]> git.ipfire.org Git - thirdparty/valgrind.git/commit
Add a perf test to measure the speed (and memory) of PDB handling.
authorPhilippe Waroquiers <philippe.waroquiers@skynet.be>
Tue, 17 Jan 2012 22:27:47 +0000 (22:27 +0000)
committerPhilippe Waroquiers <philippe.waroquiers@skynet.be>
Tue, 17 Jan 2012 22:27:47 +0000 (22:27 +0000)
commit4ea8a59e63627123152949f6a8779dad9d737860
tree9ce2638eb1791c31ccb4d70f5367742d589a9473
parentbe97cddd7a99fb359caa50c0118fc31fc87bc9f3
Add a perf test to measure the speed (and memory) of PDB handling.

* modified perf/heap.c so that it optionally creates a partially defined bytes every N bytes
* created perf/heap_pdb4.vgperf calling heap 4 (so one byte on 4 is PDB in what heap allocates).

before/after pool alloc, here are the performances on a ppc64. So, it looks like
pool alloc also significantly improves the speed of PDB handling.

perl perf/vg_perf --vg=../pool_alloc --vg=../before_pool_trunk_untouched --reps=5 perf/heap_pdb4.vgperf
-- heap_pdb4 --
heap_pdb4 pool_alloc:0.41s  no: 3.0s ( 7.3x, -----)  me:11.9s (29.0x, -----)
heap_pdb4 before_pool_trunk_untouched:0.41s  no: 3.1s ( 7.6x, -3.7%)  me:16.9s (41.1x,-41.7%)

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@12342
perf/Makefile.am
perf/heap.c
perf/heap_pdb4.vgperf [new file with mode: 0644]