Jumbo 0.6.62 is out

[joechess]

Hi Sven works normally from commanline

C:\Jumbo-0.6.62

C:\Jumbo-0.6.62>jumbo64-0.6.62-bb-sse41.exe
# JumboBB 0.6.62
# (c) 2016-18 by Sven Schuele, Berlin, Germany
#
feature done=0
# new size of TT hash: 256 MB
# new size of Pawn hash: 32 MB
# rnbqkbnr
# pppppppp
# ........
# ........
# ........
# ........
# PPPPPPPP
# RNBQKBNR
#
# Your move:

Details about my computer
Windows 7
AMD E-350D APU with Radeon HD Graphics (supports MMX (+), SSE, SSE2, SSE3, SSSE3, SSE4A, x86-64, AMD-V)
4.0GB RAM

EDIT----------------------------------------------------------------------------------------------------------------------------------------------
Ok tried it under winboard it appears to be working, played a full game no problems.
So went back to Arena changed the setting to Winboard(old) seemed to load ok, made its first move, then stuck again on second move with something like 296mb memory.

[Sven]

So went back to Arena changed the setting to Winboard(old) seemed to load ok, made its first move, then stuck again on second move with something like 296mb memory.

Thanks for the details. "Stuck" like in "hanging", or "crashed"? Is it possible to get a debug log from Arena that shows the engine output? And what exactly do you mean by "Winboard(old)"? Is the behaviour different from configuring it like "Winboard(not old)"? (After thinking more about it I can imagine that Arena is able to communicate to engines via the ancient WinBoard protocol version 1 which would be kind of surprising for Jumbo ...)

Have you set one of these options in the GUI?
ttHashSizeMB
pawnHashSizeMB

If you have then I would expect Jumbo to always crash (also with many older versions ...).

[Sven]

I found out the following:

1) All Jumbo versions 0.6.35 and later (up to 0.6.62) crash immediately on startup when trying to process one of the configuration options "ttHashSizeMB" or "pawnHashSizeMB" from "jumbo.ini" (if such an option is not commented out). The option value does not matter.

2) All Jumbo versions 0.6.35 and later (up to 0.6.62) crash when trying to process one of the configuration options "ttHashSizeMB" or "pawnHashSizeMB" passed from the GUI (via "option" protocol command), provided the new value is smaller than the default value (256 resp. 32).

Bugfix is on its way!

[Sven]

I have just released Jumbo 0.6.63 containing bugfixes for the problems mentioned above.

http://www.schuelebln.de/chess/jumbo

[Gabor Szots]

I have just released Jumbo 0.6.63 containing bugfixes for the problems mentioned above.

http://www.schuelebln.de/chess/jumbo

Thank you, Sven. My problem has gone.

[joechess]

Ok i ran the the new sse.41 bb .exe in Arena and here is the log file

2018-11-01 13:48:48.189*1*-----------------Starting engine 1 Jumbo64-0.6.63-bb-sse41------------------
2018-11-01 13:48:48.189*1*Configured Engine 1 Type: WB2
2018-11-01 13:48:48.189*1*Engine 1 dir: C:\Users\Joe\Documents\ChessBase\Chess Engines\WB & UCI ENGINE\Jumbo-0.6.63
2018-11-01 13:48:48.189*1*Engine 1 commandline: "C:\Users\Joe\Documents\ChessBase\Chess Engines\WB & UCI ENGINE\Jumbo-0.6.63\jumbo64-0.6.63-bb-sse41.exe"
2018-11-01 13:48:48.299*1*Engine 1 ProcessID: 2612
2018-11-01 13:48:48.299*1*Engine 1 Prio:32 ThreadPrio:0
2018-11-01 13:48:48.299-->1:xboard
2018-11-01 13:48:48.299-->1:protover 2
2018-11-01 13:48:48.345<--1:# JumboBB 0.6.63
2018-11-01 13:48:48.345<--1:# (c) 2016-18 by Sven Schuele, Berlin, Germany
2018-11-01 13:48:48.345<--1:#
2018-11-01 13:48:48.345<--1:feature done=0
2018-11-01 13:48:48.345-->1:accepted done
2018-11-01 13:48:48.345<--1:# new size of TT hash: 256 MB
2018-11-01 13:48:48.767<--1:# new size of Pawn hash: 32 MB
2018-11-01 13:48:48.829<--1:# rnbqkbnr
2018-11-01 13:48:48.829<--1:# pppppppp
2018-11-01 13:48:48.860<--1:# ........
2018-11-01 13:48:48.860<--1:# ........
2018-11-01 13:48:48.891<--1:# ........
2018-11-01 13:48:48.907<--1:# ........
2018-11-01 13:48:48.907<--1:# PPPPPPPP
2018-11-01 13:48:48.907<--1:# RNBQKBNR
2018-11-01 13:48:48.907<--1:#
2018-11-01 13:48:48.907<--1:# Your move:
2018-11-01 13:48:48.907<--1:feature ping=1
2018-11-01 13:48:48.907-->1:accepted ping
2018-11-01 13:48:48.907<--1:feature setboard=1
2018-11-01 13:48:48.907-->1:accepted setboard
2018-11-01 13:48:48.907<--1:feature playother=1
2018-11-01 13:48:48.907-->1:accepted playother
2018-11-01 13:48:48.923<--1:feature usermove=1
2018-11-01 13:48:48.923-->1:accepted usermove
2018-11-01 13:48:48.923<--1:feature colors=0
2018-11-01 13:48:48.923-->1:accepted colors
2018-11-01 13:48:48.923<--1:feature sigint=0
2018-11-01 13:48:48.923-->1:accepted sigint
2018-11-01 13:48:48.954<--1:feature sigterm=0
2018-11-01 13:48:48.954-->1:accepted sigterm
2018-11-01 13:48:48.985<--1:feature analyze=1
2018-11-01 13:48:48.985-->1:accepted analyze
2018-11-01 13:48:48.985<--1:feature name=0
2018-11-01 13:48:48.985-->1:accepted name
2018-11-01 13:48:49.016<--1:feature nps=0
2018-11-01 13:48:49.016-->1:accepted nps
2018-11-01 13:48:49.047<--1:feature debug=1
2018-11-01 13:48:49.047-->1:accepted debug
2018-11-01 13:48:49.079<--1:feature egt="gaviota"
2018-11-01 13:48:49.079-->1:accepted egt
2018-11-01 13:48:49.110<--1:feature smp=1
2018-11-01 13:48:49.110-->1:accepted smp
2018-11-01 13:48:49.141<--1:feature variants="normal"
2018-11-01 13:48:49.141-->1:accepted variants
2018-11-01 13:48:49.172<--1:feature option="ttHashSizeMB -spin 0 1 2048"
2018-11-01 13:48:49.172-->1:accepted option
2018-11-01 13:48:49.172<--1:feature option="pawnHashSizeMB -spin 0 1 2048"
2018-11-01 13:48:49.172-->1:accepted option
2018-11-01 13:48:49.172<--1:feature option="gtbCacheSizeMB -spin 16 0 256"
2018-11-01 13:48:49.172-->1:accepted option
2018-11-01 13:48:49.172<--1:feature option="gtbPath -path C:/gtb"
2018-11-01 13:48:49.172-->1:accepted option
2018-11-01 13:48:49.203<--1:feature option="bookFilePath -path book.bin"
2018-11-01 13:48:49.203-->1:accepted option
2018-11-01 13:48:49.235<--1:feature myname="JumboBB 0.6.63"
2018-11-01 13:48:49.235<--1:feature done=1
2018-11-01 13:48:49.235-->1:accepted done
2018-11-01 13:48:49.297-->1:egtpath gaviota C:\Users\Joe\Documents\ChessBase\TB\Gaviota4Men
2018-11-01 13:48:49.297-->1:cores 1
2018-11-01 13:48:49.297-->1:new
2018-11-01 13:48:49.297-->1:random
2018-11-01 13:48:49.297-->1:level 40 2 0
2018-11-01 13:48:49.297<--1:# initializing TB: type=gaviota path=C:\Users\Joe\Documents\ChessBase\TB\Gaviota4Men
2018-11-01 13:48:49.297-->1:post
2018-11-01 13:48:49.297-->1:hard
2018-11-01 13:48:49.297-->1:easy
2018-11-01 13:48:49.297-->1:ping 4
2018-11-01 13:48:49.313<--1: 2 [main] jumbo64-0.6.63-bb-sse41 2612 cygwin_exception::open_stackdumpfile: Dumping stack trace to jumbo64-0.6.63-bb-sse41.exe.stackdump
2018-11-01 13:49:00.123-->1:?
2018-11-01 13:49:00.123-->1:force
2018-11-01 13:49:00.123-->1:quit


And here is the stack dump file that i didnt realise was created the last time i posted in this thread

Exception: STATUS_ILLEGAL_INSTRUCTION at rip=001004349DC
rax=00000001004F8000 rbx=00000001004F7B60 rcx=0000000000000000
rdx=000000010065B970 rsi=0000000600081140 rdi=000000010065B960
r8 =000000010065B9A0 r9 =0000000000000001 r10=00000001004F7B80
r11=00000000FFFFFFEF r12=00000006000810B0 r13=0000000000000001
r14=000000010065B960 r15=00000000FFFF9BB0
rbp=000000010065DCF8 rsp=00000000FFFF96C0
program=C:\Users\Joe\Documents\ChessBase\Chess Engines\WB & UCI ENGINE\Jumbo-0.6.63\jumbo64-0.6.63-bb-sse41.exe, pid 2268, thread main
cs=0033 ds=002B es=002B fs=0053 gs=002B ss=002B
Stack trace:
Frame Function Args
0010065DCF8 001004349DC (00300000004, 006000810A0, 006000810E8, 000FFFFA030)
000FFFFA060 0010040E0F3 (00000000002, 003C8EDFF60, 001801197EB, 000FFFF9C30)
000FFFFA030 0010040F1AE (00000000000, 00000000080, 00000000000, 00000000000)
003C8EDFF60 0010044488E (00000000020, FF0700010302FF00, 00180047B51, 000FFFFCE00)
000FFFFCCB0 00180047BC2 (00000000000, 00000000000, 00000000000, 00000000000)
00000000000 00180045863 (00000000000, 00000000000, 00000000000, 00000000000)
000FFFFFFF0 00180045914 (00000000000, 00000000000, 00000000000, 00000000000)
End of stack trace

BTW the Winboard(old) option in arena just means starting an engine in the old winboard 1 option for older engines.
haven't done a log or stackdump file for that yet. i figured you would have enough info above?

[Sven]

Thanks for the details. So 0.6.63 does not run under Arena on your hardware either using the "sse41" binary. Could you try the "normal" binary without "sse41", i.e. jumbo64-0.6.63-bb.exe, and check whether the behaviour is the same? In an earlier post you showed your hardware details and that did not look like "old hardware" so you would not need the "sse41" binary.

[joechess]

Ok so here is the normal jumbo64-0.6.63-bb.exe under winboard in arena

2018-11-01 22:16:34.598 Arena 3.5.1
2018-11-01 22:17:05.580*1*--------------------Starting engine 1 Jumbo64-0.6.63-bb---------------------
2018-11-01 22:17:05.580*1*Configured Engine 1 Type: Auto
2018-11-01 22:17:05.595*1*Engine 1 dir: C:\Users\Joe\Documents\ChessBase\Chess Engines\WB & UCI ENGINE\Jumbo-0.6.63
2018-11-01 22:17:05.595*1*Engine 1 commandline: "C:\Users\Joe\Documents\ChessBase\Chess Engines\WB & UCI ENGINE\Jumbo-0.6.63\jumbo64-0.6.63-bb.exe"
2018-11-01 22:17:05.704*1*Engine 1 ProcessID: 2288
2018-11-01 22:17:05.704*1*Engine 1 Prio:32 ThreadPrio:0
2018-11-01 22:17:05.704-->1:xboard
2018-11-01 22:17:05.736<--1: 1 [main] jumbo64-0.6.63-bb 2288 cygwin_exception::open_stackdumpfile: Dumping stack trace to jumbo64-0.6.63-bb.exe.stackdump
2018-11-01 22:17:05.736-->1:uci
2018-11-01 22:17:13.738-->1:isready
2018-11-01 22:17:14.534-->1:new
2018-11-01 22:17:14.534-->1:random
2018-11-01 22:17:14.534-->1:level 40 2 0
2018-11-01 22:17:14.534-->1:post
2018-11-01 22:17:14.534-->1:hard
2018-11-01 22:17:14.534-->1:easy

and here is the stackdump file created

Exception: STATUS_ILLEGAL_INSTRUCTION at rip=00100443045
rax=00000000FFFFA038 rbx=00000000FFFFCBD0 rcx=00000000000062E8
rdx=000000060001803F rsi=0000000000000000 rdi=0000000000000000
r8 =0000000000000000 r9 =0000000000000000 r10=00000000FFFF99F0
r11=0000000000000000 r12=0000000000000001 r13=00000000FFFFCBD0
r14=0000000000000000 r15=00000001802193F0
rbp=0000000000000000 rsp=00000000FFFF9B30
program=C:\Users\Joe\Documents\ChessBase\Chess Engines\WB & UCI ENGINE\Jumbo-0.6.63\jumbo64-0.6.63-bb.exe, pid 2288, thread main
cs=0033 ds=002B es=002B fs=0053 gs=002B ss=002B
Stack trace:
Frame Function Args
00000000000 00100443045 (00000000020, FF0700010302FF00, 00180047B51, 000FFFFCE00)
000FFFFCCB0 00180047BC2 (00000000000, 00000000000, 00000000000, 00000000000)
00000000000 00180045863 (00000000000, 00000000000, 00000000000, 00000000000)
000FFFFFFF0 00180045914 (00000000000, 00000000000, 00000000000, 00000000000)
End of stack trace

[Guenther]

Thanks for the details. So 0.6.63 does not run under Arena on your hardware either using the "sse41" binary. Could you try the "normal" binary without "sse41", i.e. jumbo64-0.6.63-bb.exe, and check whether the behaviour is the same? In an earlier post you showed your hardware details and that did not look like "old hardware" so you would not need the "sse41" binary.

Hi Sven, he showed that SSE4A is his machines highest instructions level (which is below 4.1).

[Sven]

Thanks for the details. So 0.6.63 does not run under Arena on your hardware either using the "sse41" binary. Could you try the "normal" binary without "sse41", i.e. jumbo64-0.6.63-bb.exe, and check whether the behaviour is the same? In an earlier post you showed your hardware details and that did not look like "old hardware" so you would not need the "sse41" binary.

Hi Sven, he showed that SSE4A is his machines highest instructions level (which is below 4.1).

You are right, I did not read correctly, I "dreamt" having read something like 4.2.

Does 0.6.63 also crash for you?