projects / dragonfly.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 042ceb8) | patch
At Jeffrey Hsu's suggestion (who follows USENIX papers far more closely the
authorMatthew Dillon <dillon@dragonflybsd.org>
Thu, 28 Aug 2003 17:24:38 +0000 (17:24 +0000)
committerMatthew Dillon <dillon@dragonflybsd.org>
Thu, 28 Aug 2003 17:24:38 +0000 (17:24 +0000)
commit1c5ca4f32592c2f5f59a2e7dc03c125f254938bd
tree8d75cb642c9e66eafc7051c92edbe5cccce1da92tree | snapshot
parent042ceb8d5fa8cd290e7e26a3176a170db14f87f8commit | diff
At Jeffrey Hsu's suggestion (who follows USENIX papers far more closely the
I do), change the offset for new allocations out of each new zone we create
in order to spread-out L1/L2 cache use.  Without this new allocations will
tend to front-load the cpu caches resulting in non-optimal memory accesses.
sys/kern/kern_slaballoc.c diff | blob | blame | history
sys/sys/slaballoc.h diff | blob | blame | history
DragonFly Project Source