gcl-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gcl-devel] Re: OFLAG, O2FLAGS, O3FLAGS


From: Mike Thomas
Subject: [Gcl-devel] Re: OFLAG, O2FLAGS, O3FLAGS
Date: Sat, 03 Apr 2004 10:30:04 +1000
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 (ax)

Sorry to keep updating every couple of minutes but I'm keeping everyone abreast of the situation so no time is wasted in duplicating work and I'm about to be interrupted indefinitely.

I just rebuilt on my home machine with OFLAG = O2FLAGS = O3FLAGS (to save time I left gmp3 as is ie instead of make clean I did rm */*.exe */*.o */*.a).

With consistent flags set to:

======================================================================
O3FLAGS = -fdefer-pop -fmerge-constants -fthread-jumps -floop-optimize \
 -fcrossjumping -fif-conversion -fif-conversion2 -fdelayed-branch \
 -fguess-branch-probability -fcprop-registers -fomit-frame-pointer \
-fforce-mem -foptimize-sibling-calls -fstrength-reduce -fcse-follow-jumps \
 -fcse-skip-blocks -frerun-cse-after-loop  -frerun-loop-opt \
 -fgcse   -fgcse-lm   -fgcse-sm -freorder-blocks -falign-functions \
 -falign-jumps -falign-loops -falign-labels \
 -fdelete-null-pointer-checks -fexpensive-optimizations -fregmove \
 -fschedule-insns  -fschedule-insns2 -fsched-interblock -fsched-spec \
 -fcaller-saves -fpeephole2

O2FLAGS      = $(O3FLAGS)
OFLAG        = $(O3FLAGS)
======================================================================

The 306th test failure goes away.

This leaves

| #          -fstrict-aliasing \
| #   -freorder-functions \


but I'm hoping that setting all these flags consistently may be the key to all of our probs.

Cheers

MIke Thomas.


Mike Thomas wrote:



Hi Camm.

Discussing stable branch:

I've just noticed that although I was setting O2/3FLAGS yesterday to the various optimisation flags, I was not setting OFLAG.

I suspect that I should have all three settings exactly yhe same for the purposes of these tests. Do you agree?

I'm also wondering whether inconsistent optimisation settings across different parts of GCL would be likely to cause problems especially those to do with moving arguments and results on the stack and function calls around?

Cheers

Mike Thomas






reply via email to

[Prev in Thread] Current Thread [Next in Thread]