3 msgTurning a scope block into not-quite-a nested f...
2 msgget complete type
2 msgLots of soft-fp warnings
3 msgread source input file
1 msgFortran regressions
2 msgRe: PATCH: Regenerate aclocal.m4 and Makefile.in
1 msggcc-4.3-20070629 is now available
3 msgfailed to compile trunk svn rev 126124
3 msgGCC 4.3.0 Status Report (2007-06-29)
2 msgQuestion about validity of wfmath template code
1 msgSome numbers on Regressions
2 msgStatus of trunk freeze
2 msgStatic const int as array bound inside class
1 msggcc-4.2-20070627 is now available

Proposal: adding two zeros to the integer cost ...
\ J.C. Pizarro (27 Jun 2007)
. \ Rask Ingemann Lambertsen (28 Jun 2007)

7 msgcombine corrupts insns + dumps with insn cost p...
11 msg[ARM] Cirrus EP93xx Maverick Crunch Support - '...
2 msgHelp on testsuite failures, only with optimizat...
8 msg[tuples/LTO] RFC: houghts on auto-generating GS...
5 msgLTO reader support for MEMORY_PARTITION_TAG
Subject:Re: Proposal: adding two zeros to the integer cost to calibrate better.
Group:Gcc
From:Rask Ingemann Lambertsen
Date:28 Jun 2007


On Wed, Jun 27, 2007 at 09:41:08PM +0200, J.C. Pizarro wrote:

> I recommend to add 2 zeros to the integer costs as if those are 2 decimal
> zeros,
> for example,
>
> insn_cost 5: 1200 // it's 12.00
> insn_cost 6: 800 // it's 8.00
> insn_cost 7: 400 // it's 4.00
> insn_cost 8: 433 // it's 4.33 little costly than 7th, +x.xx%
> better calibrating.
> insn_cost 9: 466 // it's 4.66 little costly than 8th, +x.xx%
> better calibrating.
> insn_cost 10: 500 // it's 5.00

We already multiply the costs by four (COSTS_N_INSNS()), so you can
already add or subtract one to fine tune costs.

A much worse problem is that e.g. combine only passes small fragments of
an insn to rtx_cost(), which means that the back end had to guess what the
insn looked like to begin with, rather than simply looking at the rtx passed
to it. It is a garbage in, garbage out system.

--
Rask Ingemann Lambertsen


© 2004-2008 readlist.com