llvm-6502/lib/Transforms
Chandler Carruth a2b88163af Teach SROA how to split whole-alloca integer loads and stores into
smaller integer loads and stores.

The high-level motivation is that the frontend sometimes generates
a single whole-alloca integer load or store during ABI lowering of
splittable allocas. We need to be able to break this apart in order to
see the underlying elements and properly promote them to SSA values. The
hope is that this fixes some performance regressions on x86-32 with the
new SROA pass.

Unfortunately, this causes quite a bit of churn in the test cases, and
bloats some IR that comes out. When we see an alloca that consists soley
of bits and bytes being extracted and re-inserted, we now do some
splitting first, before building widened integer "bucket of bits"
representations. These are always well folded by instcombine however, so
this shouldn't actually result in missed opportunities.

If this splitting of all-integer allocas does cause problems (perhaps
due to smaller SSA values going into the RA), we could potentially go to
some extreme measures to only do this integer splitting trick when there
are non-integer component accesses of an alloca, but discovering this is
quite expensive: it adds yet another complete walk of the recursive use
tree of the alloca.

Either way, I will be watching build bots and LNT bots to see what
fallout there is here. If anyone gets x86-32 numbers before & after this
change, I would be very interested.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166662 91177308-0d34-0410-b5e6-96231b3b80d8
2012-10-25 04:37:07 +00:00
..
Hello Since the Hello pass is built as a loadable dynamic library, don't try to convert it to new-style registration yet. 2010-10-07 00:31:16 +00:00
InstCombine Add some cleanup to the DataLayout changes requested by Chandler. 2012-10-24 18:36:13 +00:00
Instrumentation Back out r166591, not sure why this made it through since I cancelled the command. Bleh, sorry about this! 2012-10-24 17:25:11 +00:00
IPO revert accidental change 2012-10-24 23:48:57 +00:00
Scalar Teach SROA how to split whole-alloca integer loads and stores into 2012-10-25 04:37:07 +00:00
Utils Fix a compiler warning with an unused variable. 2012-10-24 22:32:26 +00:00
Vectorize Add support for additional reduction variables: AND, OR, XOR. 2012-10-25 00:08:41 +00:00
CMakeLists.txt Add a basic-block autovectorization pass. 2012-02-01 03:51:43 +00:00
LLVMBuild.txt Add a basic-block autovectorization pass. 2012-02-01 03:51:43 +00:00
Makefile Add a basic-block autovectorization pass. 2012-02-01 03:51:43 +00:00