This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
Hi! Casts from vectors to integers of the same sizes or vice versa are NOP_EXPRs, so the assumption that there will be only modes of the same bitsize if at least one of them is vector mode after folding or tree optimizers is wrong. The following patches (HEAD and 3.4.x) fixes it. The testcase still ICEs at -O3 on x86_64 -m32 -msse2 and at -O{0,3,s} on x86_64 -m64, but that seems to be unrelated bugs, without this patch it ices with -O{0,1,2,3,s} {-m32 -msse2,-m64}. Ok to commit? Jakub
Attachment:
PR16104-HEAD
Description: Text document
Attachment:
PR16104-3.4
Description: Text document
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |