llvm-svn: 50080

getelementptr-seteq.ll into:

define i1 @test(i64 %X, %S* %P) {
	%C = icmp eq i64 %X, -1		; <i1> [#uses=1]
	ret i1 %C
}

instead of:

define i1 @test(i64 %X, %S* %P) {
	%A.idx.mask = and i64 %X, 4611686018427387903		; <i64> [#uses=1]
	%C = icmp eq i64 %A.idx.mask, 4611686018427387903		; <i1> [#uses=1]
	ret i1 %C
}

And fixes the second half of PR2235.  This speeds up the insertion sort
case by 45%, from 1.12s to 0.77s.  In practice, this will significantly
speed up for loops structured like:

for (double *P = Base + N; P != Base; --P)
  ...

Which happens frequently for C++ iterators.

llvm-svn: 50079

llvm-svn: 50078

argument. The x86-64 ABI requires the incoming value of %rdi to
be copied to %rax on exit from a function that is returning a
large C struct.

Also, add a README-X86-64 entry detailing the missed optimization
opportunity and proposing an alternative approach.

llvm-svn: 50075

empty ScheduleDAG.

llvm-svn: 50054

llvm-svn: 50053

llvm-svn: 50051

llvm-svn: 50047

llvm-svn: 50035

and simplify code that was fallout from the separation of memcpyopt and gvn.

llvm-svn: 50034

warning.

llvm-svn: 50033

llvm-svn: 50032

llvm-svn: 50031

llvm-svn: 50029

lost newlines between blocks.

llvm-svn: 50024

fixes a crash in opt on 433.milc.

llvm-svn: 50023

llvm-svn: 50022

as a global helper function.  At the same type, switch it from taking
a vector of predecessors to an arbitrary sequential input.  This allows
us to switch LoopSimplify to use a SmallVector for various temporary
vectors that it passed into SplitBlockPredecessors.

llvm-svn: 50020

nodes, removing a hack.

llvm-svn: 50019

instead of doing it after every call.

llvm-svn: 50018

llvm-svn: 50016

llvm-svn: 50015

llvm-svn: 50014

llvm-svn: 50008

llvm-svn: 50007

llvm-svn: 50006

a call to Instruction::isUsedOutsideOfBlock.

llvm-svn: 50005

whether it is used outside of some block.  This can be used to see
if there are any non-local references, for example.

llvm-svn: 50004

llvm-svn: 50003

Add a cost function.

llvm-svn: 50002

such as those in PR2235.  Right now the pass is not very
effective. :)

llvm-svn: 50000

llvm-svn: 49999

llvm-svn: 49995

This fixes the testsuite failure on ppcf128-4.ll.

llvm-svn: 49994

llvm-svn: 49993

llvm-svn: 49991

llvm-svn: 49989

function has already been codegen'd. This is required by the Java class loading
mechanism which executes Java code when materializing a function.

llvm-svn: 49988

llvm-svn: 49986

ppc long double, which doesn't work.

This may break some stuff temporarily, but I want to
avoid the propagation of code that assumes this works.

llvm-svn: 49983