gcc-4.1-200506011 compiler ICE's using -ftree-vectorize AND -fdump-tree-all-details-stats in conjunction. Leave out one of both options and you got a successful compilation. Command: /opt/gcc-4.1-20050611/bin/gcc -O3 -march=pentium4 -msse3 -save-temps test.c -o test -ftree-vectorize -ftree-vectorizer-verbose=5 -fdump-tree-all-details-stats Error message: test.c: In function 'main': test.c:2: internal compiler error: in op_iter_init_maydef, at tree-flow-inline.h:1065 Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://gcc.gnu.org/bugs.html> for instructions. Source Code (I wrote a stupid little program to reproduce the error): int main(void) { int a[1024]; int b[1024]; int c[1024]; int i, j = 0; for (i = 0; i < 1024; i++) { a[i] = i; b[i] = 1023 - i; } for (i = 0; i < 1024; i++) c[i] = a[i] + b[i]; for (i = 0; i < 1024; i++) j = j + c[i]; return j; } P.S.: Placed under GPL by Ralf Recker <Ralf_Recker@gmx.de> :-) Preprocessed source: # 1 "test.c" # 0 "<built-in>" # 1 "<command line>" # 1 "test.c" int main(void) { int a[1024]; int b[1024]; int c[1024]; int i, j = 0; for (i = 0; i < 1024; i++) { a[i] = i; b[i] = 1023 - i; } for (i = 0; i < 1024; i++) c[i] = a[i] + b[i]; for (i = 0; i < 1024; i++) j = j + c[i]; return j; }
To all attendants of the developer summit: Greet the elks (moose?) from me ;-)
This is a dup of bug 22029 which I reported. Thanks for your bug report, hopefully someone will fix this soon. *** This bug has been marked as a duplicate of 22029 ***
gcc-4.1-20050618 ICE's with the same error message.
P.S.: Why comes .c.t59.copyprop4 before .c.t72.copyprop3 (At that point the ICE seems to occur)?
(In reply to comment #4) > P.S.: Why comes .c.t59.copyprop4 before .c.t72.copyprop3 (At that point the ICE > seems to occur)? I forgot to mention this is due to how tree-optimizate works, it is really a bug (a known one), please file so we don't lose track of it.
(In reply to comment #5) > I forgot to mention this is due to how tree-optimizate works, it is really a bug (a known one), please file > so we don't lose track of it. Sorry, I didn't understood your answer fully (stupid me :-) So is it entered in the BugZilla database, shall I enter it or shall we leave it alone? Ralf Recker
(In reply to comment #6) > Sorry, I didn't understood your answer fully (stupid me :-) > So is it entered in the BugZilla database, shall I enter it or shall we leave it > alone? It is known meaning people have reported it to the GCC list before but that is it. A new bugzilla bug would be nice to keep track of it.
(In reply to comment #7) > It is known meaning people have reported it to the GCC list before but that is it. > A new bugzilla bug would be nice to keep track of it. I just found there was a bug was already filed: PR 20962.
(In reply to comment #8) > (In reply to comment #7) > > It is known meaning people have reported it to the GCC list before but that is it. > > A new bugzilla bug would be nice to keep track of it. > > I just found there was a bug was already filed: PR 20962. > Thank you very much!