llvm-6502/lib/Target/NVPTX
Jingyue Wu 2918efd551 Add straight-line strength reduction to LLVM
Summary:
Straight-line strength reduction (SLSR) is implemented in GCC but not yet in
LLVM. It has proven to effectively simplify statements derived from an unrolled
loop, and can potentially benefit many other cases too. For example,

LLVM unrolls

  #pragma unroll
  foo (int i = 0; i < 3; ++i) {
    sum += foo((b + i) * s);
  }

into

  sum += foo(b * s);
  sum += foo((b + 1) * s);
  sum += foo((b + 2) * s);

However, no optimizations yet reduce the internal redundancy of the three
expressions:

  b * s
  (b + 1) * s
  (b + 2) * s

With SLSR, LLVM can optimize these three expressions into:

  t1 = b * s
  t2 = t1 + s
  t3 = t2 + s

This commit is only an initial step towards implementing a series of such
optimizations. I will implement more (see TODO in the file commentary) in the
near future. This optimization is enabled for the NVPTX backend for now.
However, I am more than happy to push it to the standard optimization pipeline
after more thorough performance tests.

Test Plan: test/StraightLineStrengthReduce/slsr.ll

Reviewers: eliben, HaoLiu, meheff, hfinkel, jholewinski, atrick

Reviewed By: jholewinski, atrick

Subscribers: karthikthecool, jholewinski, llvm-commits

Differential Revision: http://reviews.llvm.org/D7310

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@228016 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-03 19:37:06 +00:00
..
InstPrinter Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
MCTargetDesc [NVPTX] Do not emit .weak symbols for NVPTX 2014-12-01 21:16:17 +00:00
TargetInfo
cl_common_defines.h
CMakeLists.txt [NVPTX] Add an NVPTX-specific TargetTransformInfo 2014-11-10 18:38:25 +00:00
LLVMBuild.txt Update libdeps in NVPTXCodeGen, since r225944. 2015-01-14 23:01:36 +00:00
Makefile
ManagedStringPool.h Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
NVPTX.h [PM] Remove a bunch of stale TTI creation method declarations. I nuked 2015-02-01 00:22:15 +00:00
NVPTX.td
NVPTXAllocaHoisting.cpp
NVPTXAllocaHoisting.h [LPM] Stop using the string based preservation API. It is an 2015-01-28 04:57:56 +00:00
NVPTXAsmPrinter.cpp Remove usernames from TODOs, NFC 2015-02-03 17:57:38 +00:00
NVPTXAsmPrinter.h [NVPTX] Emit .pragma "nounroll" for loops marked with nounroll 2015-02-01 02:27:45 +00:00
NVPTXAssignValidGlobalNames.cpp
NVPTXFavorNonGenericAddrSpaces.cpp
NVPTXFrameLowering.cpp Have MachineFunction cache a pointer to the subtarget to make lookups 2014-08-05 02:39:49 +00:00
NVPTXFrameLowering.h Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
NVPTXGenericToNVVM.cpp NVPTX: Use MapMetadata() instead of custom/stale/untested logic 2015-01-14 05:14:30 +00:00
NVPTXImageOptimizer.cpp [cleanup] Re-sort all the #include lines in LLVM using 2015-01-14 11:23:27 +00:00
NVPTXInstrFormats.td [NVPTX] Add more surface/texture intrinsics, including CUDA unified texture fetch 2014-07-17 11:59:04 +00:00
NVPTXInstrInfo.cpp [cleanup] Re-sort all the #include lines in LLVM using 2015-01-14 11:23:27 +00:00
NVPTXInstrInfo.h Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
NVPTXInstrInfo.td Migrate NVPTXISelDAGToDAG's getSubtarget to a runOnMachineFunction 2015-01-30 01:40:59 +00:00
NVPTXIntrinsics.td NVPTX: support direct f16 <-> f64 conversions via intrinsics. 2014-07-18 08:30:10 +00:00
NVPTXISelDAGToDAG.cpp Remove unused argument. 2015-01-30 01:41:01 +00:00
NVPTXISelDAGToDAG.h Migrate NVPTXISelDAGToDAG's getSubtarget to a runOnMachineFunction 2015-01-30 01:40:59 +00:00
NVPTXISelLowering.cpp Migrate NVPTXISelLowering to take the subtarget that it's dependent 2015-01-30 01:50:07 +00:00
NVPTXISelLowering.h Migrate NVPTXISelLowering to take the subtarget that it's dependent 2015-01-30 01:50:07 +00:00
NVPTXLowerAggrCopies.cpp
NVPTXLowerAggrCopies.h [LPM] Stop using the string based preservation API. It is an 2015-01-28 04:57:56 +00:00
NVPTXLowerStructArgs.cpp Clean up NVPTXLowerStructArgs.cpp. NFC 2014-11-06 17:05:49 +00:00
NVPTXMachineFunctionInfo.h Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
NVPTXMCExpr.cpp
NVPTXMCExpr.h Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
NVPTXPrologEpilogPass.cpp Migrate a bare getSubtarget call to query the MachineFunction 2015-01-30 01:50:09 +00:00
NVPTXRegisterInfo.cpp [NVPTX] Rename registers %fl -> %fd and %rl -> %rd 2014-07-16 16:26:58 +00:00
NVPTXRegisterInfo.h Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
NVPTXRegisterInfo.td [NVPTX] Rename registers %fl -> %fd and %rl -> %rd 2014-07-16 16:26:58 +00:00
NVPTXReplaceImageHandles.cpp [cleanup] Re-sort all the #include lines in LLVM using 2015-01-14 11:23:27 +00:00
NVPTXSection.h Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
NVPTXSubtarget.cpp Migrate NVPTXISelLowering to take the subtarget that it's dependent 2015-01-30 01:50:07 +00:00
NVPTXSubtarget.h Move DataLayout back to the TargetMachine from TargetSubtargetInfo 2015-01-26 19:03:15 +00:00
NVPTXTargetMachine.cpp Add straight-line strength reduction to LLVM 2015-02-03 19:37:06 +00:00
NVPTXTargetMachine.h [multiversion] Switch all of the targets over to use the 2015-02-01 13:20:00 +00:00
NVPTXTargetObjectFile.h Remove a bit of dead code. 2014-11-12 01:27:22 +00:00
NVPTXTargetTransformInfo.cpp [PM] Switch the TargetMachine interface from accepting a pass manager 2015-01-31 11:17:59 +00:00
NVPTXTargetTransformInfo.h [multiversion] Switch the TTI queries from TargetMachine to Subtarget 2015-02-01 14:22:17 +00:00
NVPTXutil.cpp
NVPTXutil.h Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
NVPTXUtilities.cpp [cleanup] Re-sort all the #include lines in LLVM using 2015-01-14 11:23:27 +00:00
NVPTXUtilities.h Canonicalize header guards into a common format. 2014-08-13 16:26:38 +00:00
NVPTXVector.td Replace neverHasSideEffects=1 with hasSideEffects=0 in all .td files. 2014-11-26 00:46:26 +00:00
NVVMReflect.cpp