. N@e7+ Kf8 {+5.43/3 8} 19. Q@c5 B@d6 {+8.20/2 8} 20. Bxg7+ Kxg7 {+10.73/2 8} 21. Qg5+ N@g6 {+10.47/2 8} 22. P@h6+ Kf8 {+12.99/4 8} 23. Nxg6+ hxg6 {+13.10/3 8} 24. Re1 Q@e6 {+12.27/2 8} 25. N@e4 B@e7 {+12.22/2 8} 26. Qa5 b6 {+13.56/2 8} 27. Qd2 P@c3 {+11.27/2 8} 28. Nbxc3 Rxh6 {+6.47/2 8} 29. Rxe2 Be5 {+6.09/2 8} 30. Rae1 P@b4 {+7.29/2 8} 31. P@g5 bxc3 {+10.05/2 8} 32. bxc3 Rh4 {+10.18/2 8} 33. Nf6 P@e3 {+11.42/2 7} 34. fxe3 B5xf6 {+10.51/2 7} 35. gxf6 Bxf6 {+9.87/2 7} 36. exf4 Qxa2 {+9.03/2 7} 37. Re8+ Kg7 {+11.60/4 7} 38. B@f8+ Kg8 {+8.85/2 7} 39. Be7+ Kh7 {+1.36/2 7} 40. Bxf6 N@h3+ {-0.73/2 6} 41. Kh1 N@f2+ {+6.48/2 6} 42. Qxf2 P@b2 {-1000.01/1 6} 43. B@g8+ Kh6 {-1000.02/2 6} 44. B@g5+ Nxg5 {-1000.02/2 6} 45. fxg5+ Kh5 {-4.61/2 6} 46. N@f4+ Rxf4 {-2.59/2 6} 47. Qxf4 N@f2+ {-7.46/2 6} 48. Qxf2 N@f5 {-12.95/2 5} 49. R@h6+ Nxh6 {-7.08/2 5} 50. N@g7+ Kg4 {-1000.01/3 5} 51. h3# {Black resigns} 1-0
Lost to level 1 then beat level 1 and level 2 :
https://fr.lichess.org/bnhRRYVM/black
https://fr.lichess.org/y9js4iy0
https://fr.lichess.org/jJHlBAtI

-
- Posts: 4889
- Joined: Thu Mar 09, 2006 6:34 am
- Location: Pen Argyl, Pennsylvania
Re: Stockfish 8 SSE (Special Edition)
A triple pack of refreshed SF for macOS exe's - these SHOULD work on most Macs going back the last 5 years - either compiled with clang or included static libraries - please let me know if you see otherwise.MikeB wrote:Thanks very much T for the windows exe, will place them on my GitHub page. Unfortunately, I'm on macOS which does not use LP and it ir=s not something I'm not familiar with.tpoppins wrote:SF SSE 161120 (rev 4252) Windoze bins:
TinyUpload
MediaFire
Code: Select all
sf-sse-161120-w32-old_tp.exe .... for old Win32 systems sf-sse-161120-w32-sse_tp.exe .... for Win32 systems with SSE sf-SSE-161120-x64_tp.exe ........ for old x64 systems sf-SSE-161120-x64-popcnt_tp.exe . for 'modern' x64 systems with POPCNT sf-sse-lp.patch ................. a Large Pages patch (for self-compiles) The w32 compiles have been built and tested on WinXP+SP3, the rest on Win7 x64 Pro. The x64 builds feature Large Pages support - run the execs or the GUI with admin privileges to enable it. No BMI2 compile at this time. Thanks to Marco Zerbinati for the Large Pages code http://preview.tinyurl.com/hzmqdcf Bench: 5773672
Stockfish 8 Variant Windoze bins:
TinyUpload
MediaFire
Michael, please check out the LP code patch. It generates some (non-critical) warning from the compiler, perhaps you could clean them up and incorporate the patch in your code base? I didn't use it for the Variant bins since for some reason it made the POPCNT exec slower than vanilla old x64 (!); maybe you could figure out why it worked for SSE but not Variant.Code: Select all
sf-8-variant-161120-w32-old_tp.exe .... for old Win32 systems sf-8-variant-161120-w32-sse_tp.exe .... for Win32 systems with SSE sf-8-variant-161120-x64_tp.exe ........ for old x64 systems sf-8-variant-161120-x64-popcnt_tp.exe . for 'modern' x64 systems with POPCNT The w32 compiles have been built and tested on WinXP+SP3, the rest on Win7 x64 Pro. The POPCNT build features Large Pages support - run the exec or the GUI with admin privileges to enable it. No BMI2 compile at this time. Thanks to Brice Allenbrand for Stockfish BYO. Bench: 5773672
The latest Windows exe's are now on Github
sf-sse-161120_tp_windows: https://github.com/MichaelB7/Stockfish/ ... indows.zip
sf-8-variant-161120_tp_Windows:
https://github.com/MichaelB7/Stockfish/ ... indows.zip
https://www.dropbox.com/s/umhbiturw76bz ... K.zip?dl=1
Includes:
latest developmental SF
latest SF sse version
latest SF variant (crazy house etc)
it's late , will update Github with sources tomorrow ...going to bed
-
- Posts: 919
- Joined: Tue Nov 24, 2015 9:11 pm
- Location: upstate
Re: Stockfish 8 SSE (Special Edition)
The only variants Fritz GUIs support are Chess960 (FRC) and Giveaway, AFAIK. Try Winboard instead.Rowen wrote:I cant get this crazyhouse version to work in my fritz gui ( I am still using a Fritz 12 gui ).
Presumably fritz ( 12 ) gui doesnt support crazy house ?
-
- Posts: 105
- Joined: Tue Nov 15, 2016 1:19 pm
- Location: Cheshire, England
Re: Stockfish 8 SSE (Special Edition)
Thought so. thanks for the answer.
-
- Posts: 4889
- Joined: Thu Mar 09, 2006 6:34 am
- Location: Pen Argyl, Pennsylvania
Re: Stockfish 8 SSE (Special Edition)
We have updates to both "Stockfish-8-sse" and "Stockfish-8-crazyhouse"
The "Ladder" was granualzied to now offer over 1000 different levels. Virtually every ELO level from 1200 to 2600 will now impact the number of nodes (positions) searched - but still keeping to the doubling of nodes every 100 ELO. The secret is this number "1.0069555500567" - take this number to 100th power , it is almost exactly 2. 1.0069555500567^100 ~= 2. So what that means is that can a unique number of nodes seems for virtually every ELO from 1200 to 2600 , 1200, 1201, 1202 etc. Note at the low levels, some ELOs will generate the same number of nodes to be seen, but once you play t a level over 1500 or 1600 ELO , you will see a very slight increase in the number of nodes searched with an increase of just one ELO ( ~0.69555% increase in nodes for every ELO).
in addition, for more variety of play, you now can select a range of ELO for Stockfish to play with serval new commands, "Ladder Range" (true or false ) "Ladder Lower" and "Ladder Upper". As an example if you set "Ladder Range" to be true (checked) , and select "Ladder Lower" to be 1600 and "Ladder Upper" to be 1800, on any given move it will play somewhere after seeing ~1000 to ~ 4000 nodes - so just like a human, it will play some very good moves and occasionally a clunker but not by weakening the evaluation or search ( other than reducing the number of nodes it sees. All other features previously added are maintained.
Official Stockfish: https://github.com/official-stockfish/Stockfish
MB's Github: https://github.com/MichaelB7
Stockfish sse: https://github.com/MichaelB7/Stockfish-sse
Stockfish 8 Variant: https://github.com/MichaelB7/Stockfish-8-Variant
Stockfish 8 Variant for the macOS: https://github.com/MichaelB7/Stockfish- ... -MacOS.zip
Stockfish 8 sse for macOS: https://github.com/MichaelB7/Stockfish- ... _macOS.zip
The "Ladder" was granualzied to now offer over 1000 different levels. Virtually every ELO level from 1200 to 2600 will now impact the number of nodes (positions) searched - but still keeping to the doubling of nodes every 100 ELO. The secret is this number "1.0069555500567" - take this number to 100th power , it is almost exactly 2. 1.0069555500567^100 ~= 2. So what that means is that can a unique number of nodes seems for virtually every ELO from 1200 to 2600 , 1200, 1201, 1202 etc. Note at the low levels, some ELOs will generate the same number of nodes to be seen, but once you play t a level over 1500 or 1600 ELO , you will see a very slight increase in the number of nodes searched with an increase of just one ELO ( ~0.69555% increase in nodes for every ELO).
in addition, for more variety of play, you now can select a range of ELO for Stockfish to play with serval new commands, "Ladder Range" (true or false ) "Ladder Lower" and "Ladder Upper". As an example if you set "Ladder Range" to be true (checked) , and select "Ladder Lower" to be 1600 and "Ladder Upper" to be 1800, on any given move it will play somewhere after seeing ~1000 to ~ 4000 nodes - so just like a human, it will play some very good moves and occasionally a clunker but not by weakening the evaluation or search ( other than reducing the number of nodes it sees. All other features previously added are maintained.
Official Stockfish: https://github.com/official-stockfish/Stockfish
MB's Github: https://github.com/MichaelB7
Stockfish sse: https://github.com/MichaelB7/Stockfish-sse
Stockfish 8 Variant: https://github.com/MichaelB7/Stockfish-8-Variant
Stockfish 8 Variant for the macOS: https://github.com/MichaelB7/Stockfish- ... -MacOS.zip
Stockfish 8 sse for macOS: https://github.com/MichaelB7/Stockfish- ... _macOS.zip
-
- Posts: 113
- Joined: Thu Apr 30, 2015 7:36 pm
Re: Stockfish 8 SSE (Special Edition)
We hope the kindness of T Poppins, so that compile us for windows. Thank you very much.
-
- Posts: 3410
- Joined: Sat Feb 16, 2008 7:38 am
- Full name: Peter Martan
Re: Stockfish 8 SSE (Special Edition)
Hi Diego!
There are windows-compiles at Michael's SF SSE- github- link given, you just have to cklick "view raw" to download themfantasmadel50 wrote:We hope the kindness of T Poppins, so that compile us for windows.
Peter.
-
- Posts: 3410
- Joined: Sat Feb 16, 2008 7:38 am
- Full name: Peter Martan
Re: Stockfish 8 SSE (Special Edition)
Thanks a lot, Michael!
SF SSE seems to use large pages now, doesn't it? Could you please consider to add a UCI- option to switch that off too, like there is one in Brainfish?
Having used Windows for a while you always have to restart it, if you want to let an engine compute with big hash and lp- usage, otherwise lp rather slow down the engine than making it faster
SF SSE seems to use large pages now, doesn't it? Could you please consider to add a UCI- option to switch that off too, like there is one in Brainfish?
Having used Windows for a while you always have to restart it, if you want to let an engine compute with big hash and lp- usage, otherwise lp rather slow down the engine than making it faster
Peter.