DeepToga1.9.6 Post

Public forum

Moderator: Team

DeepToga1.9.6 Post

Postby WHMoweryJr on Tue Jun 23, 2009 2:54 pm

Hello World:

DT1.9.6 is here.
Thanks to JA for the build.
It should be fast.


Bill

P.S. Pls send any feedback that you like to: WHMoweryJr
Attachments
DeepToga1.9.6.7z
(611.94 KiB) Downloaded 566 times
WHMoweryJr
 
Posts: 603
Joined: Fri Apr 11, 2008 7:32 pm
Location: Ellenton FL

Re: DeepToga1.9.6 Post

Postby Solis on Tue Jun 23, 2009 7:07 pm

Hi Bill,
I have installed new Toga on Arena interface and it is not working properly.And I think it also affected other Deep Togas on my list.
First, in the engine matches it would start and then stop somewhere around depth 17 or 18 while the count is going.Also the numbers are half of what they should be while it shows 100% use on the Quad that I have.Than the numbers go up while nothing is happening.
But is seems to me that the same bug kind of extended to other Deep Togas on my list of engines.
Lost couple of hours trying to see if it will work but no luck.
Let's see if other people have the same kind of problem.
At least I can say that TTK is doing just fine.Very nice games so far.
Thanks for your work and attention.
Solis
User avatar
Solis
 
Posts: 358
Joined: Sat Jan 10, 2009 4:05 pm

Re: DeepToga1.9.6 Post

Postby Solis on Wed Jun 24, 2009 3:32 am

Here are some more things that I have noticed.
Deep Toga works fine on dual core but not with quad.I have tried different configurations .The engine would only start when number of threads was set to 4 as well as threads per depth also on 4.After exiting the book moves it would start calculating and maybe making one move after long thinking and than after following move it would continue to calculate without move execution.Attempts to make a move are usually unsuccessful and after managing to start new game the process of calculation would be persistent with use of CPU .Few resets are usually necessary to stop the process.
Similar things are seen when set for engine matches
This looks similar to problems we had with one of the previous Toga engines on dual core.That problem was not fixed.
Maybe this report would be helpful to fix above mentioned problems.And to see if other people had the same problem on 4 cores.
Also any suggestion about engine configuration.I have tried what I have known but right information always helps.
Thanks,
Solis
User avatar
Solis
 
Posts: 358
Joined: Sat Jan 10, 2009 4:05 pm

Re: DeepToga1.9.6 Post

Postby souk on Thu Jun 25, 2009 12:18 am

I've seen crashes with 2 and 4 core's. It seems to run great in single core mode :D . The toga logs are'nt showing anything abnormal but I'll upload them anyway.
1.9.6 just seems to stop sending moves and crashes, and doesnt fully unload from chessbase gui with out mannually killing it in the task manager. All testing was done in chessbase 8 gui with no books, no tablebases, and no bitbases. Is there anyway we could get a build with the old pv just to rule out that the bug isnt in the new pv? If there is a free compiler I could use to help debug this please suggest one with link. There must be a good way to chase down this crashing mp bug .
Attachments
toga logs.rar
(34.14 KiB) Downloaded 225 times
souk
 
Posts: 88
Joined: Wed Mar 18, 2009 10:33 pm

Re: DeepToga1.9.6 Post

Postby cirebonb on Thu Jun 25, 2009 9:11 am

hmmm.

full new depth on pv is nice idea, i should try to..
great bill!
cirebonb
 
Posts: 19
Joined: Tue Aug 05, 2008 1:56 am

Re: DeepToga1.9.6 Post

Postby Damir Desevac on Thu Jun 25, 2009 7:51 pm

As Souk already pointed out this MP bug has been going on for quite some time, whether it might be with TBS being used or not, and I find that really annoying.
Today my engine crashed several times. After I changed couple of parameters the engine would not move, and crashed whether it was still in the book or out of the book, no move was delivered :x :x
Fortunately my opponent was kind enough to abort all 4-5 games so I did not lose on time...., those 5 games were all tried to be played on long time control (60+0)
PS: I think the whole Toga code should be written from scratch, as this can no longer go on like this. Better 1 good stable version regardless of playing strength than 100 bad with increased strenth of play......
Regards
Damir
Damir Desevac
 
Posts: 122
Joined: Sat Aug 16, 2008 7:25 pm

Re: DeepToga1.9.6 Post

Postby souk on Thu Jun 25, 2009 8:11 pm

I think the new pv might still be in the beta stage. In the read me file It says the old pv was bug free so my guess is the bug must be in the new pv. Lets be pacient and just be blessed we have a chance to have some part of the new toga's .
souk
 
Posts: 88
Joined: Wed Mar 18, 2009 10:33 pm

Re: DeepToga1.9.6 Post

Postby Denis P. Mendoza on Thu Jun 25, 2009 10:12 pm

:lol: Patience Damir! Even from the start, Toga is already buggy. I feel the same way you do, and there were times I already wanted to quit because of many other reasons aside from this. But this is what we have at the moment, so I just have to accept, wait and hope for the best. Cheer-up pal! This rusty engine really needs some major tuneup. And you two will be the test drivers when it's ready - like always ;) !

But talking of stable Toga derivatives, Cyclone is the one code that I feel has stability. compiles clean with no warnings. I haven't got a crash too when I had used it before. I don't know about Grapefruit code, but code is just similar to the ones here. But Vadim made a good job in giving Toga a boost in speed. Any news from your friend? He's been hibernating lately.

Anyway, just keep Bill informed on these bugs. His code is a liitle bit mixed-up, but he's working on it. There's a big difference in Bill's work as it started on his LearninToga series (from Toga 1.2.1a) way before Toga 1.4 beta5c. Thomas only added MP (from beta5c) on his code before he left.

But I agree that the more you add features, the more it becomes complicated. But because he has not much time to test, he distributes it to us to receive inputs (good or bad) in order to continue his work. The development is not always a success. It's a bit slow, but as long as somebody is working, we just need to support or it will stop and be forgotten. Let's just accept the truth and move on.
Denis P. Mendoza
 
Posts: 876
Joined: Sat Apr 12, 2008 1:33 am

Re: DeepToga1.9.6 Post

Postby Damir Desevac on Fri Jun 26, 2009 10:13 am

Sorry Denis, I really am getting frustrated with all these releases, and none of them work as they should.
I want to see Toga play as normal chess as possible without any crashes, but you re right we have to take sweet with the bitter sweet as the code is still new and needs some polishing.
I hope one day Toga can play without any problems.
Damir Desevac
 
Posts: 122
Joined: Sat Aug 16, 2008 7:25 pm

Re: DeepToga1.9.6 Post

Postby Damir Desevac on Fri Jun 26, 2009 10:19 am

PS: I don't know what happened to Vadim, haven't heared from him in months. His last release date of GrapeFruit is dated 31 January this year,
so it has been some time I heared something from him.
The good news is it seems that Norm is working on a newer version of Cyclone 4 release date ~ 7/15/2009.
I also hope Vadim too comes back and start working on a new GrapeFruit. It has been some time he released a new Version...
Damir Desevac
 
Posts: 122
Joined: Sat Aug 16, 2008 7:25 pm

Next

Return to Public

Who is online

Users browsing this forum: No registered users and 1 guest

cron