-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I note in /sys/vm/vm_page.h we have .. /* Backward compatibility for existing PQ_*CACHE config options. */ #if !defined(PQ_CACHESIZE) #if defined(PQ_HUGECACHE) #define PQ_CACHESIZE 1024 #elif defined(PQ_LARGECACHE) #define PQ_CACHESIZE 512 #elif defined(PQ_MEDIUMCACHE) #define PQ_CACHESIZE 256 #elif defined(PQ_NORMALCACHE) #define PQ_CACHESIZE 64 #elif defined(PQ_NOOPT) #define PQ_CACHESIZE 0 #else #define PQ_CACHESIZE 128 #endif #endif /* !defined(PQ_CACHESIZE) */ #if PQ_CACHESIZE >= 1024 #define PQ_PRIME1 31 /* Prime number somewhat less than PQ_L2_SIZE */ #define PQ_PRIME2 23 /* Prime number somewhat less than PQ_L2_SIZE */ #define PQ_L2_SIZE 256 /* A number of colors opt for 1M cache */ #elif PQ_CACHESIZE >= 512 [ .. snip .. ] .. given the ever-increasing availability of L2 cache (even my core duo laptop has 2MB of unified cache between two cores), are these numbers now too small? Alternatively, are these parameters subject to the laws of diminishing returns? Michael -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (FreeBSD) iD8DBQFFCqbUQv9rrgRC1JIRArorAJ9bWRe/qfKZlQ+LvWk8JrRYMP1xGACgxUEy QNnjFa+Iv6B3Lxosjq4OaG4= =EWMf -----END PGP SIGNATURE-----Received on Fri Sep 15 2006 - 11:13:00 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:00 UTC