New J.A Compiles

[Joseph]

Hi jim,

Thank you so much for compiling the latest development versions of Clarity and fixed version of Peacekeeper! I truly appreciate the time and effort you’ve put into resolving these issues on Android user. I’m about to test both engines now and will let you know how they perform. thanks for everything!

Your support means a lot—thanks again for all your help!

Best regards,
[Joseph]

[Joseph]

Hi Jim,

Here are the current test results for Peacekeeper and Clarity. Unfortunately, the issue with Peacekeeper still persists, but I’m happy to report that Clarity is working fine.

For my first test, I used the Normal book instead of the UHO book. Peacekeeper stopped working once, which I noticed after reaching 9 games without using UHO. I then restarted the test using Stefan Pohl’s UHO book. After reaching 14 games, I checked the games and saw that Peacekeeper had stopped working again—this time in a clearly winning position. It resulted in a time forfeiture and a loss.

Here are the standings:

1. Seer 2.8.0 Archimedes
4.5 (+4 =1 -1)


2. Clarity 7.2.0-dev-7aaab23 JA
3.0 (+1 =4 -1)


3. Koivisto 9.0 Archimedes
2.5 (+2 =1 -1)


4. Peacekeeper 3.01-dev JA
2.0 (+1 =2 -1)


5. Igel 3.6.0 JA
1.5 (+1 =1 -2)


6. Arasan 25.0 JA
0.5 (+0 =1 -3)



Let me know if there’s anything else I can try or if you need more details. Thank you again for your hard work—Clarity is performing wonderfully!

Best regards,
[Joseph]

[Jim Ablett]

https://github.com/WillChing01/Pingu
Pingu 4.0.0 by William Ching

Pingu 4.0.0 64 JA

Windows/Linux & Android profiled builds.

https://drive.proton.me/urls/1ZDVG9V2NG#f0TdZ3gwtSPw

Jim.

Update:

Pingu 5.0.0 64 JA

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/VT0GD4T534#s0lcQpPnKwzA

Jim.

[Jim Ablett]

https://github.com/Jochengehtab/Schoenemann
Schoenemann 0.1.3 by Jochen Gehtab

Schoenemann 0.1.3 64 JA

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/S3RS5JGYC0#GPZDjBCNRZ8L

Jim.

Update:

Schoenemann 0.2.0 64 JA

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/RZV8G6W3PW#FjjkPjIRxF5q

Jim.

Update:

Schoenemann 0.4.0 64 JA

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/R9W89EDTYG#gvC8xlM6ANiW

Jim.

[Jim Ablett]

https://github.com/Quanticade/Quanticade
Quanticade 0.8 dev.fd359fb by DarkNeutrino

Quanticade 0.8 dev.fd359fb 64 JA

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/GV05GYQ66R#XZHi38Xbt2RQ

Jim.

Update:

Quanticade 0.9 Electra 64 JA

Windows/Linux & Android clang profiled builds

https://drive.proton.me/urls/FP4A4SP5G0#7RHD7eCvZOHX

Jim.

Update:

Quanticade Chimera 1.1 64 JA

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/FRFTVK9CSG#7J5hLJ8zNZQ9

Jim.

[Carbec]

Hello Jim

Happy new year !
Thanks for the compiles. I can't test the AVX512, but the BMI2 is really faster than mine.

But, with windows, there is a bug I already met, and never could answer.
After launching zangdar, enter this : run q 45.
This starts a search on this position : "2r1r1k1/pp1q1ppp/3p1b2/3P4/3Q4/5N2/PP2RPPP/4R1K1 w - - bm Qg4", with depth_max 45.
The search stops after depth=42 most of the time, and zangdar quits, with no message.
That's why I compile with gcc.
I tried the option "-fsanitize=address", and got the error ERROR: AddressSanitizer: stack-overflow on address 0x7ffd034e5592. Even in debug mode, I didn't got the name of the routine that causes the problem.
If some one has an idea, I would be very happy.

Philippe

[Jim Ablett]

Hello Jim

Happy new year !
Thanks for the compiles. I can't test the AVX512, but the BMI2 is really faster than mine.

But, with windows, there is a bug I already met, and never could answer.
After launching zangdar, enter this : run q 45.
This starts a search on this position : "2r1r1k1/pp1q1ppp/3p1b2/3P4/3Q4/5N2/PP2RPPP/4R1K1 w - - bm Qg4", with depth_max 45.
The search stops after depth=42 most of the time, and zangdar quits, with no message.
That's why I compile with gcc.
I tried the option "-fsanitize=address", and got the error ERROR: AddressSanitizer: stack-overflow on address 0x7ffd034e5592. Even in debug mode, I didn't got the name of the routine that causes the problem.
If some one has an idea, I would be very happy.

Philippe

Hi Philippe,
Happy New year.

I just tested by increasing the stack size with this compiler switch and it seems to solve the problem >

Code: Select all

set stack for windows >

-Wl,--stack,16777216


set stack size for linux/arm   >

-Wl,-z,stack-size=16777216

Jim.

[Carbec]

Hello Jim,

Many thanks for the answer.
I would add that this is an option for the linker.
This problem is present only for windows and clang++. I don't know why, but im not expert.

Philippe

[Jim Ablett]

https://github.com/Carbecq/Zangdar
Zangdar 2.30 by Philippe Chevalier

Zangdar 2.30 64 JA

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/EPHJ76YF5R#zbeSW5PGQJG7

Jim.

Update:

Zangdar 2.31.04 64 JA

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/SEWZKM0KER#at9nR5dyXqSp

Jim.

Zangdar 3.04 64 JA

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/K6K6M2YBFC#WLRs2B4XyIcD

Jim.

I have updated the download package to fix bug in Windows builds.

same link >

https://drive.proton.me/urls/K6K6M2YBFC#WLRs2B4XyIcD

Jim.

[Jim Ablett]

Looks like its not working correctly. How does it play?
Jim.

I make no game - I think search is broken, nps are tooo high.

Time for me to get another pc I think so I can test what I am compiling

Jim.

I've got some new avx512 capable hardware now
Here's a new Windows avx512 build of Igel 3.6.0. In my tests slightly faster than the bmi2 compile.
Not by much, but every little bit helps

https://drive.proton.me/urls/0T4X53HHFW#I9HUxb4wUZ4B

Jim.