On Wed, May 13, 2009 at 03:28:19AM +0200, Roland Mainz wrote: > johan...@sun.com wrote: > > On Tue, May 12, 2009 at 01:33:15AM +0200, Roland Mainz wrote: > > > Can you check whether the memory allocator in libast performs better in > > > this case (e.g. compile with $ cc -I/usr/include/ast/ -last ... # (note: > > > libast uses a |_ast_|-prefix for all symbols and does (currently) not > > > act as |malloc()| interposer)). > > > > What would the outcome of this experiment tell us about the problem that > > Bob is facing? > > libast (at least the version which will be included with > ksh93-integration update2) includes multiple (selectable) algorithms to > handle small allocation sizes and multiple (selectable) backends for the > allocations (e.g. using |sbrk()| + using |mmap()| (either anon mem or > via /dev/zero)). The idea was that this may help figuring out whether > the choice of backend is responsible or not...
Ok. Was there a specific set of options that you thought might shed some additional light on this problem? So far, we seem to think that it's alignment related, but haven't yet narrowed the problem down any further. -j _______________________________________________ perf-discuss mailing list perf-discuss@opensolaris.org
