syzygy wrote:To prevent your programs from being swapped out, disable the page file. Windows will then release RAM caching TB data when it needs more memory.
Thanks, as obvious as it is for some reason that didn't occur to me. Tried it (I set max pagefile size to something small like 256 MB) and indeed the problem seems to be gone. Thank you.
syzygy wrote:To prevent your programs from being swapped out, disable the page file. Windows will then release RAM caching TB data when it needs more memory.
koedem wrote:
On a seperate note, any progress being made on the Syzygy 7 men? The mate in 500 or what it was did seem like it can work.
My building process is ongoing, soon I will prepare another machine to build different sets in parallel.
Ronald is already working on a file format to store large DTZ values, the generator is fine, it would just require a bit more memory to build, and such cases are not very common among 7 pieces.
Is there any hint implying more cursed wins for 7-man bases than 6-man?
ˆ
Passive alternative for 7-man bases is to purchase 20 TB HDD with the idea to use them only with cutechess adjudication. Probing 6-men + 7-man adjudication.
As of now i probe 5-men during search + immediate 6-man adjudication. Work without performance penalty because the HDD is awaken from hibernation once number of pieces present on the board is near 6-pieces.
Nordlandia wrote:Is there any hint implying more cursed wins for 7-man bases than 6-man?
ˆ
Passive alternative for 7-man bases is to purchase 20 TB HDD with the idea to use them only with cutechess adjudication. Probing 6-men + 7-man adjudication.
As of now i probe 5-men during search + immediate 6-man adjudication. Work without performance penalty because the HDD is awaken from hibernation once number of pieces present on the board is near 6-pieces.
Do we need that much HDD space for 7-pieces?
Does cutechess adjudicate 7-men already?