RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-25 12:12:47 +01:00
Code Issues Projects Releases Wiki Activity
cbadeac131
llvm-mirror/lib/Target/ARM/ARMHazardRecognizer.h

55 lines
1.6 KiB
C
Raw Normal View History

Making use of VFP / NEON floating point multiply-accumulate / subtraction is difficult on current ARM implementations for a few reasons. 1. Even though a single vmla has latency that is one cycle shorter than a pair of vmul + vadd, a RAW hazard during the first (4? on Cortex-a8) can cause additional pipeline stall. So it's frequently better to single codegen vmul + vadd. 2. A vmla folowed by a vmul, vmadd, or vsub causes the second fp instruction to stall for 4 cycles. We need to schedule them apart. 3. A vmla followed vmla is a special case. Obvious issuing back to back RAW vmla + vmla is very bad. But this isn't ideal either: vmul vadd vmla Instead, we want to expand the second vmla: vmla vmul vadd Even with the 4 cycle vmul stall, the second sequence is still 2 cycles faster. Up to now, isel simply avoid codegen'ing fp vmla / vmls. This works well enough but it isn't the optimial solution. This patch attempts to make it possible to use vmla / vmls in cases where it is profitable. A. Add missing isel predicates which cause vmla to be codegen'ed. B. Make sure the fmul in (fadd (fmul)) has a single use. We don't want to compute a fmul and a fmla. C. Add additional isel checks for vmla, avoid cases where vmla is feeding into fp instructions (except for the #3 exceptional case). D. Add ARM hazard recognizer to model the vmla / vmls hazards. E. Add a special pre-regalloc case to expand vmla / vmls when it's likely the vmla / vmls will trigger one of the special hazards. Work in progress, only A+B are enabled. llvm-svn: 120960
2010-12-05 23:04:16 +01:00
//===-- ARMHazardRecognizer.h - ARM Hazard Recognizers ----------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file defines hazard recognizers for scheduling ARM functions.
//
//===----------------------------------------------------------------------===//
#ifndef ARMHAZARDRECOGNIZER_H
#define ARMHAZARDRECOGNIZER_H
Generalize PostRAHazardRecognizer so it can be used in any pass for both forward and backward scheduling. Rename it to ScoreboardHazardRecognizer (Scoreboard is one word). Remove integer division from the scoreboard's critical path. llvm-svn: 121274
2010-12-08 21:04:29 +01:00
#include "llvm/CodeGen/ScoreboardHazardRecognizer.h"
Making use of VFP / NEON floating point multiply-accumulate / subtraction is difficult on current ARM implementations for a few reasons. 1. Even though a single vmla has latency that is one cycle shorter than a pair of vmul + vadd, a RAW hazard during the first (4? on Cortex-a8) can cause additional pipeline stall. So it's frequently better to single codegen vmul + vadd. 2. A vmla folowed by a vmul, vmadd, or vsub causes the second fp instruction to stall for 4 cycles. We need to schedule them apart. 3. A vmla followed vmla is a special case. Obvious issuing back to back RAW vmla + vmla is very bad. But this isn't ideal either: vmul vadd vmla Instead, we want to expand the second vmla: vmla vmul vadd Even with the 4 cycle vmul stall, the second sequence is still 2 cycles faster. Up to now, isel simply avoid codegen'ing fp vmla / vmls. This works well enough but it isn't the optimial solution. This patch attempts to make it possible to use vmla / vmls in cases where it is profitable. A. Add missing isel predicates which cause vmla to be codegen'ed. B. Make sure the fmul in (fadd (fmul)) has a single use. We don't want to compute a fmul and a fmla. C. Add additional isel checks for vmla, avoid cases where vmla is feeding into fp instructions (except for the #3 exceptional case). D. Add ARM hazard recognizer to model the vmla / vmls hazards. E. Add a special pre-regalloc case to expand vmla / vmls when it's likely the vmla / vmls will trigger one of the special hazards. Work in progress, only A+B are enabled. llvm-svn: 120960
2010-12-05 23:04:16 +01:00
namespace llvm {
class ARMBaseInstrInfo;
class ARMBaseRegisterInfo;
class ARMSubtarget;
class MachineInstr;
Generalize PostRAHazardRecognizer so it can be used in any pass for both forward and backward scheduling. Rename it to ScoreboardHazardRecognizer (Scoreboard is one word). Remove integer division from the scoreboard's critical path. llvm-svn: 121274
2010-12-08 21:04:29 +01:00
class ARMHazardRecognizer : public ScoreboardHazardRecognizer {
Making use of VFP / NEON floating point multiply-accumulate / subtraction is difficult on current ARM implementations for a few reasons. 1. Even though a single vmla has latency that is one cycle shorter than a pair of vmul + vadd, a RAW hazard during the first (4? on Cortex-a8) can cause additional pipeline stall. So it's frequently better to single codegen vmul + vadd. 2. A vmla folowed by a vmul, vmadd, or vsub causes the second fp instruction to stall for 4 cycles. We need to schedule them apart. 3. A vmla followed vmla is a special case. Obvious issuing back to back RAW vmla + vmla is very bad. But this isn't ideal either: vmul vadd vmla Instead, we want to expand the second vmla: vmla vmul vadd Even with the 4 cycle vmul stall, the second sequence is still 2 cycles faster. Up to now, isel simply avoid codegen'ing fp vmla / vmls. This works well enough but it isn't the optimial solution. This patch attempts to make it possible to use vmla / vmls in cases where it is profitable. A. Add missing isel predicates which cause vmla to be codegen'ed. B. Make sure the fmul in (fadd (fmul)) has a single use. We don't want to compute a fmul and a fmla. C. Add additional isel checks for vmla, avoid cases where vmla is feeding into fp instructions (except for the #3 exceptional case). D. Add ARM hazard recognizer to model the vmla / vmls hazards. E. Add a special pre-regalloc case to expand vmla / vmls when it's likely the vmla / vmls will trigger one of the special hazards. Work in progress, only A+B are enabled. llvm-svn: 120960
2010-12-05 23:04:16 +01:00
const ARMBaseInstrInfo &TII;
const ARMBaseRegisterInfo &TRI;
const ARMSubtarget &STI;
MachineInstr *LastMI;
Various bits of framework needed for precise machine-level selection DAG scheduling during isel. Most new functionality is currently guarded by -enable-sched-cycles and -enable-sched-hazard. Added InstrItineraryData::IssueWidth field, currently derived from ARM itineraries, but could be initialized differently on other targets. Added ScheduleHazardRecognizer::MaxLookAhead to indicate whether it is active, and if so how many cycles of state it holds. Added SchedulingPriorityQueue::HasReadyFilter to allowing gating entry into the scheduler's available queue. ScoreboardHazardRecognizer now accesses the ScheduleDAG in order to get information about it's SUnits, provides RecedeCycle for bottom-up scheduling, correctly computes scoreboard depth, tracks IssueCount, and considers potential stall cycles when checking for hazards. ScheduleDAGRRList now models machine cycles and hazards (under flags). It tracks MinAvailableCycle, drives the hazard recognizer and priority queue's ready filter, manages a new PendingQueue, properly accounts for stall cycles, etc. llvm-svn: 122541
2010-12-24 06:03:26 +01:00
unsigned FpMLxStalls;
Making use of VFP / NEON floating point multiply-accumulate / subtraction is difficult on current ARM implementations for a few reasons. 1. Even though a single vmla has latency that is one cycle shorter than a pair of vmul + vadd, a RAW hazard during the first (4? on Cortex-a8) can cause additional pipeline stall. So it's frequently better to single codegen vmul + vadd. 2. A vmla folowed by a vmul, vmadd, or vsub causes the second fp instruction to stall for 4 cycles. We need to schedule them apart. 3. A vmla followed vmla is a special case. Obvious issuing back to back RAW vmla + vmla is very bad. But this isn't ideal either: vmul vadd vmla Instead, we want to expand the second vmla: vmla vmul vadd Even with the 4 cycle vmul stall, the second sequence is still 2 cycles faster. Up to now, isel simply avoid codegen'ing fp vmla / vmls. This works well enough but it isn't the optimial solution. This patch attempts to make it possible to use vmla / vmls in cases where it is profitable. A. Add missing isel predicates which cause vmla to be codegen'ed. B. Make sure the fmul in (fadd (fmul)) has a single use. We don't want to compute a fmul and a fmla. C. Add additional isel checks for vmla, avoid cases where vmla is feeding into fp instructions (except for the #3 exceptional case). D. Add ARM hazard recognizer to model the vmla / vmls hazards. E. Add a special pre-regalloc case to expand vmla / vmls when it's likely the vmla / vmls will trigger one of the special hazards. Work in progress, only A+B are enabled. llvm-svn: 120960
2010-12-05 23:04:16 +01:00
unsigned ITBlockSize; // No. of MIs in current IT block yet to be scheduled.
MachineInstr *ITBlockMIs[4];
public:
ARMHazardRecognizer(const InstrItineraryData *ItinData,
const ARMBaseInstrInfo &tii,
const ARMBaseRegisterInfo &tri,
Various bits of framework needed for precise machine-level selection DAG scheduling during isel. Most new functionality is currently guarded by -enable-sched-cycles and -enable-sched-hazard. Added InstrItineraryData::IssueWidth field, currently derived from ARM itineraries, but could be initialized differently on other targets. Added ScheduleHazardRecognizer::MaxLookAhead to indicate whether it is active, and if so how many cycles of state it holds. Added SchedulingPriorityQueue::HasReadyFilter to allowing gating entry into the scheduler's available queue. ScoreboardHazardRecognizer now accesses the ScheduleDAG in order to get information about it's SUnits, provides RecedeCycle for bottom-up scheduling, correctly computes scoreboard depth, tracks IssueCount, and considers potential stall cycles when checking for hazards. ScheduleDAGRRList now models machine cycles and hazards (under flags). It tracks MinAvailableCycle, drives the hazard recognizer and priority queue's ready filter, manages a new PendingQueue, properly accounts for stall cycles, etc. llvm-svn: 122541
2010-12-24 06:03:26 +01:00
const ARMSubtarget &sti,
const ScheduleDAG *DAG) :
ScoreboardHazardRecognizer(ItinData, DAG, "post-RA-sched"), TII(tii),
TRI(tri), STI(sti), LastMI(0), ITBlockSize(0) {}
Making use of VFP / NEON floating point multiply-accumulate / subtraction is difficult on current ARM implementations for a few reasons. 1. Even though a single vmla has latency that is one cycle shorter than a pair of vmul + vadd, a RAW hazard during the first (4? on Cortex-a8) can cause additional pipeline stall. So it's frequently better to single codegen vmul + vadd. 2. A vmla folowed by a vmul, vmadd, or vsub causes the second fp instruction to stall for 4 cycles. We need to schedule them apart. 3. A vmla followed vmla is a special case. Obvious issuing back to back RAW vmla + vmla is very bad. But this isn't ideal either: vmul vadd vmla Instead, we want to expand the second vmla: vmla vmul vadd Even with the 4 cycle vmul stall, the second sequence is still 2 cycles faster. Up to now, isel simply avoid codegen'ing fp vmla / vmls. This works well enough but it isn't the optimial solution. This patch attempts to make it possible to use vmla / vmls in cases where it is profitable. A. Add missing isel predicates which cause vmla to be codegen'ed. B. Make sure the fmul in (fadd (fmul)) has a single use. We don't want to compute a fmul and a fmla. C. Add additional isel checks for vmla, avoid cases where vmla is feeding into fp instructions (except for the #3 exceptional case). D. Add ARM hazard recognizer to model the vmla / vmls hazards. E. Add a special pre-regalloc case to expand vmla / vmls when it's likely the vmla / vmls will trigger one of the special hazards. Work in progress, only A+B are enabled. llvm-svn: 120960
2010-12-05 23:04:16 +01:00
Various bits of framework needed for precise machine-level selection DAG scheduling during isel. Most new functionality is currently guarded by -enable-sched-cycles and -enable-sched-hazard. Added InstrItineraryData::IssueWidth field, currently derived from ARM itineraries, but could be initialized differently on other targets. Added ScheduleHazardRecognizer::MaxLookAhead to indicate whether it is active, and if so how many cycles of state it holds. Added SchedulingPriorityQueue::HasReadyFilter to allowing gating entry into the scheduler's available queue. ScoreboardHazardRecognizer now accesses the ScheduleDAG in order to get information about it's SUnits, provides RecedeCycle for bottom-up scheduling, correctly computes scoreboard depth, tracks IssueCount, and considers potential stall cycles when checking for hazards. ScheduleDAGRRList now models machine cycles and hazards (under flags). It tracks MinAvailableCycle, drives the hazard recognizer and priority queue's ready filter, manages a new PendingQueue, properly accounts for stall cycles, etc. llvm-svn: 122541
2010-12-24 06:03:26 +01:00
virtual HazardType getHazardType(SUnit *SU, int Stalls);
Making use of VFP / NEON floating point multiply-accumulate / subtraction is difficult on current ARM implementations for a few reasons. 1. Even though a single vmla has latency that is one cycle shorter than a pair of vmul + vadd, a RAW hazard during the first (4? on Cortex-a8) can cause additional pipeline stall. So it's frequently better to single codegen vmul + vadd. 2. A vmla folowed by a vmul, vmadd, or vsub causes the second fp instruction to stall for 4 cycles. We need to schedule them apart. 3. A vmla followed vmla is a special case. Obvious issuing back to back RAW vmla + vmla is very bad. But this isn't ideal either: vmul vadd vmla Instead, we want to expand the second vmla: vmla vmul vadd Even with the 4 cycle vmul stall, the second sequence is still 2 cycles faster. Up to now, isel simply avoid codegen'ing fp vmla / vmls. This works well enough but it isn't the optimial solution. This patch attempts to make it possible to use vmla / vmls in cases where it is profitable. A. Add missing isel predicates which cause vmla to be codegen'ed. B. Make sure the fmul in (fadd (fmul)) has a single use. We don't want to compute a fmul and a fmla. C. Add additional isel checks for vmla, avoid cases where vmla is feeding into fp instructions (except for the #3 exceptional case). D. Add ARM hazard recognizer to model the vmla / vmls hazards. E. Add a special pre-regalloc case to expand vmla / vmls when it's likely the vmla / vmls will trigger one of the special hazards. Work in progress, only A+B are enabled. llvm-svn: 120960
2010-12-05 23:04:16 +01:00
virtual void Reset();
virtual void EmitInstruction(SUnit *SU);
virtual void AdvanceCycle();
Generalize PostRAHazardRecognizer so it can be used in any pass for both forward and backward scheduling. Rename it to ScoreboardHazardRecognizer (Scoreboard is one word). Remove integer division from the scoreboard's critical path. llvm-svn: 121274
2010-12-08 21:04:29 +01:00
virtual void RecedeCycle();
Making use of VFP / NEON floating point multiply-accumulate / subtraction is difficult on current ARM implementations for a few reasons. 1. Even though a single vmla has latency that is one cycle shorter than a pair of vmul + vadd, a RAW hazard during the first (4? on Cortex-a8) can cause additional pipeline stall. So it's frequently better to single codegen vmul + vadd. 2. A vmla folowed by a vmul, vmadd, or vsub causes the second fp instruction to stall for 4 cycles. We need to schedule them apart. 3. A vmla followed vmla is a special case. Obvious issuing back to back RAW vmla + vmla is very bad. But this isn't ideal either: vmul vadd vmla Instead, we want to expand the second vmla: vmla vmul vadd Even with the 4 cycle vmul stall, the second sequence is still 2 cycles faster. Up to now, isel simply avoid codegen'ing fp vmla / vmls. This works well enough but it isn't the optimial solution. This patch attempts to make it possible to use vmla / vmls in cases where it is profitable. A. Add missing isel predicates which cause vmla to be codegen'ed. B. Make sure the fmul in (fadd (fmul)) has a single use. We don't want to compute a fmul and a fmla. C. Add additional isel checks for vmla, avoid cases where vmla is feeding into fp instructions (except for the #3 exceptional case). D. Add ARM hazard recognizer to model the vmla / vmls hazards. E. Add a special pre-regalloc case to expand vmla / vmls when it's likely the vmla / vmls will trigger one of the special hazards. Work in progress, only A+B are enabled. llvm-svn: 120960
2010-12-05 23:04:16 +01:00
};
} // end namespace llvm
#endif // ARMHAZARDRECOGNIZER_H
Reference in New Issue Copy Permalink