GCC memory consumption increased by recent patch!

Wed Dec 8 02:34:00 GMT 2004

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: 25329k
    Peak memory use before GGC: 9329k
    Peak memory use after GGC: 8641k
    Maximum of released memory in single GGC run: 2926k
    Garbage: 42601k
    Leak: 6087k
    Overhead: 5743k
    GGC runs: 354

comparing combine.c compilation at -O1 level:
  Overall memory allocated via mmap and sbrk increased from 25885k to 26145k, overall 1.00%
    Overall memory needed: 25885k -> 26145k
    Peak memory use before GGC: 9201k
    Peak memory use after GGC: 8723k
    Maximum of released memory in single GGC run: 2063k
    Garbage: 68067k
    Leak: 6484k
    Overhead: 10671k
    GGC runs: 547

comparing combine.c compilation at -O2 level:
    Overall memory needed: 29989k -> 29993k
    Peak memory use before GGC: 12705k
    Peak memory use after GGC: 12578k
    Maximum of released memory in single GGC run: 2575k
    Garbage: 82364k
    Leak: 6305k
    Overhead: 14872k
    GGC runs: 549

comparing combine.c compilation at -O3 level:
    Overall memory needed: 20764k
    Peak memory use before GGC: 12987k
    Peak memory use after GGC: 12578k
    Maximum of released memory in single GGC run: 3412k
    Garbage: 111191k
    Leak: 6832k
    Overhead: 19887k
    GGC runs: 615

comparing insn-attrtab.c compilation at -O0 level:
    Overall memory needed: 118468k
    Peak memory use before GGC: 79386k
    Peak memory use after GGC: 46137k
    Maximum of released memory in single GGC run: 43335k
    Garbage: 162034k
    Leak: 10609k
    Overhead: 21249k
    GGC runs: 295

comparing insn-attrtab.c compilation at -O1 level:
    Overall memory needed: 132256k
    Peak memory use before GGC: 85238k
    Peak memory use after GGC: 70037k
    Maximum of released memory in single GGC run: 41132k
    Garbage: 449332k
    Leak: 10955k
    Overhead: 79597k
    GGC runs: 429

comparing insn-attrtab.c compilation at -O2 level:
    Overall memory needed: 154184k
    Peak memory use before GGC: 101254k
    Peak memory use after GGC: 85454k
    Maximum of released memory in single GGC run: 42175k
    Garbage: 496336k
    Leak: 10883k
    Overhead: 87863k
    GGC runs: 363

comparing insn-attrtab.c compilation at -O3 level:
    Overall memory needed: 154200k
    Peak memory use before GGC: 101255k
    Peak memory use after GGC: 85456k
    Maximum of released memory in single GGC run: 42175k
    Garbage: 497624k
    Leak: 10930k
    Overhead: 88052k
    GGC runs: 371

comparing Gerald's testcase PR8361 compilation at -O0 level:
    Overall memory needed: 113252k
    Peak memory use before GGC: 89795k
    Peak memory use after GGC: 88905k
    Maximum of released memory in single GGC run: 19901k
    Garbage: 249005k
    Leak: 57666k
    Overhead: 45421k
    GGC runs: 362

comparing Gerald's testcase PR8361 compilation at -O1 level:
    Overall memory needed: 95796k
    Peak memory use before GGC: 88796k
    Peak memory use after GGC: 87822k
    Maximum of released memory in single GGC run: 19406k
    Garbage: 553627k -> 553631k
    Leak: 59700k
    Overhead: 115466k -> 115466k
    GGC runs: 611

comparing Gerald's testcase PR8361 compilation at -O2 level:
    Overall memory needed: 95796k
    Peak memory use before GGC: 88796k
    Peak memory use after GGC: 87822k
    Maximum of released memory in single GGC run: 19405k
    Garbage: 602694k -> 602698k
    Leak: 60281k
    Overhead: 137635k -> 137635k
    GGC runs: 655

comparing Gerald's testcase PR8361 compilation at -O3 level:
    Overall memory needed: 98380k -> 98384k
    Peak memory use before GGC: 90197k
    Peak memory use after GGC: 88649k
    Maximum of released memory in single GGC run: 20099k
    Garbage: 642690k -> 642694k
    Leak: 60622k
    Overhead: 148933k -> 148933k
    GGC runs: 647

Head of changelog is:

--- /usr/src/SpecTests/sandbox-britten-memory/x86_64/mem-result/ChangeLog	2004-12-07 20:46:04.000000000 +0000
+++ /usr/src/SpecTests/sandbox-britten-memory/gcc/gcc/ChangeLog	2004-12-08 01:29:17.000000000 +0000
@@ -1,3 +1,61 @@
+2004-12-08  Richard Guenther <richard.guenther@uni-tuebingen.de>
+
+	* doc/invoke.texi: Adjust default values for
+	large-function-growth and inline-unit-growth to match
+	params.def.  Clarify used algorithm.
+
+2004-12-07  David Mosberger  <davidm@hpl.hp.com>
+
+        PR target/18443
+        * config/ia64/ia64.c (ia64_assemble_integer): Add support for
+        emitting unaligned pointer-sized integers.
+
+2004-12-07  Steven Bosscher  <stevenb@suse.de>
+
+	PR c/18867
+	* c-typeck.c (c_start_case): Set orig_type to error_mark_node
+	when the type of the controlling expression is not a valid type.
+
+2004-12-07  Steven Bosscher  <stevenb@suse.de>
+
+	PR tree-optimization/17340
+	* tree-ssa-pre.c (compute_antic): Fix comment.
+	(compute_avail): Do not recurse, instead do a DFS using a stack
+	and a loop.
+	(execute_pre): Adjust.
+
+2004-12-07  Ziemowit Laski  <zlaski@apple.com>
+
+	* c-tree.h (struct lang_type): Rename 'objc_protocols' field
+	to 'objc_info'.
+
+2004-12-07  Eric Christopher  <echristo@redhat.com>
+
+	PR target/16317
+	* config/netbsd.h (NO_IMPLICIT_EXTERN_C): Define.
+
+2004-12-07  Eric Christopher  <echristo@redhat.com>
+
+	* config/mips/mips-protos.h (mips_valid_pointer_mode): Move
+	prototype...
+	* config/mips/mips.c: ... here. Make static.
+
+2004-12-07  Eric Christopher  <echristo@redhat.com>
+
+	PR target/18442
+	* config/mips/mips.c (mips_vector_mode_supported_p): New function.
+	(TARGET_SCALAR_MODE_SUPPORTED_P): Define to above.
+
+2004-12-07  Kazu Hirata  <kazu@cs.umass.edu>
+
+	* c-common.c, expr.c, fold-const.c, print-tree.c,
+	tree-gimple.c, tree-inline.c, tree-pretty-print.c,
+	tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c: Replace uses of
+	first_rtl_op with TREE_CODE_LENGTH.
+	* tree.c (first_rtl_op): Remove.
+	Replace uses of first_rtl_op with TREE_CODE_LENGTH.
+	* tree.h: Remove the prototype for first_rtl_op.
+
 2004-12-07  David Edelsohn  <edelsohn@gnu.org>
 
 	* config/rs6000/t-aix43, config/rs6000/t-aix52 (SHLIB_LINK):
--- /usr/src/SpecTests/sandbox-britten-memory/x86_64/mem-result/ChangeLog.cp	2004-12-07 16:11:24.000000000 +0000
+++ /usr/src/SpecTests/sandbox-britten-memory/gcc/gcc/cp/ChangeLog	2004-12-08 01:29:20.000000000 +0000
@@ -1,3 +1,12 @@
+2004-12-07  Ziemowit Laski  <zlaski@apple.com>
+
+	* cp-tree.h (struct lang_type_class): Rename 'objc_protocols'
+	field to 'objc_info'.
+
+2004-12-07  Kazu Hirata  <kazu@cs.umass.edu>
+
+	* pt.c: Replace a use of first_rtl_op with TREE_CODE_LENGTH.
+
 2004-12-07  Roger Sayle  <roger@eyesopen.com>
 
 	* name-lookup.c (leave_scope): We only need to update

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.

