[PATCH] Proto-patch to make insn-attrtab.c much smaller (was Re: [PATCH] Remove define_function_unit)

[Paolo Bonzini]

> I would say it's definitely worth finding out whether that time and
> space buys us anything.

If anybody wants to experiment, here's a patch.  It does not remove 
*all* optimizations, only those based in optimize_attrs (250 lines of 
code), but it's enough for...

wc insn-attrtab.c.*
   50293  245411 2026436 insn-attrtab.c.opt
                  795200 insn-attrtab.o.opt
   13140   71926  563531 insn-attrtab.c.noopt
                  228328 insn-attrtab.o.noopt

This is basically the same ratio as Michael Matz's message (with a 
different patch) which can be found at 
http://gcc.gnu.org/ml/gcc-patches/2003-09/msg00458.html

Jeff Law wrote in the same thread:

 > I believe the "canonical" test for genattrtab optimizations was always
 > the MIPS port.  IIRC the MIPS port was consistently the most
 > problematical.

But, MIPS had a lot of define_function_units, and I guess Jeff was 
referring to that.  Anyhow, just for reference, Michael's results were

Bootstrap time (for C,C++) x86 HEAD (-j4)       user    39m22.830s
with no optimizations in genattrtab:            user    37m15.600s
kdecore.ii (HEAD)                               user     3m43.130s
kdecore.ii no genattrtab optimization           user     3m43.260s
kdecore.ii mips, HEAD                           user      4m0.640s
kdecore.ii mips, no genattrtab optimization     user      4m1.490s

If anyone wants to pick this patch up and do more timings, that'd be good.

Paolo
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: remove-genattrtab-opts.patch
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20040720/fc8c3a0a/attachment.ksh>