Colossus 2008b available

[MartinBryant]

Hello,

This version is much better than the precedent ver., but unfortunately it often crashes, ca. 1 / 6. What a pity!

Hopefully it will be quickly corrected !

Regards
Gregory

Hi Greg,
I've not known it to crash for years.
What GUI are you using?

[Le Fou numerique]

Hello,

This version is much better than the precedent ver., but unfortunately it often crashes, ca. 1 / 6. What a pity!

Hopefully it will be quickly corrected !

Regards
Gregory

Hi Greg,
I've not known it to crash for years.
What GUI are you using?

Hi,

I don't have crash in the 126 games for the "tournoi d'entrée" with Arena.

Also with Fritz 11 GUI, Colossus 2008b played 15 games in a tournament in running and no crash.

Regards,
Patrick

[mclane]

i am testing in arena and had no crash of colossus versions.
also not 2008b.

so maybe your GUI is buggy. or your PC has a problem-

[gerold]

Hello,

This version is much better than the precedent ver., but unfortunately it often crashes, ca. 1 / 6. What a pity!

Hopefully it will be quickly corrected !

Regards
Gregory

Hi Greg,
I've not known it to crash for years.
What GUI are you using?

Using it in Arena here Martin. No crashs here. Thanks again
for the fine program.

Best to you,

Gerold.

[Gregory Owett]

Hi Martin,

I used CB (F9) on an AMD dual core, Win XP64, book: Perfect 14, (other used-HS-Masterbook), (T = 2'+1") with the following results:

Colossus 08b-
- ET Chess 08 2.5 - 3.5
- Jonny 2.83x64 2.5 - 3.5
- Booot 4140 2 - 4
- PRODEO 1.6 5 - 1
- Hamsters 07 4.5 - 1.5
- Slow 2.1 3 - 3
- Alanic 7.07 4 - 2
- Frenzee Fx64 5.5 - 0.5
- Delfi 5.4 4.5 - 1.5

tot: 33.5 - 20.5 62.03%

Against Delfi 5.4 it crashed 3 times.

On another machine, (a single trial), once on 60 games.
Colossus 08b-
- ET Chess 08 3.5 - 2.5
- Jonny 2.83x64 3.5 - 2.5
- Booot 4140 2.5 - 3.5
- PRODEO 1.6 4 - 2
- Hamsters 07 4 - 2
- Pharaon 3.5 3.5 - 2.5
- Pharaoh 3.5.1 4.5 - 1.5
- 2 JA Scorpio 4.5 - 1.5
- Frenzee F 4.5 - 1.5
- Wildcat 8 2.5 - 3.5

tot: 37 - 23 61.66%

I'll check again in a few series against Delfi 5.4

[mclane]

hi martin. on the FAST machines

Code: Select all

                                        2008a     2008b
Intel core2duo 6300                     9,5/51    15/51
Intel core2quad 6600                   32,5/100   35/86

this is a BIG difference.





the new 2008b version is much better than the 2008a version.
the tournaments are not yet finished, but as you can see the tendency is similar on the faster machines. on my athlon2400+ the later version was little weaker from the result.

maybe this has to do with the changes martin did.
if they are search related it can be that they need kind of
hardware to differenciate enough.

[Jouni]

Hi Martin,

In endgametests like EET and "Echecs" 2000b is really good! BTW I see it uses TBs heavily.

Jouni

[Kaj Soderberg]

Hi Martin,

Been playing around a lot with Colossus 2008b since Friday, running 3 2008b engines at the same time on quite fast cores.

Looks really good, and what i find very nice is that C is quite creative. It for instance gives away a pawn to achieve an endgame with two rooks and a bishop and lots of pawns in order to try to draw due to the bishops of different colours, instead of staying under quite heavy positional pressure. Now that i find well implemented chess knowlegde.

The only less nice thing i have found out is that, in analysis mode, the generally low branching factor suddenly can go up steeply (to 5 or so), after say 1 to 2 hours. It looks a bit like the tree is blown at a certain depth, but you of course must have a better view on that.

Anyway, so far nice to get to know Colossus.

Best regards,

Kaj