llvm dev - Aug 2012 - [LLVMdev] RFC: Adding pass in X86PassConfig::addPreEmitPass for LEA optimization on Atom

Hi,

We are getting ready to implement several heuristics for correctly using LEAs to
avoid stalls in the address generator on Atom. Our plan is to:


1.       Disabling LEA generation on Atom in X86ISelDAGToDAG:: SelectLEAAddr()
for all but a few pseudo-instructions

2.       Identify loads and stores in a X86PassConfig::addPreEmitPass() pass and
examine several preceding instructions to determine if an add, subtract, or mov
can profitably be turned into an LEA.

The heuristics for using LEAs efficiently must know how many cycles pass between
the generation of an address and its use. This requires LEAs to be added after
scheduling and register allocation. Also, LEAs should not be used for math
operations due to a 3 cycle stall between the execution stage and the address
generator.

Attached is an incomplete patch that disables isel LEA generation and includes
an empty pre-emit pass that will contain the LEA selection heuristics.

Any feedback you may have on this plan is welcome.

Sincerely,

Tyler Nowicki
Intel
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20120810/3697e3c9/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: UseEarlyAG_Template_svn.patch
Type: application/octet-stream
Size: 4455 bytes
Desc: UseEarlyAG_Template_svn.patch
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20120810/3697e3c9/attachment.obj>