Stupid error, crazy performance boost

Discussion of chess software programming and technical issues.

Moderators: hgm, Rebel, chrisw

User avatar
xr_a_y
Posts: 1871
Joined: Sat Nov 25, 2017 2:28 pm
Location: France

Stupid error, crazy performance boost

Post by xr_a_y »

After months of diagnostic, bug tracking and tuning on Weini, trying to get some strength, my eyes became blind to mostly everything ... So I used CLOP a little, and added a few elo. But nothing spectacular at all ...
Yesterday, the single threaded version of Weini, with ponder on, was "only" +150elo (69%) against fairy-max at bullet or atomic TC but was loosing badly versus Isa or Dorpsgek.

So I decided I shall take some time to read again old, basic, stuff in the code... And of course, I've just discovered something bad, very bad. In QSearch, move generation was performed too soon, before standard cut-offs... Having fixed this, I ran a match versus fairy-max : the speed boost is quite interesting of course : more than 30% in knps ; but the elo gain seems crazy : now around >80% winrate (>250elo) versus fairy-max.

Some days are needed to ensure everything is correct, but that was clearly the silly little kind of things I was looking for for months ... I feel a little better now !
User avatar
CMCanavessi
Posts: 1142
Joined: Thu Dec 28, 2017 4:06 pm
Location: Argentina

Re: Stupid error, crazy performance boost

Post by CMCanavessi »

Nice, a new engine! What's the current estimated strength? I might include it in next season's CCLS.
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls
User avatar
xr_a_y
Posts: 1871
Joined: Sat Nov 25, 2017 2:28 pm
Location: France

Re: Stupid error, crazy performance boost

Post by xr_a_y »

I'd be glade to participate. Weini should be around 2200 for now.
tpoppins
Posts: 919
Joined: Tue Nov 24, 2015 9:11 pm
Location: upstate

Re: Stupid error, crazy performance boost

Post by tpoppins »

No Windoze bins? I've been unable to compile with Cygwin. Stuck at the Cmake stage, with an error about /usr/bin/c++.exe being broken.
Tirsa Poppins
CCRL
Dann Corbit
Posts: 12537
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Stupid error, crazy performance boost

Post by Dann Corbit »

It uses backtrace() which is Linux only and it also has lots of template substitutions that are not legal.
No way to compile it under Windows.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
User avatar
xr_a_y
Posts: 1871
Joined: Sat Nov 25, 2017 2:28 pm
Location: France

Re: Stupid error, crazy performance boost

Post by xr_a_y »

No issue with visual studio here. The traceback call is inside a #ifdef. Indeed, I've never tested cygwin. You can probably remove the content of backtrace function if requiered. I'll do something for the next release for cygwin.
User avatar
xr_a_y
Posts: 1871
Joined: Sat Nov 25, 2017 2:28 pm
Location: France

Re: Stupid error, crazy performance boost

Post by xr_a_y »

Dann Corbit wrote: Tue Jul 24, 2018 8:31 pm it also has lots of template substitutions that are not legal.
No way to compile it under Windows.
Can you elaborate a little about this template issue please ?
Sven
Posts: 4052
Joined: Thu May 15, 2008 9:57 pm
Location: Berlin, Germany
Full name: Sven Schüle

Re: Stupid error, crazy performance boost

Post by Sven »

xr_a_y wrote: Tue Jul 24, 2018 10:42 pm
Dann Corbit wrote: Tue Jul 24, 2018 8:31 pm it also has lots of template substitutions that are not legal.
No way to compile it under Windows.
Can you elaborate a little about this template issue please ?
On the Github page you have linked to, I only see a README file, was that intended?
Sven Schüle (engine author: Jumbo, KnockOut, Surprise)
Dann Corbit
Posts: 12537
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Stupid error, crazy performance boost

Post by Dann Corbit »

xr_a_y wrote: Tue Jul 24, 2018 10:42 pm
Dann Corbit wrote: Tue Jul 24, 2018 8:31 pm it also has lots of template substitutions that are not legal.
No way to compile it under Windows.
Can you elaborate a little about this template issue please ?
g++ output is attached.
You can get a stack trace like this in Windows:
https://stackoverflow.com/questions/620 ... unning-app
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
Sesse
Posts: 300
Joined: Mon Apr 30, 2018 11:51 pm

Re: Stupid error, crazy performance boost

Post by Sesse »

All the compile errors complain that you have defined a backtrace() that returns void, it seems. It complains on this line over and over again:

Code: Select all

             (*_os) << backtrace() << std::endl;
Nothing illegal about that, as long as backtrace() returns something that can be printed (ie., not void).