Codegen floating point stores of constants into integer instructions. This
authorChris Lattner <sabre@nondot.org>
Fri, 7 May 2004 21:18:15 +0000 (21:18 +0000)
committerChris Lattner <sabre@nondot.org>
Fri, 7 May 2004 21:18:15 +0000 (21:18 +0000)
commite7a31c98db05a955ef558d4c685b143c48063b68
tree809c44b3ddd0124cb9c6527ef721b84b08ac9fcc
parent2c9a94cd4f83b9a997f2aca86345a4cb203a1efa
Codegen floating point stores of constants into integer instructions.  This
allows us to compile:

store float 10.0, float* %P

into:
        mov DWORD PTR [%EAX], 1092616192

instead of:

.CPItest_0:                                     # float 0x4024000000000000
.long   1092616192      # float 10
...
        fld DWORD PTR [.CPItest_0]
        fstp DWORD PTR [%EAX]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@13409 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/X86/InstSelectSimple.cpp
lib/Target/X86/X86ISelSimple.cpp