Обсуждение: test_fsync on MinGW(win32)
Hi. I wanted to see this difference. It is somewhat ad hoc. However, It is more effective than it doesn't show it at all. -- MinGW(win32) -- $ test_fsync Simple write timing: write 3.644000 Compare fsync times on write() and non-write() descriptor: (If the times are similar, fsync() can sync data written on a different descriptor.) write, fsync, close 40.543000 write, close, fsync 47.033000 Compare one o_sync write to two: (o_sync unavailable) open o_dsync, write 0.094000 (fdatasync unavailable) write, fsync, 26.700000 Compare file sync methods with 2 8k writes: open o_dsync, write 0.203000 (fdatasync unavailable) write, fsync, 27.920000 Regards, Hiroshi Saito
Вложения
Patch applied. Thanks. --------------------------------------------------------------------------- Hiroshi Saito wrote: > Hi. > > I wanted to see this difference. It is somewhat ad hoc. > However, It is more effective than it doesn't show it at all. > > -- MinGW(win32) -- > $ test_fsync > Simple write timing: > write 3.644000 > > Compare fsync times on write() and non-write() descriptor: > (If the times are similar, fsync() can sync data written > on a different descriptor.) > write, fsync, close 40.543000 > write, close, fsync 47.033000 > > Compare one o_sync write to two: > (o_sync unavailable) > open o_dsync, write 0.094000 > > (fdatasync unavailable) > write, fsync, 26.700000 > > Compare file sync methods with 2 8k writes: > open o_dsync, write 0.203000 > (fdatasync unavailable) > write, fsync, 27.920000 > > > Regards, > Hiroshi Saito [ Attachment, skipping... ] > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Have you searched our list archives? > > http://archives.postgresql.org -- Bruce Momjian bruce@momjian.us EnterpriseDB http://www.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +