[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnubg] How best to run a large GNU v GNU session/Memory Leaks?

From: grefow
Subject: Re: [Bug-gnubg] How best to run a large GNU v GNU session/Memory Leaks?
Date: Sat, 18 Oct 2003 05:07:33 -0400

Whenever I have gnubg play a +100-game session against itself (gnubg0 vs gnubg1; match length = 1; player strength = Supremo for both)  it causes my Win98 system to become unstable. In fact, I now automatically reboot the system immediately upon finishing a gnubg session of any length. The couple of times that I let it play much longer sessions, my system locked up completely.
Here's a copy of a report previously submitted, since it's the kind of thing I typically see when gnubg craps out on me:
     bgnubg 0.14 (Jul 2 build)
     2003/10/02 3:09pm EDT
     While analysing bg00859 session.
     GNUBG caused an invalid page fault in
     module KERNEL32.DLL at 0177:bff7a4a3.
     EAX=00000010 CS=0177 EIP=bff7a4a3 EFLGS=00010206
     EBX=00000010 SS=017f ESP=00dc9270 EBP=00dc9280
     ECX=00de0038 DS=017f ESI=00de0000 FS=5f97
     EDX=00000664 ES=017f EDI=00de000c GS=0000
     Bytes at CS:EIP:
     8b 03 25 fc ff ff 0f 3b 45 0c 0f 83 81 00 00 00 
     Stack dump:
     00f18780 00000660 00000660 a2bc9b37 00dc9294 bff88d9b 00de0000 00000664 00000000 00dc92d4 78001426
     00de0000 00000000 00000660 00f18780 00000080
Is there any other kind of detail that I can provide for you when this occurs? Any way to configure gnubg so that it auto-generates it's own error log, tracks it's own memory use, etc?

> Message: 13
> Date: Tue, 07 Oct 2003 12:26:56 +0100
> From: Jon Kinsey <
> Subject: Re: [Bug-gnubg] How best to run a large GNU v GNU session
> To: Achim Mueller <
address@hidden>,bug-gnubg <address@hidden>
> Message-ID: <
> Content-Type: text/plain; charset="us-ascii"; format=flowed
> I've checked the code and there aren't any obvious memory leaks.  I was
> able to run lots of games (100,000+ with a nogui version) and the allocated
> memory (at any one time) didn't go above 10Mb.
> I didn't use the highest level of evaluation, I used the world class
> setting, so there could be something else going wrong in your particular test.
> Jon

> >
> > ===== Original Message From address@hidden =====
> > I was wondering how Joseph Heled ran his 500000 test session between
> > 0.14 and 0.13 weights because even using the no gui and game record off
> > and display computer moves off my PC with 256mb Ram really strts
> > straining after ~4500 games.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]