On i686-apple-darwin9 the test gcc/testsuite/gcc.dg/graphite/block-3.c introduced in revision 143159 fails with: [ibook-dhum] f90/bug% gcc44 /opt/gcc/_gcc_clean/gcc/testsuite/gcc.dg/graphite/block-3.c /opt/gcc/_gcc_clean/gcc/testsuite/gcc.dg/graphite/block-3.c:6: error: size of array 'A' is too large
I have forgotten to say that the failure occurs in 32 bit mode, but disappears with -m64.
Subject: Re: FAIL: gcc.dg/graphite/block-3.c (test for excess errors) Does the attached patch fix the fail? Thanks, Sebastian
Created attachment 17072 [details] 1317_pr38791.diff
> Does the attached patch fix the fail? With the patch the test compiles (it does with M up to 812) and the "Strip Mining" is done for the second nested loops: for (s_1=0;s_1<=1;s_1++) { for (s_3=0;s_3<=1;s_3++) { for (s_5=0;s_5<=1;s_5++) { for (s_7=64*s_1;s_7<=min(64*s_1+63,99);s_7++) { for (s_9=64*s_3;s_9<=min(64*s_3+63,99);s_9++) { for (s_11=64*s_5;s_11<=min(64*s_5+63,99);s_11++) { S13(s_1,s_3,s_5,s_7,s_9,s_11) ; } } } } } } I don't know how I can test this file alone without regtesting all gcc (I tried: make -k check-gcc RUNTESTFLAGS="dg.exp=graphite/block-3.c" without success).
> I don't know how I can test this file alone without regtesting all gcc (I > tried: make -k check-gcc RUNTESTFLAGS="dg.exp=graphite/block-3.c" without > success). Try make -k check-gcc RUNTESTFLAGS="graphite.exp=block-3.c"
> Try make -k check-gcc RUNTESTFLAGS="graphite.exp=block-3.c" Thanks, then I get: Running /opt/gcc/gcc-4.4-work/gcc/testsuite/gcc.dg/graphite/graphite.exp ... === gcc Summary === # of expected passes 1 /Volumes/MacBook/opt/gcc/i686-darwin/gcc/xgcc version 4.4.0 20090110 (experimental) [trunk revision 143247p1] (GCC)
Before closing this pr as fixed, I have a question: usually tests having -fdump-* in dg-options are doing some search of patterns in the dumped file, e.g. in gcc/testsuite/gcc.dg/pr35729.c /* { dg-options "-Os -fdump-rtl-loop2_invariant" } */ ... /* { dg-final { scan-rtl-dump-times "Decided to move invariant" 0 "loop2_invariant" } } */ I noticed that gcc/testsuite/gcc.dg/graphite/block-3.c has only the cleaning dg-final, but no scan-* one(s). I don't see anything in gcc/testsuite/gcc.dg/graphite/graphite.exp that could supply it either. Is this the intended behavior or is there something missing in this test (and possibly other graphite ones)?
Subject: Re: FAIL: gcc.dg/graphite/block-3.c (test for excess errors) > Before closing this pr as fixed, I have a question: usually tests having > -fdump-* in dg-options are doing some search of patterns in the dumped file, > e.g. in gcc/testsuite/gcc.dg/pr35729.c > > /* { dg-options "-Os -fdump-rtl-loop2_invariant" } */ > ... > /* { dg-final { scan-rtl-dump-times "Decided to move invariant" 0 > "loop2_invariant" } } */ > > I noticed that gcc/testsuite/gcc.dg/graphite/block-3.c has only the cleaning > dg-final, but no scan-* one(s). I don't see anything in > gcc/testsuite/gcc.dg/graphite/graphite.exp that could supply it either. > Is this the intended behavior or is there something missing in this test (and > possibly other graphite ones)? The test for loop blocking is missing in block-3.c. We will have to clean up the graphite testsuite and making the tests more reliable, but probably this will be done in GCC4.5. Sebastian
Fixed.