[llvm-commits] [llvm] r40015 - in /llvm/trunk: lib/Target/X86/X86InstrSSE.td test/CodeGen/X86/sse-align-0.ll test/CodeGen/X86/sse-align-1.ll test/CodeGen/X86/sse-align-10.ll test/CodeGen/X86/sse-align-2.ll test/CodeGen/X86/sse-align-3.ll test/CodeGen/X86/sse-align-4.ll test/CodeGen/X86/sse-align-5.ll test/CodeGen/X86/sse-align-6.ll test/CodeGen/X86/sse-align-7.ll test/CodeGen/X86/sse-align-8.ll test/CodeGen/X86/sse-align-9.ll
Chris Lattner
clattner at apple.com
Wed Aug 1 10:14:25 PDT 2007
On Jul 18, 2007, at 1:23 PM, Dan Gohman wrote:
> Implement initial memory alignment awareness for SSE instructions.
> Vector loads
> and stores that have a specified alignment of less than 16 bytes
> now use
> instructions that support misaligned memory references.
This is *very* cool.
-Chris
More information about the llvm-commits
mailing list