diff options
author | Andrew Trick <atrick@apple.com> | 2011-03-05 08:00:22 +0000 |
---|---|---|
committer | Andrew Trick <atrick@apple.com> | 2011-03-05 08:00:22 +0000 |
commit | e0ef509aeb47b396cf1bdc170ca4f468f799719f (patch) | |
tree | be309e6cc6e8a6f933206cba78de2930eba7a668 /lib/Target/X86 | |
parent | a4a1c3f0d92f8737b5ff72bfd4b44bf19c859cee (diff) | |
download | external_llvm-e0ef509aeb47b396cf1bdc170ca4f468f799719f.zip external_llvm-e0ef509aeb47b396cf1bdc170ca4f468f799719f.tar.gz external_llvm-e0ef509aeb47b396cf1bdc170ca4f468f799719f.tar.bz2 |
Increased the register pressure limit on x86_64 from 8 to 12
regs. This is the only change in this checkin that may affects the
default scheduler. With better register tracking and heuristics, it
doesn't make sense to artificially lower the register limit so much.
Added -sched-high-latency-cycles and X86InstrInfo::isHighLatencyDef to
give the scheduler a way to account for div and sqrt on targets that
don't have an itinerary. It is currently defaults to 10 (the actual
number doesn't matter much), but only takes effect on non-default
schedulers: list-hybrid and list-ilp.
Added several heuristics that can be individually disabled for the
non-default sched=list-ilp mode. This helps us determine how much
better we can do on a given benchmark than the default
scheduler. Certain compute intensive loops run much faster in this
mode with the right set of heuristics, and it doesn't seem to have
much negative impact elsewhere. Not all of the heuristics are needed,
but we still need to experiment to decide which should be disabled by
default for sched=list-ilp.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@127067 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'lib/Target/X86')
-rw-r--r-- | lib/Target/X86/X86ISelLowering.cpp | 2 | ||||
-rw-r--r-- | lib/Target/X86/X86InstrInfo.cpp | 16 | ||||
-rw-r--r-- | lib/Target/X86/X86InstrInfo.h | 2 |
3 files changed, 13 insertions, 7 deletions
diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp index 722202d..c21aa43 100644 --- a/lib/Target/X86/X86ISelLowering.cpp +++ b/lib/Target/X86/X86ISelLowering.cpp @@ -1284,7 +1284,7 @@ X86TargetLowering::getRegPressureLimit(const TargetRegisterClass *RC, case X86::GR32RegClassID: return 4 - FPDiff; case X86::GR64RegClassID: - return 8 - FPDiff; + return 12 - FPDiff; case X86::VR128RegClassID: return Subtarget->is64Bit() ? 10 : 4; case X86::VR64RegClassID: diff --git a/lib/Target/X86/X86InstrInfo.cpp b/lib/Target/X86/X86InstrInfo.cpp index 76a9b12..21df57c 100644 --- a/lib/Target/X86/X86InstrInfo.cpp +++ b/lib/Target/X86/X86InstrInfo.cpp @@ -3085,12 +3085,8 @@ void X86InstrInfo::getNoopForMachoTarget(MCInst &NopInst) const { NopInst.setOpcode(X86::NOOP); } -bool X86InstrInfo:: -hasHighOperandLatency(const InstrItineraryData *ItinData, - const MachineRegisterInfo *MRI, - const MachineInstr *DefMI, unsigned DefIdx, - const MachineInstr *UseMI, unsigned UseIdx) const { - switch (DefMI->getOpcode()) { +bool X86InstrInfo::isHighLatencyDef(int opc) const { + switch (opc) { default: return false; case X86::DIVSDrm: case X86::DIVSDrm_Int: @@ -3120,6 +3116,14 @@ hasHighOperandLatency(const InstrItineraryData *ItinData, } } +bool X86InstrInfo:: +hasHighOperandLatency(const InstrItineraryData *ItinData, + const MachineRegisterInfo *MRI, + const MachineInstr *DefMI, unsigned DefIdx, + const MachineInstr *UseMI, unsigned UseIdx) const { + return isHighLatencyDef(DefMI->getOpcode()); +} + namespace { /// CGBR - Create Global Base Reg pass. This initializes the PIC /// global base register for x86-32. diff --git a/lib/Target/X86/X86InstrInfo.h b/lib/Target/X86/X86InstrInfo.h index f4581da..81bbd3d 100644 --- a/lib/Target/X86/X86InstrInfo.h +++ b/lib/Target/X86/X86InstrInfo.h @@ -858,6 +858,8 @@ public: const SmallVectorImpl<MachineOperand> &MOs, unsigned Size, unsigned Alignment) const; + bool isHighLatencyDef(int opc) const; + bool hasHighOperandLatency(const InstrItineraryData *ItinData, const MachineRegisterInfo *MRI, const MachineInstr *DefMI, unsigned DefIdx, |