This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
New test for g77 regression
- From: law at redhat dot com
- To: gcc-patches at gcc dot gnu dot org
- Cc: gcc at gcc dot gnu dot org
- Date: Thu, 07 Mar 2002 10:56:06 -0700
- Subject: New test for g77 regression
- Reply-to: law at redhat dot com
I've committed this test to the mainline's testsuite. It will fail with
the mainline compiler on ppc-linux native with -O2.
The test (and the larger source from which it was derived) compiles just
fine with 2.95.3, so this represents a regression compared to our older
compilers.
I'm not currently debugging this problem.
* g77.f-torture/compile/20020307-1.f: New test.
SUBROUTINE SWEEP
PARAMETER(MAXDIM=4,MAXVEC=4**3*8,MAXT=20)
REAL*8 B,W1,W2,BNORM,BINV,WT,W0,C1,C2,R1,R2
DIMENSION B(MAXVEC,0:3),W1(MAXVEC,0:3),W2(MAXVEC,0:3)
DIMENSION BNORM(MAXVEC),BINV(MAXVEC),WT(MAXVEC),W0(MAXVEC)
DIMENSION C1(MAXVEC),C2(MAXVEC),R1(MAXVEC),R2(MAXVEC)
DO 200 ILAT=1,2**IDIM
DO 200 I1=1,IDIM
DO 220 I2=1,IDIM
CALL INTACT(ILAT,I1,I1,W1)
220 CONTINUE
DO 310 IATT=1,IDIM
DO 311 I=1,100
WT(I)=ONE + C1(I)*LOG(EPS+R1(I))
IF( R2(I)**2 .LE. (ONE-WT(I)**2) )THEN
W0(I)=WT(I)
ENDIF
311 CONTINUE
310 CONTINUE
200 CONTINUE
END