Bootstrap of revision 131654. Testsuite has one failure. Executing on host: /home/kargl/gcc/obj4x/gcc/testsuite/gfortran/../../gfortran -B/home/kargl/gcc/obj4x/ gcc/testsuite/gfortran/../../ /home/kargl/gcc/gcc4x/gcc/testsuite/gfortran.dg/vect/fast-math-pr33299.f9 0 -O2 -ftree-vectorize -fno-vect-cost-model -ftree-vectorizer-verbose=4 -fdump-tree-vect-stats -msse2 -ffast-math -L/home/kargl/gcc/obj4x/i386-unknown-freebsd8.0/./libgfortran/.libs -L/home/kargl/gcc/obj 4x/i386-unknown-freebsd8.0/./libgfortran/.libs -L/home/kargl/gcc/obj4x/i386-unknown-freebsd8.0/./libibe rty -lm -o ./fast-math-pr33299.exe (timeout = 300) PASS: gfortran.dg/vect/fast-math-pr33299.f90 (test for excess errors) Setting LD_LIBRARY_PATH to .:/home/kargl/gcc/obj4x/i386-unknown-freebsd8.0/./libgfortran/.libs:/home/ka rgl/gcc/obj4x/i386-unknown-freebsd8.0/./libgfortran/.libs:/home/kargl/gcc/obj4x/gcc:.:/home/kargl/gcc/o bj4x/i386-unknown-freebsd8.0/./libgfortran/.libs:/home/kargl/gcc/obj4x/i386-unknown-freebsd8.0/./libgfo rtran/.libs:/home/kargl/gcc/obj4x/gcc FAIL: gfortran.dg/vect/fast-math-pr33299.f90 execution test This is a -ffast-math issue as illustrated by 2 command lines. I copied the guilty program into tmp/ kargl[205] gfc4x -o z -O2 -ftree-vectorize -fno-vect-cost-model -ftree-vectorize r-verbose=4 -fdump-tree-vect-stats -fdump-tree-vect-stats -msse2 fast-math-pr332 99.f90 kargl[206] ./z kargl[207] gfc4x -o z -O2 -ftree-vectorize -fno-vect-cost-model -ftree-vectorize r-verbose=4 -fdump-tree-vect-stats -fdump-tree-vect-stats -msse2 -ffast-math fa st-math-pr33299.f90 kargl[208] ./z Illegal instruction (core dumped)
What instructions is causing the crash? Are you testing on a machine which has SSE2 ?
Subject: Re: fast-math-pr33299.f90 failure with illegal instruction due to -ffast-math. On Sun, Jan 20, 2008 at 04:09:19AM -0000, pinskia at gcc dot gnu dot org wrote: > > What instructions is causing the crash? > > Are you testing on a machine which has SSE2 ? > Nope. From FreeBSD's dmesg: CPU: AMD Athlon(tm) Processor (1208.75-MHz 686-class CPU) Origin = "AuthenticAMD" Id = 0x642 Stepping = 2 Features=0x183f9ff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,MMX,FXSR> AMD Features=0xc0440800<SYSCALL,<b18>,MMX+,3DNow!+,3DNow!> SSE2 isn't one of the listed features.
There is the issue, the testcase should be not run on your computer as it is using SSE2. So this is a testsuite issue.
(In reply to comment #3) > There is the issue, the testcase should be not run on your computer as it is > using SSE2. So this is a testsuite issue. Please look at gcc.dg/vect/ how this should be done. There is a check_vect () call before each runtime test (check_vect is in tree-vect.h). I think that fortran vectorizer tests should call this test too - perhaps via C bindings?
*** Bug 34987 has been marked as a duplicate of this bug. ***
Isn't this a dup of 34168?
Subject: Re: fast-math-pr33299.f90 failure with illegal instruction due to -ffast-math. On Tue, Feb 26, 2008 at 05:53:52PM -0000, ghazi at gcc dot gnu dot org wrote: > > ------- Comment #6 from ghazi at gcc dot gnu dot org 2008-02-26 17:53 ------- > Isn't this a dup of 34168? > It appears to be.
(In reply to comment #7) > > Isn't this a dup of 34168? > > > It appears to be. Can you implement Uros' suggestion?
*** This bug has been marked as a duplicate of 34168 ***