This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch RFC] SH: Use FRAME_GROWS_DOWNWARD
Joern RENNECKE <joern.rennecke@st.com> wrote:
> Come to think of it, I can't actually recall why we ended up with
> an inverted allocation pattern for -fomit-frame-pointer, and hence
> I can't verify we still do by merely looking at the source.
> The partial spilling code might also have changed the correlation
> between offset and usage frequency for -fno-omit-frame-pointer.
> This probably calls for some experiments.
I've done a simple experimentation with compiling gcc itself by
the original and the patched compiler. Some object sizes with
-O2 are here:
original compiler patched compiler
no-omit-frame omit-frame no-omit-frame omit-frame
alias.o 35180 34336 35144 34300
alloc-pool.o 1368 1336 1368 1336
attribs.o 3260 3260 3260 3260
bb-reorder.o 19720 19592 20148 19772
bitmap.o 10840 10616 10900 10612
bt-load.o 18476 18368 18068 17984
builtins.o 140084 138252 140324 138444
c-aux-info.o 6184 6216 6224 6212
c-common.o 158660 156188 159148 155984
c-convert.o 2204 2204 2204 2204
c-cppbuiltin.o 12992 12916 13008 12932
c-decl.o 102108 100456 102356 100884
c-dump.o 1184 1152 1184 1152
c-errors.o 1620 1620 1620 1620
c-format.o 56612 56088 55864 55520
c-gimplify.o 2744 2776 2744 2776
c-incpath.o 5260 5260 5292 5260
c-lang.o 14404 14404 14404 14404
c-lex.o 12288 12180 12272 12220
c-objc-common.o 4092 3996 4092 3996
c-opts.o 26108 25284 26108 25284
c-parser.o 74548 73648 74484 73596
c-pch.o 8340 8308 8340 8308
c-ppoutput.o 6796 6764 6796 6764
c-pragma.o 11328 11416 11384 11288
c-pretty-print.o 24824 23736 24824 23736
c-semantics.o 2020 2020 2052 2020
c-typeck.o 99940 98856 99856 98480
caller-save.o 11536 11416 11600 11536
calls.o 38180 37892 37928 37608
cfg.o 17280 16864 17272 16920
cfganal.o 10708 10472 10708 10472
It seems that there is no clear law between the original and
the patched one, though the size of the object may be not so
good measure.
Regards,
kaz