Elena Demikhovsky 05e61f7113 X86: optimized i64 vector multiply with constant ... When we multiply two 64-bit vectors, we extract lower and upper part and use the PMULUDQ instruction. When one of the operands is a constant, the upper part may be zero, we know this at compile time. Example: %a = mul <4 x i64> %b, <4 x i64> < i64 5, i64 5, i64 5, i64 5>. I'm checking the value of the upper part and prevent redundant "multiply", "shift" and "add" operations. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@239802 91177308-0d34-0410-b5e6-96231b3b80d8		2015-06-16 06:07:24 +00:00
..
Analysis	Move logic from JumpThreading into LazyValue info to simplify caller.	2015-06-16 00:49:59 +00:00
AsmParser	modules: Add explicit dependency on intrinsics_gen	2015-06-16 00:44:12 +00:00
Bitcode	modules: Add explicit dependency on intrinsics_gen	2015-06-16 00:44:12 +00:00
CodeGen
DebugInfo
ExecutionEngine	modules: Add explicit dependency on intrinsics_gen	2015-06-16 00:44:12 +00:00
Fuzzer
IR
IRReader	modules: Add explicit dependency on intrinsics_gen	2015-06-16 00:44:12 +00:00
LibDriver
LineEditor
Linker	modules: Add explicit dependency on intrinsics_gen	2015-06-16 00:44:12 +00:00
LTO
MC
Object	modules: Add explicit dependency on intrinsics_gen	2015-06-16 00:44:12 +00:00
Option
Passes
ProfileData	modules: Add explicit dependency on intrinsics_gen	2015-06-16 00:44:12 +00:00
Support
TableGen
Target	X86: optimized i64 vector multiply with constant	2015-06-16 06:07:24 +00:00
Transforms	Revert 239795	2015-06-16 01:20:53 +00:00
CMakeLists.txt
LLVMBuild.txt
Makefile