projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: dc985ef) | patch
ARM: implement some simple f64 materializations.
authorTim Northover <tnorthover@apple.com>
Tue, 20 Aug 2013 08:57:11 +0000 (08:57 +0000)
committerTim Northover <tnorthover@apple.com>
Tue, 20 Aug 2013 08:57:11 +0000 (08:57 +0000)
commitf79c3a5aefefaf7de1b4c5787871f09d947d85f1
tree37332ee67624c468b2c2379b3218dec63b8fd78etree | snapshot
parentdc985ef0af6fe5d73bfc5bd5b470b6edddd4ac77commit | diff
ARM: implement some simple f64 materializations.

Previously we used a const-pool load for virtually all 64-bit floating values.
Actually, we can get quite a few common values (including 0.0, 1.0) via "vmov"
instructions of one stripe or another.

llvm-svn: 188773
llvm/lib/Target/ARM/ARMISelLowering.cpp diff | blob | history
llvm/test/CodeGen/ARM/constantfp.ll [new file with mode: 0644] blob
llvm/test/CodeGen/ARM/reg_sequence.ll diff | blob | history
Domain: System / Toolchain;