Closed Marius0509 closed 5 months ago
Oooh, that's embarrassing. This appears to be related to microthreading. I'm disabling microthreading in (ea381c1fca023421bcc02badd5fba28dae19f03e) until I can get to the bottom of this. Let me know if this resolves it for you. And thank you very much for the excellent report.
This is fixed in (e40618da4ce869063238dce7553c6f8e1c916af4). Thank you for the defect report. Microthreading has been enabled once more.
The scramble R2 U2 R2 U2 D2 R2 U2 L2 is 8 move optimal in <R2,U2,L2,D2>.
If I clear the cache and run the command ./twsearch --scramblealg "R2 U2 R2 U2 D2 R2 U2 L2" --moves R2,U2,L2,D2 --writeprunetables never samples/3x3x3.tws multiple times, I randomly get "max depth 8" (2%-3% of the time) or "max depth 12".
If I clear the cache and run the command ./twsearch --scramblealg "R2 U2 R2 U2 D2 R2 U2 L2" --moves R2,U2,L2,D2 --writeprunetables always samples/3x3x3.tws multiple times, the first time I get "max depth 12", and from then on (when the pruning table is already generated) I get "max depth 8".