llvm-project/llvm/lib/Target/AArch64
Adam Nemet 6d8beeca53 [LoopDataPrefetch/Aarch64] Allow selective prefetching of large-strided accesses
Summary:
And use this TTI for Cyclone.  As it was explained in the original RFC
(http://thread.gmane.org/gmane.comp.compilers.llvm.devel/92758), the HW
prefetcher work up to 2KB strides.

I am also adding tests for this and the previous change (D17943):

* Cyclone prefetching accesses with a large stride
* Cyclone not prefetching accesses with a small stride
* Generic Aarch64 subtarget not prefetching either

Reviewers: hfinkel

Subscribers: aemerson, rengolin, llvm-commits, mzolotukhin

Differential Revision: http://reviews.llvm.org/D17945

llvm-svn: 263771
2016-03-18 00:27:38 +00:00
..
AsmParser Move MCTargetAsmParser.h to llvm/MC/MCParser where it belongs. 2016-01-27 10:01:28 +00:00
Disassembler Remove autoconf support 2016-01-26 21:29:08 +00:00
InstPrinter Remove autoconf support 2016-01-26 21:29:08 +00:00
MCTargetDesc AArch64: rename compact unwind forms back to UNWIND_ARM64_*. NFC. 2016-02-23 21:49:05 +00:00
TargetInfo Remove autoconf support 2016-01-26 21:29:08 +00:00
Utils Remove autoconf support 2016-01-26 21:29:08 +00:00
AArch64.h [AArch64] Add pass to remove redundant copy after RA 2016-02-16 20:02:39 +00:00
AArch64.td AArch64: remove CRC feature from Cyclone. 2016-02-24 18:10:17 +00:00
AArch64A53Fix835769.cpp AArch64: Remove implicit ilist iterator conversions, NFC 2015-10-13 20:02:15 +00:00
AArch64A57FPLoadBalancing.cpp [AArch64] Slight cleanup in FPLoadBalancing 2016-02-26 09:10:53 +00:00
AArch64AddressTypePromotion.cpp Simplify some boolean conditional return statements in AArch64. 2016-02-29 22:50:49 +00:00
AArch64AdvSIMDScalarPass.cpp AArch64: Remove implicit ilist iterator conversions, NFC 2015-10-13 20:02:15 +00:00
AArch64AsmPrinter.cpp Clean up redundant copies of Triple objects. NFC 2015-06-16 15:44:21 +00:00
AArch64BranchRelaxation.cpp AArch64: Remove implicit ilist iterator conversions, NFC 2015-10-13 20:02:15 +00:00
AArch64CallLowering.cpp [GlobalISel] Re-apply r260922-260923 with MSVC-friendly code. 2016-02-16 19:26:02 +00:00
AArch64CallLowering.h [GlobalISel] Re-apply r260922-260923 with MSVC-friendly code. 2016-02-16 19:26:02 +00:00
AArch64CallingConvention.h Replace uint16_t with the MCPhysReg typedef in many places. A lot of physical register arrays already use this typedef. 2015-12-05 07:13:35 +00:00
AArch64CallingConvention.td Add support for a preserve_most calling convention to the AArch64 backend. 2016-03-10 04:35:09 +00:00
AArch64CleanupLocalDynamicTLSPass.cpp Revert "CodeGen: MachineInstr::getIterator() => getInstrIterator(), NFC" 2016-02-22 20:49:58 +00:00
AArch64CollectLOH.cpp [AArch64][CollectLOH] Remove an invalid assertion and add a test case exposing it. 2015-08-31 19:02:00 +00:00
AArch64ConditionOptimizer.cpp Fix AArch64ConditionOptimizer 2016-01-15 00:06:58 +00:00
AArch64ConditionalCompares.cpp WIP: CodeGen: Use MachineInstr& in MachineInstrBundle.h, NFC 2016-02-27 17:05:33 +00:00
AArch64DeadRegisterDefinitionsPass.cpp [AArch64] Register AArch64DeadRegisterDefinition pass with LLVM pass manager. 2015-08-05 17:35:34 +00:00
AArch64ExpandPseudoInsts.cpp [AArch64] Register (existing) AArch64ExpandPseudo pass with LLVM pass manager. 2015-08-05 14:22:53 +00:00
AArch64FastISel.cpp Simplify some boolean conditional return statements in AArch64. 2016-02-29 22:50:49 +00:00
AArch64FrameLowering.cpp [AArch64] Refactor AArch64FrameLowering::emitPrologue. NFC. 2016-03-14 18:24:34 +00:00
AArch64FrameLowering.h [AArch64] Clean up callee-save CFI emission. NFC. 2016-02-25 16:36:08 +00:00
AArch64ISelDAGToDAG.cpp Simplify some boolean conditional return statements in AArch64. 2016-02-29 22:50:49 +00:00
AArch64ISelLowering.cpp [DAG] use !isUndef() ; NFCI 2016-03-14 18:09:43 +00:00
AArch64ISelLowering.h [GlobalISel] Re-apply r260922-260923 with MSVC-friendly code. 2016-02-16 19:26:02 +00:00
AArch64InstrAtomics.td
AArch64InstrFormats.td AArch64: Add missing schedinfo, check completeness for cyclone 2016-03-01 21:20:31 +00:00
AArch64InstrInfo.cpp [AArch64] Optimize compare and branch sequence when the compare's constant operand is power of 2 2016-03-10 17:54:55 +00:00
AArch64InstrInfo.h [AArch64] Move helper functions into TII, so they can be reused elsewhere. NFC. 2016-03-09 17:29:48 +00:00
AArch64InstrInfo.td AArch64: remove pseudo-instructions used only for their patterns. 2016-03-10 18:46:12 +00:00
AArch64LoadStoreOptimizer.cpp [AArch64] Move helper functions into TII, so they can be reused elsewhere. NFC. 2016-03-09 17:29:48 +00:00
AArch64MCInstLower.cpp Convert some AArch64 code to foreach loops. NFC. 2015-08-03 19:04:32 +00:00
AArch64MCInstLower.h Revert r240137 (Fixed/added namespace ending comments using clang-tidy. NFC) 2015-06-23 09:49:53 +00:00
AArch64MachineFunctionInfo.h [AArch64] Break the dependency between FP and SP when possible. 2016-03-14 18:17:41 +00:00
AArch64PBQPRegAlloc.cpp CodeGen: Take MachineInstr& in SlotIndexes and LiveIntervals, NFC 2016-02-27 06:40:41 +00:00
AArch64PBQPRegAlloc.h Revert r240137 (Fixed/added namespace ending comments using clang-tidy. NFC) 2015-06-23 09:49:53 +00:00
AArch64PerfectShuffle.h
AArch64PromoteConstant.cpp Simplify some boolean conditional return statements in AArch64. 2016-02-29 22:50:49 +00:00
AArch64RedundantCopyElimination.cpp AArch64: Use MachineInstr& in guaranteesZeroRegInBlock(), NFC 2016-02-27 19:12:54 +00:00
AArch64RegisterInfo.cpp Add support for a preserve_most calling convention to the AArch64 backend. 2016-03-10 04:35:09 +00:00
AArch64RegisterInfo.h CXX_FAST_TLS calling convention: performance improvement for AArch64. 2015-12-16 21:04:19 +00:00
AArch64RegisterInfo.td [AArch64] Add ARMv8.2-A FP16 vector instructions 2015-12-08 12:16:10 +00:00
AArch64SchedA53.td AArch64: Reenable CompleteModel for A53, A57 and Kryo models 2016-03-01 21:55:35 +00:00
AArch64SchedA57.td AArch64: Reenable CompleteModel for A53, A57 and Kryo models 2016-03-01 21:55:35 +00:00
AArch64SchedA57WriteRes.td
AArch64SchedCyclone.td AArch64: Add missing schedinfo, check completeness for cyclone 2016-03-01 21:20:31 +00:00
AArch64SchedKryo.td AArch64: Reenable CompleteModel for A53, A57 and Kryo models 2016-03-01 21:55:35 +00:00
AArch64SchedKryoDetails.td [AArch64] Add support for Qualcomm Kryo CPU. 2016-02-12 15:51:51 +00:00
AArch64SchedM1.td [AArch64] Minor reformatting (NFC). 2016-03-09 19:56:38 +00:00
AArch64Schedule.td AArch64: Add missing schedinfo, check completeness for cyclone 2016-03-01 21:20:31 +00:00
AArch64SelectionDAGInfo.cpp Remove getDataLayout() from TargetSelectionDAGInfo (had no users) 2015-07-09 02:10:08 +00:00
AArch64SelectionDAGInfo.h Rename TargetSelectionDAGInfo into SelectionDAGTargetInfo and move it to CodeGen/ 2016-01-27 16:32:26 +00:00
AArch64StorePairSuppress.cpp [TII] Allow getMemOpBaseRegImmOfs() to accept negative offsets. NFC. 2016-03-09 16:00:35 +00:00
AArch64Subtarget.cpp [GlobalISel] Re-apply r260922-260923 with MSVC-friendly code. 2016-02-16 19:26:02 +00:00
AArch64Subtarget.h [GlobalISel] Re-apply r260922-260923 with MSVC-friendly code. 2016-02-16 19:26:02 +00:00
AArch64TargetMachine.cpp [Aarch64] Add pass LoopDataPrefetch for Cyclone 2016-03-18 00:27:29 +00:00
AArch64TargetMachine.h Replace string GNU Triples with llvm::Triple in TargetMachine. NFC. 2015-06-11 19:41:26 +00:00
AArch64TargetObjectFile.cpp MC: Clean up MCExpr naming. NFC. 2015-05-30 01:25:56 +00:00
AArch64TargetObjectFile.h [AsmPrinter][TLOF] 32-bit MachO support for replacing GOT equivalents 2015-03-06 13:49:05 +00:00
AArch64TargetTransformInfo.cpp [LoopDataPrefetch/Aarch64] Allow selective prefetching of large-strided accesses 2016-03-18 00:27:38 +00:00
AArch64TargetTransformInfo.h [LoopDataPrefetch/Aarch64] Allow selective prefetching of large-strided accesses 2016-03-18 00:27:38 +00:00
CMakeLists.txt [AArch64] Add pass to remove redundant copy after RA 2016-02-16 20:02:39 +00:00
LLVMBuild.txt [AArch64] Plug the beginning of the GlobalISel pipeline. 2016-02-11 19:35:06 +00:00