[LLVMbugs] [Bug 23632] New: llc optimises unaligned intrinsic to aligned load

[bugzilla-daemon at llvm.org]

https://llvm.org/bugs/show_bug.cgi?id=23632

            Bug ID: 23632
           Summary: llc optimises unaligned intrinsic to aligned load
           Product: tools
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: llc
          Assignee: unassignedbugs at nondot.org
          Reporter: Vsevolod.Livinskij at frtk.ru
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified

Created attachment 14370
  --> https://llvm.org/bugs/attachment.cgi?id=14370&action=edit
Reproducer.

@llvm.x86.avx512.mask.compress.store.d.512 with -1 mask is optimized to
vmovdqa32 (i.e. aligned move), which seems to be incorrect, as alignment is
unknown in general case.

Compile: llc -mcpu=knl repr.ll
clang++ -march=knl main.cpp repr.s

Run: sde -knl -- ./a.out

llc -version
LLVM (http://llvm.org/):
  LLVM version 3.7.0svn
  Optimized build with assertions.
  Built May 21 2015 (15:01:04).
  Default target: x86_64-unknown-linux-gnu
  Host CPU: sandybridge

clang version 3.7.0 (trunk 237892)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20150522/871f4664/attachment.html>