On Fri, Jan 13, 2012 at 5:20 PM, Pavel Ivanov <paiva...@gmail.com> wrote:
> > At this point I've exhausted my avenues of research. Does anyone have > any > > further ideas as to what may be going on which could cause this increase > in > > memory management activity? Or any suggestions of where to look next? > > IIRC, SQLite 3.7.9 changed its PCACHE interface and thus page cache > implementation. It introduced one level of indirection into each page > in the cache. But or course I don't think it should result in such > huge amount of additional mallocs. > Also as you see effects only inside Delphi and not in command line > utility it suggests me that Delphi installed its own page cache. And > as it uses old (deprecated in 3.7.9) interface it may have some bad > interaction with new SQLite. > > And just a thought: number of calls to memcpy with 3.7.9 is larger > than with 3.6.17 roughly on the same amount as number of mallocs. > Which suggests that all new calls are actually calls to realloc() (is > number of calls to free() also bigger on 6M?). The number of calls to _free are 6,968,427 in 3.7.9. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
