This is the mail archive of the
gcc-regression@gcc.gnu.org
mailing list for the GCC project.
GCC memory consumption increased by recent patch!
- From: gcctest at suse dot de
- To: jh at suse dot cz, gcc-regression at gcc dot gnu dot org
- Date: Sun, 12 Sep 2004 22:34:13 +0000
- Subject: GCC memory consumption increased by recent patch!
Hi,
Comparing memory consumption on compilation of combine.i and generate-3.4.ii I got:
comparing combine.c compilation at -O0 level:
Overall memory needed: 17820k
Peak memory use before GGC: 9294k
Peak memory use after GGC: 8606k
Maximum of released memory in single GGC run: 2867k
Garbage: 42475k
Leak: 6107k
Overhead: 5590k
GGC runs: 363
comparing combine.c compilation at -O1 level:
Overall memory needed: 18540k
Peak memory use before GGC: 9573k
Peak memory use after GGC: 8663k
Maximum of released memory in single GGC run: 2067k
Garbage: 78747k
Leak: 6483k
Overhead: 13868k
GGC runs: 589
comparing combine.c compilation at -O2 level:
Overall memory needed: 22076k
Peak memory use before GGC: 12756k
Peak memory use after GGC: 12610k
Maximum of released memory in single GGC run: 2576k
Garbage: 94713k
Leak: 6304k
Overhead: 18777k
GGC runs: 580
comparing combine.c compilation at -O3 level:
Overall memory needed: 23972k
Peak memory use before GGC: 13246k
Peak memory use after GGC: 12610k
Maximum of released memory in single GGC run: 3483k
Garbage: 126026k
Leak: 6852k
Overhead: 24652k
GGC runs: 646
comparing insn-attrtab.c compilation at -O0 level:
Overall memory needed: 132860k
Peak memory use before GGC: 76388k
Peak memory use after GGC: 45185k
Maximum of released memory in single GGC run: 41417k
Garbage: 157790k
Leak: 10620k
Overhead: 19800k
GGC runs: 310
comparing insn-attrtab.c compilation at -O1 level:
Overall memory needed: 151600k
Peak memory use before GGC: 94307k
Peak memory use after GGC: 71401k
Maximum of released memory in single GGC run: 40513k
Garbage: 474239k
Leak: 10968k
Overhead: 84931k
GGC runs: 461
comparing insn-attrtab.c compilation at -O2 level:
Overall memory needed: 237408k
Peak memory use before GGC: 109880k
Peak memory use after GGC: 86974k
Maximum of released memory in single GGC run: 35488k
Garbage: 525180k
Leak: 11150k
Overhead: 95096k
GGC runs: 383
comparing insn-attrtab.c compilation at -O3 level:
Overall memory needed: 237384k
Peak memory use before GGC: 109882k
Peak memory use after GGC: 86975k
Maximum of released memory in single GGC run: 35488k
Garbage: 527525k
Leak: 11223k
Overhead: 95873k
GGC runs: 392
comparing Gerald's testcase PR8361 compilation at -O0 level:
Overall memory allocated via mmap and sbrk increased from 113132k to 114844k, overall 1.51%
Peak amount of GGC memory allocated before garbage collecting increased from 90409k to 92008k, overall 1.77%
Peak amount of GGC memory still allocated after garbage collectin increased from 88883k to 90475k, overall 1.79%
Amount of produced GGC garbage increased from 267696k to 270774k, overall 1.15%
Amount of memory still referenced at the end of compilation increased from 57972k to 58324k, overall 0.61%
Overall memory needed: 113132k -> 114844k
Peak memory use before GGC: 90409k -> 92008k
Peak memory use after GGC: 88883k -> 90475k
Maximum of released memory in single GGC run: 20325k -> 20896k
Garbage: 267696k -> 270774k
Leak: 57972k -> 58324k
Overhead: 34596k -> 34949k
GGC runs: 553 -> 552
comparing Gerald's testcase PR8361 compilation at -O1 level:
Overall memory allocated via mmap and sbrk increased from 115408k to 120284k, overall 4.23%
Peak amount of GGC memory allocated before garbage collecting increased from 88921k to 96217k, overall 8.21%
Peak amount of GGC memory still allocated after garbage collectin increased from 88039k to 89741k, overall 1.93%
Amount of memory still referenced at the end of compilation increased from 60153k to 60665k, overall 0.85%
Overall memory needed: 115408k -> 120284k
Peak memory use before GGC: 88921k -> 96217k
Peak memory use after GGC: 88039k -> 89741k
Maximum of released memory in single GGC run: 19486k -> 20047k
Garbage: 671734k -> 672246k
Leak: 60153k -> 60665k
Overhead: 144644k -> 145158k
GGC runs: 841 -> 835
comparing Gerald's testcase PR8361 compilation at -O2 level:
Overall memory allocated via mmap and sbrk increased from 116812k to 121448k, overall 3.97%
Peak amount of GGC memory allocated before garbage collecting increased from 88922k to 96218k, overall 8.20%
Peak amount of GGC memory still allocated after garbage collectin increased from 88039k to 89741k, overall 1.93%
Amount of memory still referenced at the end of compilation increased from 60762k to 61246k, overall 0.80%
Overall memory needed: 116812k -> 121448k
Peak memory use before GGC: 88922k -> 96218k
Peak memory use after GGC: 88039k -> 89741k
Maximum of released memory in single GGC run: 19487k -> 20048k
Garbage: 726419k -> 726805k
Leak: 60762k -> 61246k
Overhead: 166904k -> 167636k
GGC runs: 876 -> 870
comparing Gerald's testcase PR8361 compilation at -O3 level:
Overall memory allocated via mmap and sbrk increased from 119032k to 120080k, overall 0.88%
Peak amount of GGC memory allocated before garbage collecting increased from 90120k to 92007k, overall 2.09%
Peak amount of GGC memory still allocated after garbage collectin increased from 89108k to 90540k, overall 1.61%
Amount of memory still referenced at the end of compilation increased from 61369k to 61633k, overall 0.43%
Overall memory needed: 119032k -> 120080k
Peak memory use before GGC: 90120k -> 92007k
Peak memory use after GGC: 89108k -> 90540k
Maximum of released memory in single GGC run: 19982k -> 20814k
Garbage: 770533k -> 766821k
Leak: 61369k -> 61633k
Overhead: 181353k -> 180906k
GGC runs: 866 -> 851
Head of changelog is:
--- /usr/src/SpecTests/sandbox-britten-memory/x86_64/mem-result/ChangeLog 2004-09-12 20:28:04.000000000 +0000
+++ /usr/src/SpecTests/sandbox-britten-memory/gcc/gcc/ChangeLog 2004-09-12 21:43:27.000000000 +0000
@@ -1,3 +1,9 @@
+2004-09-12 Richard Henderson <rth@redhat.com>
+
+ PR c++/16254
+ * fold-const.c (fold) <case CLEANUP_POINT_EXPR>: Remove.
+ * tree.c, tree.h (has_cleanups): Remove.
+
2004-09-12 Zdenek Dvorak <rakdver@atrey.karlin.mff.cuni.cz>
* tree-ssa-loop-manip.c (split_loop_exit_edge): Handle non-ssaname
--- /usr/src/SpecTests/sandbox-britten-memory/x86_64/mem-result/ChangeLog.cp 2004-09-12 07:16:19.000000000 +0000
+++ /usr/src/SpecTests/sandbox-britten-memory/gcc/gcc/cp/ChangeLog 2004-09-12 21:43:28.000000000 +0000
@@ -1,3 +1,9 @@
+2004-09-12 Richard Henderson <rth@redhat.com>
+
+ PR c++/16254
+ * semantics.c (maybe_cleanup_point_expr): Don't call fold.
+ * typeck.c (condition_conversion): Likewise.
+
2004-09-11 Richard Henderson <rth@redhat.com>
PR c++/17404
I am friendly script caring about memory consumption in GCC. Please contact
jh@suse.cz if something is going wrong.
The results can be reproduced by building compiler with
--enable-gather-detailed-mem-stats targetting x86-64 and compiling preprocessed
combine.c or testcase from PR8632 with:
-fmem-report --param=ggc-min-heapsize=1024 --param=ggc-min-expand=1 -Ox -Q
The memory consumption summary appears in the dump after detailed listing of
the places they are allocated in. Peak memory consumption is actually computed
by looking for maximal value in {GC XXXX -> YYYY} report.
Yours testing script.