Hi. I resend and shortened this email because previous emails are gone on the list ☹
I set first only 2048 buffers as i supposed buffers does not matter on RAM DISK because read from file is really read from RAM. But to my surprise (maybe not so big surprise) the results are >3 times slower (vs cache 262144 buffers). Is this overhead because of using Windows file read API compared to direct read from cache or something else involved? This can be also because of cache processing. You know, if something is not in the cache, it must be read by Windows file API (here from memory). Then old data must be removed from cache and new one stored in the cache. Can someone profile this and catch where time is spend? On Win API or on cache storing/removing or somewhere else? PS> FB4 looks faster than FB3 ~7% - i test only one time FB4 then i must retest it. 48 isql sessions Server version: WI-T4.0.0.1435 Firebird 4.0 Beta 1 SUCCESSFUL TIMES DONE: 282094 vs 901743 Regards, Karol Bieniaszewski
Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
