projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 3986924) | patch
Spill/restore X86 floating point stack registers with 64-bits of precision
authorChris Lattner <sabre@nondot.org>
Thu, 2 Dec 2004 18:17:31 +0000 (18:17 +0000)
committerChris Lattner <sabre@nondot.org>
Thu, 2 Dec 2004 18:17:31 +0000 (18:17 +0000)
commit45de191b0bb337142a3f2a09f5f7410844e46a81
tree5f153d8e81aa4621ce3f48410817936c71f4323ftree | snapshot
parent3986924e0b572e90ab0403a4412967da2c5f78d6commit | diff
Spill/restore X86 floating point stack registers with 64-bits of precision
instead of 80-bits of precision.  This fixes PR467.

This change speeds up fldry on X86 with LLC from 7.32s on apoc to 4.68s.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@18433 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/X86/X86RegisterInfo.cpp diff | blob | history
lib/Target/X86/X86RegisterInfo.td diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors