------- Additional Comments From bernie at develer dot com 2005-01-24 13:15 ------- Subject: Re: [4.0 Regression] avr-gcc 4.0, multiplication by constant, very long code
Bernardo Innocenti wrote: > marekm at amelek dot gda dot pl wrote: > >>------- Additional Comments From marekm at amelek dot gda dot pl 2005-01-24 >>09:24 ------- >>Subject: Re: [4.0 Regression] avr-gcc 4.0, multiplication by constant, very >>long code >> >>On Mon, Jan 24, 2005 at 08:56:46AM -0000, giovannibajo at libero dot it wrote: >> >> >> >>>Marek, can you review this patch please? >> >> >>>http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19597 >> >> >>Thanks. Reviewing this will take some time - I agree the current rtx >>costs are not perfect, but changing them can affect generated code in >>unexpected ways. It would be good to test it a lot on various test >>cases, to make sure it doesn't introduce new code size regressions... > > > I'm building avr-gcc right now with your two patches and > this one applied. I'll let you know shortly. Not good. With these two patches applied, the size of four big AVR applications increased slightly. These were built with -Os (the second one shows a minor improvement): text data bss dec hex filename 8008 136 401 8545 2161 images-orig/dspslave.elf 8032 136 401 8569 2179 images-patched/dspslave.elf text data bss dec hex filename 18448 536 692 19676 4cdc images-orig/dspmaster.elf 18428 536 692 19656 4cc8 images-patched/dspmaster.elf These with -O2: text data bss dec hex filename 60454 1832 1562 63848 f968 images-orig/kfront.elf 60488 1832 1562 63882 f98a images-patched/kfront.elf text data bss dec hex filename 36160 900 1713 38773 9775 images-orig/kcntrl.elf 36344 900 1713 38957 982d images-patched/kcntrl.elf Would you like to see some diffs? -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19597
