This is the mail archive of the
gcc-regression@gcc.gnu.org
mailing list for the GCC project.
A recent patch increased GCC's memory consumption!
- From: gcctest at suse dot de
- To: jh at suse dot cz, gcc-regression at gcc dot gnu dot org
- Date: Wed, 02 Apr 2008 15:15:12 +0000
- Subject: A recent patch increased GCC's memory consumption!
Hi,
I am a friendly script caring about memory consumption in GCC. Please
contact jh@suse.cz if something is going wrong.
Comparing memory consumption on compilation of combine.i, insn-attrtab.i,
and generate-3.4.ii I got:
comparing empty function compilation at -O0 level:
Overall memory needed: 8150k
Peak memory use before GGC: 1102k
Peak memory use after GGC: 1001k
Maximum of released memory in single GGC run: 129k
Garbage: 235k
Leak: 959k
Overhead: 80k
GGC runs: 2
Testing has produced no results
Testing has produced no results
comparing empty function compilation at -O0 -g level:
Overall memory needed: 8166k
Peak memory use before GGC: 1129k
Peak memory use after GGC: 1034k
Maximum of released memory in single GGC run: 126k
Garbage: 237k
Leak: 990k
Overhead: 84k
GGC runs: 4
Testing has produced no results
Testing has produced no results
comparing empty function compilation at -O1 level:
Overall memory needed: 8206k
Peak memory use before GGC: 1102k
Peak memory use after GGC: 1001k
Maximum of released memory in single GGC run: 131k
Garbage: 237k
Leak: 959k
Overhead: 81k
GGC runs: 2
Pre-IPA-Garbage: 225k
Pre-IPA-Leak: 962k
Pre-IPA-Overhead: 79k
Post-IPA-Garbage: 225k
Post-IPA-Leak: 962k
Post-IPA-Overhead: 79k
comparing empty function compilation at -O2 level:
Overall memory needed: 8226k
Peak memory use before GGC: 1102k
Peak memory use after GGC: 1002k
Maximum of released memory in single GGC run: 134k
Garbage: 241k
Leak: 960k
Overhead: 81k
GGC runs: 2
Pre-IPA-Garbage: 225k
Pre-IPA-Leak: 963k
Pre-IPA-Overhead: 79k
Post-IPA-Garbage: 225k
Post-IPA-Leak: 963k
Post-IPA-Overhead: 79k
comparing empty function compilation at -O3 level:
Overall memory needed: 8226k
Peak memory use before GGC: 1102k
Peak memory use after GGC: 1002k
Maximum of released memory in single GGC run: 134k
Garbage: 241k
Leak: 960k
Overhead: 81k
GGC runs: 2
Pre-IPA-Garbage: 225k
Pre-IPA-Leak: 963k
Pre-IPA-Overhead: 79k
Post-IPA-Garbage: 225k
Post-IPA-Leak: 963k
Post-IPA-Overhead: 79k
comparing combine.c compilation at -O0 level:
Overall memory needed: 22634k
Peak memory use before GGC: 7799k
Peak memory use after GGC: 7175k
Maximum of released memory in single GGC run: 1546k
Garbage: 37767k
Leak: 5509k
Overhead: 4605k
GGC runs: 372
Testing has produced no results
Testing has produced no results
comparing combine.c compilation at -O0 -g level:
Overall memory needed: 24510k
Peak memory use before GGC: 9499k
Peak memory use after GGC: 8762k
Maximum of released memory in single GGC run: 1831k
Garbage: 38122k
Leak: 7914k
Overhead: 4997k
GGC runs: 346
Testing has produced no results
Testing has produced no results
comparing combine.c compilation at -O1 level:
Overall memory needed: 32870k -> 32866k
Peak memory use before GGC: 16267k
Peak memory use after GGC: 16084k
Maximum of released memory in single GGC run: 1375k
Garbage: 51557k
Leak: 5657k
Overhead: 5802k
GGC runs: 440
Pre-IPA-Garbage: 13924k
Pre-IPA-Leak: 17105k
Pre-IPA-Overhead: 2131k
Post-IPA-Garbage: 13924k
Post-IPA-Leak: 17105k
Post-IPA-Overhead: 2131k
comparing combine.c compilation at -O2 level:
Amount of produced GGC garbage increased from 70951k to 71481k, overall 0.75%
Overall memory needed: 34670k -> 34674k
Peak memory use before GGC: 16367k -> 16364k
Peak memory use after GGC: 16201k
Maximum of released memory in single GGC run: 1305k
Garbage: 70951k -> 71481k
Leak: 5987k
Overhead: 8055k -> 8156k
GGC runs: 510 -> 512
Pre-IPA-Garbage: 14078k
Pre-IPA-Leak: 17166k
Pre-IPA-Overhead: 2145k
Post-IPA-Garbage: 14078k
Post-IPA-Leak: 17166k
Post-IPA-Overhead: 2145k
comparing combine.c compilation at -O3 level:
Amount of produced GGC garbage increased from 87983k to 88531k, overall 0.62%
Overall memory needed: 37390k -> 37570k
Peak memory use before GGC: 16499k
Peak memory use after GGC: 16206k
Maximum of released memory in single GGC run: 2006k
Garbage: 87983k -> 88531k
Leak: 6050k -> 6042k
Overhead: 9993k -> 10096k
GGC runs: 545 -> 547
Pre-IPA-Garbage: 14080k
Pre-IPA-Leak: 17178k
Pre-IPA-Overhead: 2146k
Post-IPA-Garbage: 14080k
Post-IPA-Leak: 17178k
Post-IPA-Overhead: 2146k
comparing insn-attrtab.c compilation at -O0 level:
Overall memory needed: 137442k -> 137438k
Peak memory use before GGC: 56968k
Peak memory use after GGC: 31393k
Maximum of released memory in single GGC run: 32968k
Garbage: 128704k
Leak: 8125k
Overhead: 14274k
GGC runs: 293
Testing has produced no results
Testing has produced no results
comparing insn-attrtab.c compilation at -O0 -g level:
Overall memory needed: 138594k
Peak memory use before GGC: 58027k
Peak memory use after GGC: 32452k
Maximum of released memory in single GGC run: 32968k
Garbage: 128694k
Leak: 9812k
Overhead: 14512k
GGC runs: 289
Testing has produced no results
Testing has produced no results
comparing insn-attrtab.c compilation at -O1 level:
Overall memory needed: 142870k -> 142866k
Peak memory use before GGC: 56190k
Peak memory use after GGC: 49970k
Maximum of released memory in single GGC run: 23960k
Garbage: 205125k
Leak: 9262k
Overhead: 24695k
GGC runs: 315
Pre-IPA-Garbage: 50369k
Pre-IPA-Leak: 49664k
Pre-IPA-Overhead: 7545k
Post-IPA-Garbage: 50369k
Post-IPA-Leak: 49664k
Post-IPA-Overhead: 7545k
comparing insn-attrtab.c compilation at -O2 level:
Amount of produced GGC garbage increased from 245673k to 246848k, overall 0.48%
Overall memory needed: 184170k -> 184358k
Peak memory use before GGC: 56811k
Peak memory use after GGC: 51536k
Maximum of released memory in single GGC run: 22857k -> 22858k
Garbage: 245673k -> 246848k
Leak: 10341k
Overhead: 30329k -> 30605k
GGC runs: 347
Pre-IPA-Garbage: 50441k
Pre-IPA-Leak: 49667k
Pre-IPA-Overhead: 7553k
Post-IPA-Garbage: 50441k
Post-IPA-Leak: 49667k
Post-IPA-Overhead: 7553k
comparing insn-attrtab.c compilation at -O3 level:
Amount of produced GGC garbage increased from 269382k to 270550k, overall 0.43%
Overall memory needed: 190058k -> 189778k
Peak memory use before GGC: 66353k
Peak memory use after GGC: 62274k
Maximum of released memory in single GGC run: 23354k -> 23355k
Garbage: 269382k -> 270550k
Leak: 10379k
Overhead: 32117k -> 32391k
GGC runs: 346 -> 347
Pre-IPA-Garbage: 50441k
Pre-IPA-Leak: 49667k
Pre-IPA-Overhead: 7553k
Post-IPA-Garbage: 50441k
Post-IPA-Leak: 49667k
Post-IPA-Overhead: 7553k
comparing Gerald's testcase PR8361 compilation at -O0 level:
Overall memory needed: 147667k -> 147635k
Peak memory use before GGC: 83441k
Peak memory use after GGC: 82614k
Maximum of released memory in single GGC run: 16522k
Garbage: 198620k
Leak: 47798k
Overhead: 23753k
GGC runs: 417
Pre-IPA-Garbage: 109267k
Pre-IPA-Leak: 70553k
Pre-IPA-Overhead: 12198k
Post-IPA-Garbage: 109267k
Post-IPA-Leak: 70553k
Post-IPA-Overhead: 12198k
comparing Gerald's testcase PR8361 compilation at -O0 -g level:
Overall memory needed: 166695k
Peak memory use before GGC: 93664k
Peak memory use after GGC: 92737k
Maximum of released memory in single GGC run: 16605k
Garbage: 204260k
Leak: 65692k
Overhead: 26796k
GGC runs: 393
Pre-IPA-Garbage: 109911k
Pre-IPA-Leak: 81801k
Pre-IPA-Overhead: 13755k
Post-IPA-Garbage: 109911k
Post-IPA-Leak: 81801k
Post-IPA-Overhead: 13755k
comparing Gerald's testcase PR8361 compilation at -O1 level:
Overall memory needed: 122092k -> 122085k
Peak memory use before GGC: 83693k
Peak memory use after GGC: 82863k
Maximum of released memory in single GGC run: 15729k
Garbage: 287811k
Leak: 47081k
Overhead: 29690k
GGC runs: 514
Pre-IPA-Garbage: 154418k
Pre-IPA-Leak: 87525k
Pre-IPA-Overhead: 17990k
Post-IPA-Garbage: 154418k
Post-IPA-Leak: 87525k
Post-IPA-Overhead: 17990k
comparing Gerald's testcase PR8361 compilation at -O2 level:
Amount of produced GGC garbage increased from 353950k to 354927k, overall 0.28%
Overall memory needed: 124588k -> 124589k
Peak memory use before GGC: 83899k -> 83912k
Peak memory use after GGC: 83032k
Maximum of released memory in single GGC run: 15732k
Garbage: 353950k -> 354927k
Leak: 48192k
Overhead: 36760k -> 36951k
GGC runs: 587 -> 589
Pre-IPA-Garbage: 156969k
Pre-IPA-Leak: 87662k
Pre-IPA-Overhead: 18269k
Post-IPA-Garbage: 156969k
Post-IPA-Leak: 87662k
Post-IPA-Overhead: 18269k
comparing Gerald's testcase PR8361 compilation at -O3 level:
Amount of produced GGC garbage increased from 377247k to 378284k, overall 0.27%
Overall memory needed: 127412k -> 127417k
Peak memory use before GGC: 85163k -> 85171k
Peak memory use after GGC: 84296k
Maximum of released memory in single GGC run: 16125k
Garbage: 377247k -> 378284k
Leak: 48766k
Overhead: 38883k -> 39087k
GGC runs: 609 -> 611
Pre-IPA-Garbage: 159684k
Pre-IPA-Leak: 87944k
Pre-IPA-Overhead: 18525k
Post-IPA-Garbage: 159684k
Post-IPA-Leak: 87944k
Post-IPA-Overhead: 18525k
comparing PR rtl-optimization/28071 testcase compilation at -O0 level:
Overall memory needed: 380356k
Peak memory use before GGC: 100958k
Peak memory use after GGC: 56611k
Maximum of released memory in single GGC run: 50583k
Garbage: 175620k
Leak: 8934k
Overhead: 30537k
GGC runs: 107
Testing has produced no results
Testing has produced no results
comparing PR rtl-optimization/28071 testcase compilation at -O0 -g level:
Overall memory needed: 381056k
Peak memory use before GGC: 101501k
Peak memory use after GGC: 57155k
Maximum of released memory in single GGC run: 50582k
Garbage: 175735k
Leak: 10245k
Overhead: 30757k
GGC runs: 110
Testing has produced no results
Testing has produced no results
comparing PR rtl-optimization/28071 testcase compilation at -O1 level:
Overall memory needed: 428198k -> 427926k
Peak memory use before GGC: 78563k
Peak memory use after GGC: 70443k
Maximum of released memory in single GGC run: 37595k
Garbage: 231505k
Leak: 16425k
Overhead: 33115k
GGC runs: 106
Pre-IPA-Garbage: 52922k
Pre-IPA-Leak: 49243k
Pre-IPA-Overhead: 7123k
Post-IPA-Garbage: 52922k
Post-IPA-Leak: 49243k
Post-IPA-Overhead: 7123k
comparing PR rtl-optimization/28071 testcase compilation at -O2 level:
Amount of produced GGC garbage increased from 242300k to 246014k, overall 1.53%
Overall memory needed: 353802k -> 353654k
Peak memory use before GGC: 75430k
Peak memory use after GGC: 70444k
Maximum of released memory in single GGC run: 31528k
Garbage: 242300k -> 246014k
Leak: 16598k
Overhead: 35527k -> 36251k
GGC runs: 117
Pre-IPA-Garbage: 91834k
Pre-IPA-Leak: 77300k
Pre-IPA-Overhead: 11796k
Post-IPA-Garbage: 91834k
Post-IPA-Leak: 77300k
Post-IPA-Overhead: 11796k
comparing PR rtl-optimization/28071 testcase compilation at -O3 -fno-tree-pre -fno-tree-fre level:
Amount of produced GGC garbage increased from 368322k to 368798k, overall 0.13%
Overall memory needed: 1197190k -> 1199094k
Peak memory use before GGC: 136591k
Peak memory use after GGC: 127949k
Maximum of released memory in single GGC run: 59897k -> 59981k
Garbage: 368322k -> 368798k
Leak: 25797k
Overhead: 49768k -> 49863k
GGC runs: 104
Pre-IPA-Garbage: 91834k
Pre-IPA-Leak: 77300k
Pre-IPA-Overhead: 11796k
Post-IPA-Garbage: 91834k
Post-IPA-Leak: 77300k
Post-IPA-Overhead: 11796k
Head of the ChangeLog is:
--- /usr/src/SpecTests/sandbox-britten-memory/x86_64/mem-result/ChangeLog 2008-04-02 08:13:21.000000000 +0000
+++ /usr/src/SpecTests/sandbox-britten-memory/gcc/gcc/ChangeLog 2008-04-02 12:02:10.000000000 +0000
@@ -1,3 +1,45 @@
+2008-04-02 Paolo Bonzini <bonzini@gnu.org>
+
+ PR bootstrap/35752
+ * Makefile.in (objdir): Set it here.
+ * configure.ac: Not here. Find dynamic linker characteristics.
+ * exec-tool.in: Use them.
+ * aclocal.m4: Regenerate.
+ * configure: Regenerate.
+
+2008-04-02 Paolo Bonzini <bonzini@gnu.org>
+
+ * expr.c (expand_var): Delete it.
+ * expr.h (expand_var): Delete prototype.
+ * function.c (expand_function_start): Use expand_decl instead.
+ * cfgexpand.c (expand_one_static_var, expand_one_var): Don't call
+ langhook.
+
+2008-04-02 Andy Hutchinson <hutchinsonamdy@aim.com>
+
+ PR rtl-optimization/35542
+ * fwprop.c (forward_propagate_and_simplify): Replace
+ loc_reg_mentioned_in_p with reg_mentioned_p.
+
+2008-04-02 Paolo Bonzini <bonzini@gnu.org>
+
+ PR rtl-optimization/35281
+ * fwprop.c (PR_CAN_APPEAR, PR_HANDLE_MEM): New.
+ (propagate_rtx_1): Handle PR_HANDLE_MEM.
+ (propagate_rtx): Pass PR_HANDLE_MEM if appropriate.
+ (varying_mem_p): Move above propagate_rtx.
+ (all_uses_available_at): Do not check MEMs.
+
+2008-04-02 Rafael Espindola <espindola@google.com>
+
+ * tree-vrp.c (extract_code_and_val_from_cond): Remove.
+ (register_edge_assert_for_2): Split the cond argument.
+ (register_edge_assert_for_1): Adjust for the change in
+ register_edge_assert_for_2.
+ (register_edge_assert_for): Split the cond argument.
+ (find_switch_asserts): Adjust for the change in
+ register_edge_assert_for.
+
2008-04-02 Kai Tietz <kai.tietz@onevision.com>
* config.gcc: Add for x86_64-*-mingw* the t-crtfm to tbuild.
The results can be reproduced by building a 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.
Your testing script.