search

Quuxplusone / LLVMBugzillaTest

0 stars 0 forks source link

Clang fails to optimize in-sequence extract operations #11398

Open Quuxplusone opened 13 years ago

Quuxplusone commented 13 years ago
Bugzilla Link PR11211
Status NEW
Importance P normal
Reported by maister@archlinux.us
Reported on 2011-10-22 18:59:39 -0700
Last modified on 2011-10-23 08:39:05 -0700
Version trunk
Hardware PC Linux
CC geek4civic@gmail.com, llvm-bugs@lists.llvm.org
Fixed by commit(s)
Attachments attachment.txt (3536 bytes, text/plain)
attach2.txt (4506 bytes, text/plain)
Blocks
Blocked by
See also 
Created attachment 7502
C and disasm

Implementing a trivial float to int16_t conversion routine, Clang is unable to
optimize 8 subsequent extract/store operations into one single unaligned store.

Code and x86_64 disasm in attachment.
Quuxplusone commented 13 years ago

Attached attachment.txt (3536 bytes, text/plain): C and disasm

Quuxplusone commented 13 years ago

Attached attach2.txt (4506 bytes, text/plain): int16_t -> float conversion test case.

Quuxplusone commented 13 years ago

Added similar bug. This one for int16_t -> float conversion.