Evan Cheng a15fdaa292 Optimize some 64-bit multiplication by constants into two lea's or one lea + shl since imulq is slow (latency 5). e.g. ...

x * 40
=>
shlq    $3, %rdi
leaq    (%rdi,%rdi,4), %rax

This has the added benefit of allowing more multiply to be folded into addressing mode. e.g.
a * 24 + b
=>
leaq    (%rdi,%rdi,2), %rax
leaq    (%rsi,%rax,8), %rax

llvm-svn: 67917

2009-03-28 05:57:29 +00:00

..

Analysis

Before deleting a basic block, give other loop passes a chance cleanup analysis values, related to the instructions in the basic block.

2009-03-25 23:57:48 +00:00

Archive

Add the private linkage.

2009-01-15 20:18:42 +00:00

AsmParser

Fix a bug in our autoupgrade support: in an argument list to a function

2009-03-25 06:36:36 +00:00

Bitcode

Fix internal representation of fp80 to be the

2009-03-23 21:16:53 +00:00

CodeGen

Optimize some 64-bit multiplication by constants into two lea's or one lea + shl since imulq is slow (latency 5). e.g.

2009-03-28 05:57:29 +00:00

CompilerDriver

-write-graph now can be used with -o.

2009-03-27 12:57:14 +00:00

Debugger

Oops...I committed too much.

2009-03-13 04:39:26 +00:00

ExecutionEngine

fix one more fp80 case (used only by Interpreter)

2009-03-24 18:16:17 +00:00

Linker

Apply ODR linkage changes accidentally dropped during

2009-03-08 13:35:23 +00:00

Support

move a large method out of line.

2009-03-28 02:08:47 +00:00

System

fix a bug Alexei Svitkine pointed out.

2009-03-23 06:46:20 +00:00

Target

Optimize some 64-bit multiplication by constants into two lea's or one lea + shl since imulq is slow (latency 5). e.g.

2009-03-28 05:57:29 +00:00

Transforms

Revert r67798: it breaks llvm-gcc bootstrap on x86-64-linux, presumably due to

2009-03-27 14:56:47 +00:00

VMCore

"ghostify" the ilist<Function> sentinel

2009-03-27 22:28:33 +00:00

Makefile

Reorganize llvmc code.

2009-03-02 09:01:14 +00:00