projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 9f528e6) | patch
call libc memcpy/memset if array size is bigger then threshold.
authorRafael Espindola <rafael.espindola@gmail.com>
Mon, 27 Aug 2007 10:18:20 +0000 (10:18 +0000)
committerRafael Espindola <rafael.espindola@gmail.com>
Mon, 27 Aug 2007 10:18:20 +0000 (10:18 +0000)
commit6b83b5d1ae07dcd1c5987f7548ed2cf4be73b6a1
treec66d22904e779c42f651ed9057b19a632f7fa62ftree | snapshot
parent9f528e628090ee0ffca35d4577c23b6544f9a119commit | diff
call libc memcpy/memset if array size is bigger then threshold.

Coping 100MB array (after a warmup) shows that glibc 2.6.1 implementation on
x86-64 (core 2) is 30% faster (from 0.270917s to 0.188079s)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@41479 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/X86/X86ISelLowering.cpp diff | blob | history
test/CodeGen/X86/2004-02-12-Memcpy.llx diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors