Page 2 of 59

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 5:18 pm
by noobpwnftw
About the "find_val:not found" problem happened at the third reduction, I think this line may need to be changed into

Code: Select all

j = CAPT_CWIN_RED + 1 + i - REDUCE_PLY_RED;
I'm still not quite sure about this.

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 7:45 pm
by Dann Corbit
noobpwnftw wrote:About the "find_val:not found" problem happened at the third reduction, I think this line may need to be changed into

Code: Select all

j = CAPT_CWIN_RED + 1 + i - REDUCE_PLY_RED;
I'm still not quite sure about this.
Once you get the code fixed, I can do some runs on a high end machine if you like. I just bought a bunch of SSDs including 2x2TB and 5 smaller ones.

I think it would be fun to have my very own 7 man file set. I have a great big NAS cabinet, and I can just add on 12 TB disks as they get built. I guess that storage costs will continue to fall over time so it may not be as expensive as one might think.

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 7:50 pm
by syzygy
noobpwnftw wrote:About the "find_val:not found" problem happened at the third reduction, I think this line may need to be changed into

Code: Select all

j = CAPT_CWIN_RED + 1 + i - REDUCE_PLY_RED;
I'm still not quite sure about this.
An error in this line can certainly cause a "find_val: not found" problem, but stats_val is necessary here to keep j in the correct range independent of the number of reductions.

Perhaps there is no new "i" in this reduction iteration, i.e. the max winning DTZ was reached in the previous iteration. That value has now been shifted out of the window, so won't be found in the table. However, lines 250 and 256 check for that...

So I'm not sure what goes wrong here.

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 8:02 pm
by syzygy
The values i and j should correspond to what is done here (assuming num_saves > 0):
https://github.com/syzygy1/tb/blob/2ba7 ... #L156-L159
To reduce confusion I replace j with k in these lines:

Code: Select all

      for &#40;k = 0; k < n; k++) &#123;
        total_stats&#91;1 + stats_val + k&#93; += stats&#91;CAPT_CWIN_RED + k + 2&#93;;
        total_stats&#91;1022 - stats_val - k - 1&#93; += stats&#91;LOSS_IN_ONE - k - 1&#93;;
      &#125;
Our i is equal to 1 + stats_val + k for the highest k < n for which total_stats[1 + stats_val + k] != 0.

Now we need j = CAPT_CWIN_RED + k + 2 for that same value of k.

So j = CAPT_CWIN_RED + k + 2 = CAPT_CWIN_RED + i - 1 - stats_val + 2 = CAPT_CWIN_RED + i - stats_val + 1 = CAPT_CWIN_RED + 1 + i - stats_val.

So line 242 is correct.

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 8:03 pm
by Nordlandia
Just a quick question regarding torrent files.

Do the torrent need to be updated for new sets generated apart from those available for now?

Or simply make another torrent for another generated sets when you feel for it.

Ideally it would been nice if the original torrent can be updated as more sets get generated :)

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 8:07 pm
by syzygy
Wait a minute... 512 is probably simply too low?

In the pawn code this is configurable. Probably a good idea to do the same here.

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 8:09 pm
by syzygy
Nordlandia wrote:Just a quick question regarding torrent files.

Do the torrent need to be updated for new sets generated apart from those available for now?

Or simply make another torrent for another generated sets when you feel for it.

Ideally it would been nice if the original torrent can be updated as more sets get generated :)
I think it is better to wait with large-scale distribution until there is a file format that is certain to work for all 7-piece tables.

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 8:54 pm
by Nordlandia
Concerning DTM bases. Have you checked whether placing them on regular hard disk (HDD) inflict/cause noticeabe performance drop. Similarly like dtz works fine on HDD without noticeable performance penalty.

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 10:09 pm
by syzygy
Nordlandia wrote:Concerning DTM bases. Have you checked whether placing them on regular hard disk (HDD) inflict/cause noticeabe performance drop. Similarly like dtz works fine on HDD without noticeable performance penalty.
HDD is fine if they are probed only at the root (although finding the whole winning line can take a while).

Probing DTM during the search really needs them to be on SSD.

Re: 7-men Syzygy attempt

Posted: Sun Mar 11, 2018 11:46 pm
by phhnguyen
Dann Corbit wrote:
noobpwnftw wrote:About the "find_val:not found" problem happened at the third reduction, I think this line may need to be changed into

Code: Select all

j = CAPT_CWIN_RED + 1 + i - REDUCE_PLY_RED;
I'm still not quite sure about this.
Once you get the code fixed, I can do some runs on a high end machine if you like. I just bought a bunch of SSDs including 2x2TB and 5 smaller ones.

I think it would be fun to have my very own 7 man file set. I have a great big NAS cabinet, and I can just add on 12 TB disks as they get built. I guess that storage costs will continue to fall over time so it may not be as expensive as one might think.
Look like you have been solving storage problem for 7 men!!!

How about RAM? Will you soon have 1 TB? ;)