[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Jakub Jelinek changed: What|Removed |Added CC||jakub at gcc dot gnu.org Status|NEW |RESOLVED Target Milestone|8.5 |9.0 Resolution|--- |FIXED --- Comment #10 from Jakub Jelinek --- The GCC 8 branch is being closed, fixed in GCC 9.1.
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Jakub Jelinek changed: What|Removed |Added Target Milestone|8.4 |8.5 --- Comment #9 from Jakub Jelinek --- GCC 8.4.0 has been released, adjusting target milestone.
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Richard Biener changed: What|Removed |Added Target Milestone|7.5 |8.4 --- Comment #8 from Richard Biener --- The GCC 7 branch is being closed, re-targeting to GCC 8.4.
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Richard Biener changed: What|Removed |Added Priority|P3 |P2
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Martin Liška changed: What|Removed |Added Keywords|needs-bisection | --- Comment #7 from Martin Liška --- Started probably in r228194 (change from 400MB -> 1400MB RAM usage).
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Martin Liška changed: What|Removed |Added Status|WAITING |NEW CC||marxin at gcc dot gnu.org Known to work|6.4.0 |8.2.0, 9.0 Known to fail||6.4.0 --- Comment #6 from Martin Liška --- Fixed on trunk in r251412 which is a revision where I introduced switch expansion as a GIMPLE pass.
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Richard Biener changed: What|Removed |Added Target Milestone|7.4 |7.5
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Richard Biener changed: What|Removed |Added Keywords||needs-bisection Last reconfirmed|2017-10-02 00:00:00 |2018-11-22 --- Comment #5 from Richard Biener --- Reconfirmed on x86_64-linux - -m32 -march=i586 is the key: > /usr/bin/time gcc-7 piglit-util-gl-enum-gen.i -S -O -m323.17user 0.03system > 0:03.21elapsed 100%CPU (0avgtext+0avgdata 83788maxresident)k 0inputs+1448outputs (0major+20119minor)pagefaults 0swaps > /usr/bin/time gcc-7 piglit-util-gl-enum-gen.i -S -O -m32 -march=i586 4.85user 0.48system 0:05.34elapsed 99%CPU (0avgtext+0avgdata 1388540maxresident)k 0inputs+1288outputs (0major+389610minor)pagefaults 0swaps > /usr/bin/time gcc-7 piglit-util-gl-enum-gen.i -S -O2 -m32 -march=i586 7.35user 1.89system 0:09.25elapsed 99%CPU (0avgtext+0avgdata 6201368maxresident)k 0inputs+1232outputs (0major+1688989minor)pagefaults 0swaps > /usr/bin/time gcc-7 piglit-util-gl-enum-gen.i -S -O3 -m32 -march=i586 7.51user 1.79system 0:09.30elapsed 99%CPU (0avgtext+0avgdata 6201424maxresident)k 8inputs+1392outputs (0major+1689484minor)pagefaults 0swaps > /usr/bin/time gcc-8 piglit-util-gl-enum-gen.i -S -O3 -m32 -march=i586 2.39user 0.08system 0:02.48elapsed 99%CPU (0avgtext+0avgdata 133140maxresident)k 1632inputs+1552outputs (1major+66249minor)pagefaults 0swaps somehow feels like a memory leak. Well, mem-report shows the following. simplify-rtx.c:403 (simplify_gen_ternary) 15636000: 2.3% 0: 0.0% 0: 0.0% 0: 0.0%488625 emit-rtl.c:1087 (ensure_regno_capacity) 0: 0.0% 21294000: 8.1% 21299200: 31.7% 9042864: 94.6%13 rtlanal.c:5564 (canonicalize_condition)23604360: 3.4% 0: 0.0% 0: 0.0% 0: 0.0%983515 config/i386/i386.c:22616 (ix86_expand_int_compar 58644264: 8.5% 0: 0.0%100944: 0.2% 0: 0.0% 2447717 config/i386/i386.c:22617 (ix86_expand_int_compar 58644264: 8.5% 0: 0.0%100944: 0.2% 0: 0.0% 2447717 config/i386/i386.c:22621 (ix86_expand_int_compar 58644264: 8.5% 0: 0.0%100944: 0.2% 0: 0.0% 2447717 optabs.c:3863 (prepare_cmp_insn) 58745208: 8.5% 0: 0.0% 0: 0.0% 0: 0.0% 2447717 simplify-rtx.c:419 (simplify_gen_relational) 58745304: 8.5% 0: 0.0% 0: 0.0% 0: 0.0% 2447721 rtl.c:319 (copy_rtx) 64517840: 9.3% 0: 0.0%24: 0.0% 0: 0.0% 1466370 emit-rtl.c:436 (gen_raw_REG) 58772328: 8.5%24: 0.0% 23521224: 35.0% 0: 0.0% 3428899 emit-rtl.c:3864 (make_insn_raw) 218936256: 31.6% 0: 0.0%497152: 0.7% 0: 0.0% 3428647 would be interesting to see what fixed it on the GCC 8 branch. Generated code seems to be similar.
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Richard Biener changed: What|Removed |Added Target Milestone|7.3 |7.4 --- Comment #4 from Richard Biener --- GCC 7.3 is being released, adjusting target milestone.
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Richard Biener changed: What|Removed |Added Known to work||6.4.0 Known to fail||7.2.0 --- Comment #3 from Richard Biener --- I can't reproduce with GCC 7.2.0 or 7.2.1 r253932 on x86_64 with -O[12] -m32: > /usr/bin/time /space/rguenther/install/gcc-7.2/bin/gcc > piglit-util-gl-enum-gen.i -S -m32 -O 3.33user 0.02system 0:03.35elapsed 100%CPU (0avgtext+0avgdata 84760maxresident)k 0inputs+1416outputs (0major+17649minor)pagefaults 0swaps > /usr/bin/time /space/rguenther/install/gcc-7.2/bin/gcc > piglit-util-gl-enum-gen.i -S -m32 -O2 4.37user 0.04system 0:04.44elapsed 99%CPU (0avgtext+0avgdata 103948maxresident)k 2384inputs+1232outputs (4major+33840minor)pagefaults 0swaps Note that specifying -march=i586 increases memory use quite a bit (10x): > /usr/bin/time /space/rguenther/install/gcc-7.2/bin/gcc > piglit-util-gl-enum-gen.i -S -m32 -O -march=i586 5.28user 0.39system 0:05.68elapsed 99%CPU (0avgtext+0avgdata 1389360maxresident)k 0inputs+1288outputs (0major+390160minor)pagefaults 0swaps but I see similar values with GCC 6: > /usr/bin/time /space/rguenther/install/gcc-6.4/bin/gcc > piglit-util-gl-enum-gen.i -S -m32 -O -march=i586 4.47user 0.31system 0:05.16elapsed 92%CPU (0avgtext+0avgdata 1030020maxresident)k 39416inputs+1288outputs (43major+293486minor)pagefaults 0swaps so - please report the GCC output for the "failing" case when you append -v to the command.
[Bug middle-end/82364] [7 Regression] Enormous memory usage when building for 32bit i386 with >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82364 Adrian Bunk changed: What|Removed |Added Summary|[7 Regression] Enormous |[7 Regression] Enormous |memory usage when building |memory usage when building |on 32bit i386 with >= -O1 |for 32bit i386 with >= -O1 --- Comment #2 from Adrian Bunk --- (In reply to Richard Biener from comment #1) > Can you provide output of the compile command with -v appended so we can see > the different default flags passed for both the 32bit compile and the 64bit > -m32 compile? Thanks for asking, looking at the output I realized that I made the stupid mistake of running gcc 6 in the 64bit -m32 compile. The 64bit -m32 compile succeeds with gcc-7, but it actually has 5.5 GB memory usage.
