Daydreamer 1.7

[Aaron Becker]

The original rc2 binary I posted contains a small bug that can cause gtbs to fail to load in certain situations. I've replaced it with a corrected binary, and updated the tarball I posted to match it.

[Elorejano]

Congratulations
rc2 its not a candidate anymore!
polyglots books: works perfect (sorry, Sedat)
ctg books various: works perfect
egbb: works
gtb: works with "Load tablebases in a separate thread=false", exists somebody that have one CPU and ponder=off in this world.
Well, Aaron, you have one of the most complete engines , if you fort this to Winboard protocol (at least as Don Daly plan), you will have the most fans
One request: the known distance to mate issue, the nice are to see Mate in 33
So, till now, polyglots logs are not needed. Good work
Cheers

[swami]

http://sites.google.com/site/strategict ... st-results

Daydreamer 1.74

674/1000.

7650 / 10000 in partial credit calculations

4th Rank from 53 Engines.

Output from STS Stat 2.5

Code: Select all

-------------------------------------------------------------------------------------------------------------------------------------------------------------------
|  STS1 |  STS2 |  STS3 |  STS4 |  STS5 |  STS6 |  STS7 |  STS8 |  STS9 |  STS10 |
-------------------------------------------------------------------------------------------------------------------------------------------------------------------
Daydreamer_175_rc1 |    74 |    67 |    69 |    71 |    76 |    66 |    71 |    49 |    59 |    72 |
-------------------------------------------------------------------------------------------------------------------------------------------------------------------
Grade |    A- |    B+ |    B+ |    A- |    A |    B+ |    A- |    D |    C+ |    A- |
Score / 1000 |    811 |    718 |    811 |    798 |    799 |    824 |    782 |    610 |    696 |    801 |
Grade |    A+ |    A- |    A+ |    A |    A |    A+ |    A |    B |    B+ |    A+ |
Best Move :  674 / 1000  67.4 %  Grade : B+
Total :           7650 / 10000  76.5 %  Grade : A
-------------------------------------------------------------------------------------------------------------------------------------------------------------------

PS: Daydreamer 1.61 scored 629/1000 and was ranked 24th from 53 Engines.

[Aaron Becker]

Congratulations
rc2 its not a candidate anymore!
polyglots books: works perfect (sorry, Sedat)
ctg books various: works perfect
egbb: works
gtb: works with "Load tablebases in a separate thread=false", exists somebody that have one CPU and ponder=off in this world.
Well, Aaron, you have one of the most complete engines , if you fort this to Winboard protocol (at least as Don Daly plan), you will have the most fans
One request: the known distance to mate issue, the nice are to see Mate in 33
So, till now, polyglots logs are not needed. Good work
Cheers

Great! Thanks for your efforts testing this Carlos. You're a big help. I just had a good idea for fixing the distance to mate issue, so I think that can be included in 1.75 as well.

[Aaron Becker]

http://sites.google.com/site/strategict ... st-results

Daydreamer 1.74

674/1000.

7650 / 10000 in partial credit calculations

4th Rank from 53 Engines.

PS: Daydreamer 1.61 scored 629/1000 and was ranked 24th from 53 Engines.

Wow, this is a much better result than I expected. The partial results are especially good. If Daydreamer can come anywhere close to equalling this result in direct playing strength, I'll be extremely happy. Thanks very much for your testing, Swami.

PS: I'll add that I figured out the reason that Arena was displaying -M500 scores when verbosity level is increased and I fixed the problem, so in new versions of Daydreamer you should be able to get nicely formatted pv outputs again by increasing the verbosity to medium.

[Werner]

Ok, I think I have this ctg problem figured out. A move can be both recommended (normally this means it's marked green), and given an annotation with a '?' (including ?! and !?). In these cases the move will be marked blue, but Daydreamer will still treat these moves as if they were green. I think that if I change it so that it treats these moves as non-recommended, we'll get the behavior we want.

I think I also figured out the gtb problem you identified, Werner. I'll code up these fixes and release a new candidate version as soon as I get the time.

edit: I should also add that I've removed the dependency on pthreads, so these 1.75 releases should not require the pthreads dll to be present.

Hi Aaron,
just tested RC2 candidate here on my dual:
- now Daydreamer wins that 5men endgame
- when I load Daydreamer at once 50% CPU usage is shown!
- when Daydreamer calculates taskmanager shows 100% CPU usage (2CPUs) - although load tbs into a separate thread is not checked
- it seems Daydreamer gives the CPU free when another engine (opponent with 2CPUs or from another match) is running.
- is it possible to stop that thread using CPU time when Daydremer is just loaded??

[Aaron Becker]

Hi Aaron,
just tested RC2 candidate here on my dual:
- now Daydreamer wins that 5men endgame
- when I load Daydreamer at once 50% CPU usage is shown!
- when Daydreamer calculates taskmanager shows 100% CPU usage (2CPUs) - although load tbs into a separate thread is not checked
- it seems Daydreamer gives the CPU free when another engine (opponent with 2CPUs or from another match) is running.
- is it possible to stop that thread using CPU time when Daydremer is just loaded??

Hi Werner, thanks for this report. I'm pretty sure I understand the problem, but I can't test my solution directly because I don't have a multicore Windows machine. So, this solution is somewhat experimental. I can at least say that it solves the problem of Daydreamer using CPU time when it's not being actively used. The new release candidate with this fix applied is available here.

[Werner]

Hi Aaron,
this was a very fast answer!!

And RC3 is working for me ok so far with the CPU usage.
Maybe you do not need the option "Load tbs in a separate thread" anymore - as I see not difference when checked (I do not need this option )

[swami]

Wow, this is a much better result than I expected. The partial results are especially good. If Daydreamer can come anywhere close to equalling this result in direct playing strength, I'll be extremely happy. Thanks very much for your testing, Swami.

I haven't tested Daydreamer in engine vs engine matches. I'd like to just to see how clearly the result parallels the one that was posted.

I'd think that it's on par with Ruffian 1.05, Movei, Bugchess, Delfi, Aristarch.

PS: I'll add that I figured out the reason that Arena was displaying -M500 scores when verbosity level is increased and I fixed the problem, so in new versions of Daydreamer you should be able to get nicely formatted pv outputs again by increasing the verbosity to medium.

Great. Looking forward to seeing the stylish PV lines again

[Dann Corbit]

A new release candidate is now available here. It includes new ctg code that I hope will fix the problems Carlos has reported, and new gtb code that fixes the endgame problem that Werner reported. It's now possible to force all gtb work to happen in the main search thread, limiting all work to one CPU, by setting the "Load tablebases in a separate thread" option to false.

So far, I've had very bad results testing my gtb implementation on Windows, but I'm not sure if this is due to bad implementation on my part, the relatively low number of games I've been able to try on Windows, or the bad performance characteristics of the virtual machine I'm using.

Can you coax git into providing a tarball?

I have a git client installed on my machine, but I'll be darned if I can pull source from your git repository.

Happy to, you can download it here. Pulling from my repository wouldn't have helped you anyway, because I hadn't pushed my changes out to github until just now. Out of curiosity, what happens when you do
"git clone http://github.com/AaronBecker/daydreamer.git"? I'd like to make it as easy as possible to get a copy of the source, so if there's a problem pulling source I want to fix it.

That works. The problem was procedural on my part.