--- Additional Comments From tobi at gcc dot gnu dot org 2004-10-11 00:12 ---
BTW, afaics it is surprising that -fno-ivopts fixes this, as the infinite loop
problem had been present even before Paul rewrote code generation for loops to
allow better loop optimization.
--
http://gcc.gnu
--- Additional Comments From tobi at gcc dot gnu dot org 2004-10-11 00:09 ---
(In reply to comment #4)
> Does -fno-ivopts help?
Yes. recompiling slarrv.f with '-Os -fno-ivopts -g' fixed this.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17926
--- Additional Comments From tobi at gcc dot gnu dot org 2004-10-11 00:01 ---
One more thing I should add: I have a patch in tree which makes gfortran output
sensible debug information for loops. This patch has been already posted
http://gcc.gnu.org/ml/fortran/2004-10/msg00073.html
I act
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-10 23:56
---
Does -fno-ivopts help?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17926
--- Additional Comments From tobi at gcc dot gnu dot org 2004-10-10 23:52 ---
(In reply to comment #0)
> The relevant parts form the loop looks something like this:
>40CONTINUE
> IF( NDONE.LT.IM ) THEN
> OLDNCL = NCLUS
> NCLUS = 0
> LSB
--- Additional Comments From tobi at gcc dot gnu dot org 2004-10-10 23:51 ---
Created an attachment (id=7320)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7320&action=view)
output of objdump -s
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17926
--- Additional Comments From tobi at gcc dot gnu dot org 2004-10-10 23:50 ---
Created an attachment (id=7319)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7319&action=view)
LAPACK source file which shows the infinite loop
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17926
