[CodeGen][NPM] Port ProcessImplicitDefs to NPM #138829
Open
+75
−23
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This was referenced May 7, 2025
|
Warning This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
This stack of pull requests is managed by Graphite. Learn more about stacking. |
|
@llvm/pr-subscribers-backend-amdgpu @llvm/pr-subscribers-llvm-regalloc Author: Akshat Oke (optimisan) ChangesAMDGPU/llc-pipeline-npm.ll test update is because the dummy class name is now changed to its CL name. Patch is 22.56 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/138829.diff 8 Files Affected:
diff --git a/llvm/include/llvm/InitializePasses.h b/llvm/include/llvm/InitializePasses.h
index e75f9c7a2cfe8..9a380994c2886 100644
--- a/llvm/include/llvm/InitializePasses.h
+++ b/llvm/include/llvm/InitializePasses.h
@@ -246,7 +246,7 @@ void initializePostRASchedulerLegacyPass(PassRegistry &);
void initializePreISelIntrinsicLoweringLegacyPassPass(PassRegistry &);
void initializePrintFunctionPassWrapperPass(PassRegistry &);
void initializePrintModulePassWrapperPass(PassRegistry &);
-void initializeProcessImplicitDefsPass(PassRegistry &);
+void initializeProcessImplicitDefsLegacyPass(PassRegistry &);
void initializeProfileSummaryInfoWrapperPassPass(PassRegistry &);
void initializePromoteLegacyPassPass(PassRegistry &);
void initializeRABasicPass(PassRegistry &);
diff --git a/llvm/include/llvm/Passes/CodeGenPassBuilder.h b/llvm/include/llvm/Passes/CodeGenPassBuilder.h
index 648437fe12c91..0704ae3b2ec1d 100644
--- a/llvm/include/llvm/Passes/CodeGenPassBuilder.h
+++ b/llvm/include/llvm/Passes/CodeGenPassBuilder.h
@@ -71,6 +71,7 @@
#include "llvm/CodeGen/PeepholeOptimizer.h"
#include "llvm/CodeGen/PostRASchedulerList.h"
#include "llvm/CodeGen/PreISelIntrinsicLowering.h"
+#include "llvm/CodeGen/ProcessImplicitDefs.h"
#include "llvm/CodeGen/RegAllocEvictionAdvisor.h"
#include "llvm/CodeGen/RegAllocFast.h"
#include "llvm/CodeGen/RegAllocGreedyPass.h"
diff --git a/llvm/include/llvm/Passes/MachinePassRegistry.def b/llvm/include/llvm/Passes/MachinePassRegistry.def
index 8717b79b26968..9d75e4642db00 100644
--- a/llvm/include/llvm/Passes/MachinePassRegistry.def
+++ b/llvm/include/llvm/Passes/MachinePassRegistry.def
@@ -184,6 +184,7 @@ MACHINE_FUNCTION_PASS("print<machine-uniformity>",
MachineUniformityPrinterPass(errs()))
MACHINE_FUNCTION_PASS("print<slot-indexes>", SlotIndexesPrinterPass(errs()))
MACHINE_FUNCTION_PASS("print<virtregmap>", VirtRegMapPrinterPass(errs()))
+MACHINE_FUNCTION_PASS("process-imp-defs", ProcessImplicitDefsPass())
MACHINE_FUNCTION_PASS("prolog-epilog", PrologEpilogInserterPass())
MACHINE_FUNCTION_PASS("reg-usage-collector", RegUsageInfoCollectorPass())
MACHINE_FUNCTION_PASS("reg-usage-propagation", RegUsageInfoPropagationPass())
@@ -316,7 +317,6 @@ DUMMY_MACHINE_FUNCTION_PASS("static-data-splitter", StaticDataSplitter)
DUMMY_MACHINE_FUNCTION_PASS("machine-function-splitter", MachineFunctionSplitterPass)
DUMMY_MACHINE_FUNCTION_PASS("machineinstr-printer", MachineFunctionPrinterPass)
DUMMY_MACHINE_FUNCTION_PASS("mirfs-discriminators", MIRAddFSDiscriminatorsPass)
-DUMMY_MACHINE_FUNCTION_PASS("processimpdefs", ProcessImplicitDefsPass)
DUMMY_MACHINE_FUNCTION_PASS("prologepilog-code", PrologEpilogCodeInserterPass)
DUMMY_MACHINE_FUNCTION_PASS("ra-basic", RABasicPass)
DUMMY_MACHINE_FUNCTION_PASS("ra-pbqp", RAPBQPPass)
diff --git a/llvm/lib/CodeGen/CodeGen.cpp b/llvm/lib/CodeGen/CodeGen.cpp
index 065fd4704ccfb..6e12edc3d3a5b 100644
--- a/llvm/lib/CodeGen/CodeGen.cpp
+++ b/llvm/lib/CodeGen/CodeGen.cpp
@@ -110,7 +110,7 @@ void llvm::initializeCodeGen(PassRegistry &Registry) {
initializePostRAMachineSinkingLegacyPass(Registry);
initializePostRASchedulerLegacyPass(Registry);
initializePreISelIntrinsicLoweringLegacyPassPass(Registry);
- initializeProcessImplicitDefsPass(Registry);
+ initializeProcessImplicitDefsLegacyPass(Registry);
initializeRABasicPass(Registry);
initializeRAGreedyLegacyPass(Registry);
initializeRegAllocFastPass(Registry);
diff --git a/llvm/lib/CodeGen/ProcessImplicitDefs.cpp b/llvm/lib/CodeGen/ProcessImplicitDefs.cpp
index be81ecab9c897..54fd7814ef4f3 100644
--- a/llvm/lib/CodeGen/ProcessImplicitDefs.cpp
+++ b/llvm/lib/CodeGen/ProcessImplicitDefs.cpp
@@ -6,6 +6,7 @@
//
//===----------------------------------------------------------------------===//
+#include "llvm/CodeGen/ProcessImplicitDefs.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
@@ -26,24 +27,15 @@ using namespace llvm;
namespace {
/// Process IMPLICIT_DEF instructions and make sure there is one implicit_def
/// for each use. Add isUndef marker to implicit_def defs and their uses.
-class ProcessImplicitDefs : public MachineFunctionPass {
- const TargetInstrInfo *TII = nullptr;
- const TargetRegisterInfo *TRI = nullptr;
- MachineRegisterInfo *MRI = nullptr;
-
- SmallSetVector<MachineInstr*, 16> WorkList;
-
- void processImplicitDef(MachineInstr *MI);
- bool canTurnIntoImplicitDef(MachineInstr *MI);
-
+class ProcessImplicitDefsLegacy : public MachineFunctionPass {
public:
static char ID;
- ProcessImplicitDefs() : MachineFunctionPass(ID) {
- initializeProcessImplicitDefsPass(*PassRegistry::getPassRegistry());
+ ProcessImplicitDefsLegacy() : MachineFunctionPass(ID) {
+ initializeProcessImplicitDefsLegacyPass(*PassRegistry::getPassRegistry());
}
- void getAnalysisUsage(AnalysisUsage &au) const override;
+ void getAnalysisUsage(AnalysisUsage &AU) const override;
bool runOnMachineFunction(MachineFunction &MF) override;
@@ -52,15 +44,29 @@ class ProcessImplicitDefs : public MachineFunctionPass {
MachineFunctionProperties::Property::IsSSA);
}
};
+
+class ProcessImplicitDefs {
+ const TargetInstrInfo *TII = nullptr;
+ const TargetRegisterInfo *TRI = nullptr;
+ MachineRegisterInfo *MRI = nullptr;
+
+ SmallSetVector<MachineInstr *, 16> WorkList;
+
+ void processImplicitDef(MachineInstr *MI);
+ bool canTurnIntoImplicitDef(MachineInstr *MI);
+
+public:
+ bool run(MachineFunction &MF);
+};
} // end anonymous namespace
-char ProcessImplicitDefs::ID = 0;
-char &llvm::ProcessImplicitDefsID = ProcessImplicitDefs::ID;
+char ProcessImplicitDefsLegacy::ID = 0;
+char &llvm::ProcessImplicitDefsID = ProcessImplicitDefsLegacy::ID;
-INITIALIZE_PASS(ProcessImplicitDefs, DEBUG_TYPE,
+INITIALIZE_PASS(ProcessImplicitDefsLegacy, DEBUG_TYPE,
"Process Implicit Definitions", false, false)
-void ProcessImplicitDefs::getAnalysisUsage(AnalysisUsage &AU) const {
+void ProcessImplicitDefsLegacy::getAnalysisUsage(AnalysisUsage &AU) const {
AU.setPreservesCFG();
AU.addPreserved<AAResultsWrapperPass>();
MachineFunctionPass::getAnalysisUsage(AU);
@@ -133,9 +139,24 @@ void ProcessImplicitDefs::processImplicitDef(MachineInstr *MI) {
LLVM_DEBUG(dbgs() << "Keeping physreg: " << *MI);
}
+bool ProcessImplicitDefsLegacy::runOnMachineFunction(MachineFunction &MF) {
+ return ProcessImplicitDefs().run(MF);
+}
+
+PreservedAnalyses
+ProcessImplicitDefsPass::run(MachineFunction &MF,
+ MachineFunctionAnalysisManager &MFAM) {
+ if (!ProcessImplicitDefs().run(MF))
+ return PreservedAnalyses::all();
+
+ return getMachineFunctionPassPreservedAnalyses()
+ .preserveSet<CFGAnalyses>()
+ .preserve<AAManager>();
+}
+
/// processImplicitDefs - Process IMPLICIT_DEF instructions and turn them into
/// <undef> operands.
-bool ProcessImplicitDefs::runOnMachineFunction(MachineFunction &MF) {
+bool ProcessImplicitDefs::run(MachineFunction &MF) {
LLVM_DEBUG(dbgs() << "********** PROCESS IMPLICIT DEFS **********\n"
<< "********** Function: " << MF.getName() << '\n');
diff --git a/llvm/lib/Passes/PassBuilder.cpp b/llvm/lib/Passes/PassBuilder.cpp
index a9d192a7fad55..29cbaf62c2d88 100644
--- a/llvm/lib/Passes/PassBuilder.cpp
+++ b/llvm/lib/Passes/PassBuilder.cpp
@@ -145,6 +145,7 @@
#include "llvm/CodeGen/PostRAHazardRecognizer.h"
#include "llvm/CodeGen/PostRASchedulerList.h"
#include "llvm/CodeGen/PreISelIntrinsicLowering.h"
+#include "llvm/CodeGen/ProcessImplicitDefs.h"
#include "llvm/CodeGen/RegAllocEvictionAdvisor.h"
#include "llvm/CodeGen/RegAllocFast.h"
#include "llvm/CodeGen/RegAllocGreedyPass.h"
diff --git a/llvm/test/CodeGen/AMDGPU/llc-pipeline-npm.ll b/llvm/test/CodeGen/AMDGPU/llc-pipeline-npm.ll
index 634a7fb6eb8e9..9c1d3dcf080ee 100644
--- a/llvm/test/CodeGen/AMDGPU/llc-pipeline-npm.ll
+++ b/llvm/test/CodeGen/AMDGPU/llc-pipeline-npm.ll
@@ -9,9 +9,9 @@
; GCN-O0: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(atomic-expand,verify,gc-lowering,lower-constant-intrinsics,unreachableblockelim,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(lower-switch,lower-invoke,unreachableblockelim,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-control-flow,amdgpu-rewrite-undef-for-phi,lcssa,require<uniformity>,callbr-prepare,safe-stack,stack-protector,verify)),cgscc(function(machine-function(amdgpu-isel,si-fix-sgpr-copies,si-i1-copies,finalize-isel,localstackalloc))),require<reg-usage>,cgscc(function(machine-function(reg-usage-propagation,phi-node-elimination,two-address-instruction,regallocfast,si-fix-vgpr-copies,remove-redundant-debug-values,fixup-statepoint-caller-saved,prolog-epilog,post-ra-pseudos,fentry-insert,xray-instrumentation,patchable-function,si-memory-legalizer,si-insert-waitcnts,si-late-branch-lowering,post-RA-hazard-rec,amdgpu-wait-sgpr-hazards,branch-relaxation,reg-usage-collector,remove-loads-into-fake-uses,live-debug-values,machine-sanmd,stack-frame-layout,verify),invalidate<machine-function-info>))
-; GCN-O2: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,function(amdgpu-image-intrinsic-opt),expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(infer-address-spaces,amdgpu-atomic-optimizer,atomic-expand,amdgpu-promote-alloca,separate-const-offset-from-gep<>,slsr,early-cse<>,nary-reassociate,early-cse<>,amdgpu-codegenprepare,verify,loop-mssa(loop-reduce),mergeicmps,expand-memcmp,gc-lowering,lower-constant-intrinsics,unreachableblockelim,consthoist,replace-with-veclib,partially-inline-libcalls,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,early-cse<>,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(codegenprepare,load-store-vectorizer,lower-switch,lower-invoke,unreachableblockelim,flatten-cfg,sink,amdgpu-late-codegenprepare,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-control-flow,amdgpu-rewrite-undef-for-phi,lcssa)),amdgpu-perf-hint,cgscc(function(require<uniformity>,callbr-prepare,safe-stack,stack-protector,verify)),cgscc(function(machine-function(amdgpu-isel,si-fix-sgpr-copies,si-i1-copies,finalize-isel,early-tailduplication,opt-phis,stack-coloring,localstackalloc,dead-mi-elimination,early-machinelicm,machine-cse,machine-sink,peephole-opt,dead-mi-elimination,si-fold-operands,gcn-dpp-combine,si-load-store-opt,si-peephole-sdwa,early-machinelicm,machine-cse,si-fold-operands,dead-mi-elimination,si-shrink-instructions))),require<reg-usage>,cgscc(function(machine-function(reg-usage-propagation,detect-dead-lanes,dead-mi-elimination,init-undef,ProcessImplicitDefsPass,unreachable-mbb-elimination,require<live-vars>,si-opt-vgpr-liverange,require<machine-loops>,phi-node-elimination,si-lower-control-flow,two-address-instruction,register-coalescer,rename-independent-subregs,amdgpu-rewrite-partial-reg-uses,machine-scheduler,amdgpu-pre-ra-optimizations,si-wqm,si-optimize-exec-masking-pre-ra,si-form-memory-clauses,amdgpu-pre-ra-long-branch-reg,greedy<sgpr>,virt-reg-rewriter<no-clear-vregs>,stack-slot-coloring,si-lower-sgpr-spills,si-pre-allocate-wwm-regs,greedy<wwm>,si-lower-wwm-copies,virt-reg-rewriter<no-clear-vregs>,amdgpu-reserve-wwm-regs,greedy<vgpr>,amdgpu-nsa-reassign,virt-reg-rewriter,amdgpu-mark-last-scratch-load,machine-cp,machinelicm,si-fix-vgpr-copies,si-optimize-exec-masking,remove-redundant-debug-values,fixup-statepoint-caller-saved,postra-machine-sink,shrink-wrap,prolog-epilog,branch-folder,tailduplication,machine-latecleanup,machine-cp,post-ra-pseudos,postmisched,block-placement,fentry-insert,xray-instrumentation,patchable-function,gcn-create-vopd,si-memory-legalizer,si-insert-waitcnts,si-late-branch-lowering,si-pre-emit-peephole,post-RA-hazard-rec,amdgpu-wait-sgpr-hazards,amdgpu-insert-delay-alu,branch-relaxation,reg-usage-collector,remove-loads-into-fake-uses,live-debug-values,machine-sanmd,stack-frame-layout,verify),invalidate<machine-function-info>))
+; GCN-O2: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,function(amdgpu-image-intrinsic-opt),expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(infer-address-spaces,amdgpu-atomic-optimizer,atomic-expand,amdgpu-promote-alloca,separate-const-offset-from-gep<>,slsr,early-cse<>,nary-reassociate,early-cse<>,amdgpu-codegenprepare,verify,loop-mssa(loop-reduce),mergeicmps,expand-memcmp,gc-lowering,lower-constant-intrinsics,unreachableblockelim,consthoist,replace-with-veclib,partially-inline-libcalls,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,early-cse<>,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(codegenprepare,load-store-vectorizer,lower-switch,lower-invoke,unreachableblockelim,flatten-cfg,sink,amdgpu-late-codegenprepare,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-control-flow,amdgpu-rewrite-undef-for-phi,lcssa)),amdgpu-perf-hint,cgscc(function(require<uniformity>,callbr-prepare,safe-stack,stack-protector,verify)),cgscc(function(machine-function(amdgpu-isel,si-fix-sgpr-copies,si-i1-copies,finalize-isel,early-tailduplication,opt-phis,stack-coloring,localstackalloc,dead-mi-elimination,early-machinelicm,machine-cse,machine-sink,peephole-opt,dead-mi-elimination,si-fold-operands,gcn-dpp-combine,si-load-store-opt,si-peephole-sdwa,early-machinelicm,machine-cse,si-fold-operands,dead-mi-elimination,si-shrink-instructions))),require<reg-usage>,cgscc(function(machine-function(reg-usage-propagation,detect-dead-lanes,dead-mi-elimination,init-undef,process-imp-defs,unreachable-mbb-elimination,require<live-vars>,si-opt-vgpr-liverange,require<machine-loops>,phi-node-elimination,si-lower-control-flow,two-address-instruction,register-coalescer,rename-independent-subregs,amdgpu-rewrite-partial-reg-uses,machine-scheduler,amdgpu-pre-ra-optimizations,si-wqm,si-optimize-exec-masking-pre-ra,si-form-memory-clauses,amdgpu-pre-ra-long-branch-reg,greedy<sgpr>,virt-reg-rewriter<no-clear-vregs>,stack-slot-coloring,si-lower-sgpr-spills,si-pre-allocate-wwm-regs,greedy<wwm>,si-lower-wwm-copies,virt-reg-rewriter<no-clear-vregs>,amdgpu-reserve-wwm-regs,greedy<vgpr>,amdgpu-nsa-reassign,virt-reg-rewriter,amdgpu-mark-last-scratch-load,machine-cp,machinelicm,si-fix-vgpr-copies,si-optimize-exec-masking,remove-redundant-debug-values,fixup-statepoint-caller-saved,postra-machine-sink,shrink-wrap,prolog-epilog,branch-folder,tailduplication,machine-latecleanup,machine-cp,post-ra-pseudos,postmisched,block-placement,fentry-insert,xray-instrumentation,patchable-function,gcn-create-vopd,si-memory-legalizer,si-insert-waitcnts,si-late-branch-lowering,si-pre-emit-peephole,post-RA-hazard-rec,amdgpu-wait-sgpr-hazards,amdgpu-insert-delay-alu,branch-relaxation,reg-usage-collector,remove-loads-into-fake-uses,live-debug-values,machine-sanmd,stack-frame-layout,verify),invalidate<machine-function-info>))
-; GCN-O3: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,function(amdgpu-image-intrinsic-opt),expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(infer-address-spaces,amdgpu-atomic-optimizer,atomic-expand,amdgpu-promote-alloca,separate-const-offset-from-gep<>,slsr,gvn<>,nary-reassociate,early-cse<>,amdgpu-codegenprepare,verify,loop-mssa(loop-reduce),mergeicmps,expand-memcmp,gc-lowering,lower-constant-intrinsics,unreachableblockelim,consthoist,replace-with-veclib,partially-inline-libcalls,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,gvn<>,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(codegenprepare,load-store-vectorizer,lower-switch,lower-invoke,unreachableblockelim,flatten-cfg,sink,amdgpu-late-codegenprepare,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-control-flow,amdgpu-rewrite-undef-for-phi,lcssa)),amdgpu-perf-hint,cgscc(function(require<uniformity>,callbr-prepare,safe-stack,stack-protector,verify)),cgscc(function(machine-function(amdgpu-isel,si-fix-sgpr-copies,si-i1-copies,finalize-isel,early-tailduplication,opt-phis,stack-coloring,localstackalloc,dead-mi-elimination,early-machinelicm,machine-cse,machine-sink,peephole-opt,dead-mi-elimination,si-fold-operands,gcn-dpp-combine,si-load-store-opt,si-peephole-sdwa,early-machinelicm,machine-cse,si-fold-operands,dead-mi-elimination,si-shrink-instructions))),require<reg-usage>,cgscc(function(machine-function(reg-usage-propagation,detect-dead-lanes,dead-mi-elimination,init-undef,ProcessImplicitDefsPass,unreachable-mbb-elimination,require<live-vars>,si-opt-vgpr-liverange,require<machine-loops>,phi-node-elimination,si-lower-control-flow,two-address-instruction,register-coalescer,rename-independent-subregs,amdgpu-rewrite-partial-reg-uses,machine-scheduler,amdgpu-pre-ra-optimizations,si-wqm,si-optimize-exec-masking-pre-ra,si-form-memory-clauses,amdgpu-pre-ra-long-branch-reg,greedy<sgpr>,virt-reg-rewriter<no-clear-vregs>,stack-slot-coloring,si-lower-sgpr-spills,si-pre-allocate-wwm-regs,greedy<wwm>,si-lower-wwm-copies,virt-reg-rewriter<no-clear-vregs>,amdgpu-reserve-wwm-regs,greedy<vgpr>,amdgpu-nsa-reassign,virt-reg-rewriter,amdgpu-mark-last-scratch-load,machine-cp,machinelicm,si-fix-vgpr-copies,si-optimize-exec-masking,remove-redundant-debug-values,fixup-statepoint-caller-saved,postra-machine-sink,shrink-wrap,prolog-epilog,branch-folder,tailduplication,machine-latecleanup,machine-cp,post-ra-pseudos,postmisched,block-placement,fentry-insert,xray-instrumentation,patchable-function,gcn-create-vopd,si-memory-legalizer,si-insert-waitcnts,si-late-branch-lowering,si-pre-emit-peephole,post-RA-hazard-rec,amdgpu-wait-sgpr-hazards,amdgpu-insert-delay-alu,branch-relaxation,reg-usage-collector,remove-loads-into-fake-uses,live-debug-values,machine-sanmd,stack-frame-layout,verify),invalidate<machine-function-info>))
+; GCN-O3: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,function(amdgpu-image-intrinsic-opt),expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(infer-address-spaces,amdgpu-atomic-optimizer,atomic-expand,amdgpu-promote-alloca,separate-const-offset-from-gep<>,slsr,gvn<>,nary-reassociate,early-cse<>,amdgpu-codegenprepare,verify,loop-mssa(loop-reduce),mergeicmps,expand-memcmp,gc-lowering,lower-constant-intrinsics,unreachableblockelim,consthoist,replace-with-veclib,partially-inline-libcalls,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,gvn<>,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(codegenprepare,load-store-vectorizer,lower-switch,lower-invoke,unreachableblockelim,flatten-cfg,sink,amdgpu-late-codegenprepare,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-contr...
[truncated]
|
|
@llvm/pr-subscribers-backend-x86 Author: Akshat Oke (optimisan) ChangesAMDGPU/llc-pipeline-npm.ll test update is because the dummy class name is now changed to its CL name. Patch is 22.56 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/138829.diff 8 Files Affected:
diff --git a/llvm/include/llvm/InitializePasses.h b/llvm/include/llvm/InitializePasses.h
index e75f9c7a2cfe8..9a380994c2886 100644
--- a/llvm/include/llvm/InitializePasses.h
+++ b/llvm/include/llvm/InitializePasses.h
@@ -246,7 +246,7 @@ void initializePostRASchedulerLegacyPass(PassRegistry &);
void initializePreISelIntrinsicLoweringLegacyPassPass(PassRegistry &);
void initializePrintFunctionPassWrapperPass(PassRegistry &);
void initializePrintModulePassWrapperPass(PassRegistry &);
-void initializeProcessImplicitDefsPass(PassRegistry &);
+void initializeProcessImplicitDefsLegacyPass(PassRegistry &);
void initializeProfileSummaryInfoWrapperPassPass(PassRegistry &);
void initializePromoteLegacyPassPass(PassRegistry &);
void initializeRABasicPass(PassRegistry &);
diff --git a/llvm/include/llvm/Passes/CodeGenPassBuilder.h b/llvm/include/llvm/Passes/CodeGenPassBuilder.h
index 648437fe12c91..0704ae3b2ec1d 100644
--- a/llvm/include/llvm/Passes/CodeGenPassBuilder.h
+++ b/llvm/include/llvm/Passes/CodeGenPassBuilder.h
@@ -71,6 +71,7 @@
#include "llvm/CodeGen/PeepholeOptimizer.h"
#include "llvm/CodeGen/PostRASchedulerList.h"
#include "llvm/CodeGen/PreISelIntrinsicLowering.h"
+#include "llvm/CodeGen/ProcessImplicitDefs.h"
#include "llvm/CodeGen/RegAllocEvictionAdvisor.h"
#include "llvm/CodeGen/RegAllocFast.h"
#include "llvm/CodeGen/RegAllocGreedyPass.h"
diff --git a/llvm/include/llvm/Passes/MachinePassRegistry.def b/llvm/include/llvm/Passes/MachinePassRegistry.def
index 8717b79b26968..9d75e4642db00 100644
--- a/llvm/include/llvm/Passes/MachinePassRegistry.def
+++ b/llvm/include/llvm/Passes/MachinePassRegistry.def
@@ -184,6 +184,7 @@ MACHINE_FUNCTION_PASS("print<machine-uniformity>",
MachineUniformityPrinterPass(errs()))
MACHINE_FUNCTION_PASS("print<slot-indexes>", SlotIndexesPrinterPass(errs()))
MACHINE_FUNCTION_PASS("print<virtregmap>", VirtRegMapPrinterPass(errs()))
+MACHINE_FUNCTION_PASS("process-imp-defs", ProcessImplicitDefsPass())
MACHINE_FUNCTION_PASS("prolog-epilog", PrologEpilogInserterPass())
MACHINE_FUNCTION_PASS("reg-usage-collector", RegUsageInfoCollectorPass())
MACHINE_FUNCTION_PASS("reg-usage-propagation", RegUsageInfoPropagationPass())
@@ -316,7 +317,6 @@ DUMMY_MACHINE_FUNCTION_PASS("static-data-splitter", StaticDataSplitter)
DUMMY_MACHINE_FUNCTION_PASS("machine-function-splitter", MachineFunctionSplitterPass)
DUMMY_MACHINE_FUNCTION_PASS("machineinstr-printer", MachineFunctionPrinterPass)
DUMMY_MACHINE_FUNCTION_PASS("mirfs-discriminators", MIRAddFSDiscriminatorsPass)
-DUMMY_MACHINE_FUNCTION_PASS("processimpdefs", ProcessImplicitDefsPass)
DUMMY_MACHINE_FUNCTION_PASS("prologepilog-code", PrologEpilogCodeInserterPass)
DUMMY_MACHINE_FUNCTION_PASS("ra-basic", RABasicPass)
DUMMY_MACHINE_FUNCTION_PASS("ra-pbqp", RAPBQPPass)
diff --git a/llvm/lib/CodeGen/CodeGen.cpp b/llvm/lib/CodeGen/CodeGen.cpp
index 065fd4704ccfb..6e12edc3d3a5b 100644
--- a/llvm/lib/CodeGen/CodeGen.cpp
+++ b/llvm/lib/CodeGen/CodeGen.cpp
@@ -110,7 +110,7 @@ void llvm::initializeCodeGen(PassRegistry &Registry) {
initializePostRAMachineSinkingLegacyPass(Registry);
initializePostRASchedulerLegacyPass(Registry);
initializePreISelIntrinsicLoweringLegacyPassPass(Registry);
- initializeProcessImplicitDefsPass(Registry);
+ initializeProcessImplicitDefsLegacyPass(Registry);
initializeRABasicPass(Registry);
initializeRAGreedyLegacyPass(Registry);
initializeRegAllocFastPass(Registry);
diff --git a/llvm/lib/CodeGen/ProcessImplicitDefs.cpp b/llvm/lib/CodeGen/ProcessImplicitDefs.cpp
index be81ecab9c897..54fd7814ef4f3 100644
--- a/llvm/lib/CodeGen/ProcessImplicitDefs.cpp
+++ b/llvm/lib/CodeGen/ProcessImplicitDefs.cpp
@@ -6,6 +6,7 @@
//
//===----------------------------------------------------------------------===//
+#include "llvm/CodeGen/ProcessImplicitDefs.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
@@ -26,24 +27,15 @@ using namespace llvm;
namespace {
/// Process IMPLICIT_DEF instructions and make sure there is one implicit_def
/// for each use. Add isUndef marker to implicit_def defs and their uses.
-class ProcessImplicitDefs : public MachineFunctionPass {
- const TargetInstrInfo *TII = nullptr;
- const TargetRegisterInfo *TRI = nullptr;
- MachineRegisterInfo *MRI = nullptr;
-
- SmallSetVector<MachineInstr*, 16> WorkList;
-
- void processImplicitDef(MachineInstr *MI);
- bool canTurnIntoImplicitDef(MachineInstr *MI);
-
+class ProcessImplicitDefsLegacy : public MachineFunctionPass {
public:
static char ID;
- ProcessImplicitDefs() : MachineFunctionPass(ID) {
- initializeProcessImplicitDefsPass(*PassRegistry::getPassRegistry());
+ ProcessImplicitDefsLegacy() : MachineFunctionPass(ID) {
+ initializeProcessImplicitDefsLegacyPass(*PassRegistry::getPassRegistry());
}
- void getAnalysisUsage(AnalysisUsage &au) const override;
+ void getAnalysisUsage(AnalysisUsage &AU) const override;
bool runOnMachineFunction(MachineFunction &MF) override;
@@ -52,15 +44,29 @@ class ProcessImplicitDefs : public MachineFunctionPass {
MachineFunctionProperties::Property::IsSSA);
}
};
+
+class ProcessImplicitDefs {
+ const TargetInstrInfo *TII = nullptr;
+ const TargetRegisterInfo *TRI = nullptr;
+ MachineRegisterInfo *MRI = nullptr;
+
+ SmallSetVector<MachineInstr *, 16> WorkList;
+
+ void processImplicitDef(MachineInstr *MI);
+ bool canTurnIntoImplicitDef(MachineInstr *MI);
+
+public:
+ bool run(MachineFunction &MF);
+};
} // end anonymous namespace
-char ProcessImplicitDefs::ID = 0;
-char &llvm::ProcessImplicitDefsID = ProcessImplicitDefs::ID;
+char ProcessImplicitDefsLegacy::ID = 0;
+char &llvm::ProcessImplicitDefsID = ProcessImplicitDefsLegacy::ID;
-INITIALIZE_PASS(ProcessImplicitDefs, DEBUG_TYPE,
+INITIALIZE_PASS(ProcessImplicitDefsLegacy, DEBUG_TYPE,
"Process Implicit Definitions", false, false)
-void ProcessImplicitDefs::getAnalysisUsage(AnalysisUsage &AU) const {
+void ProcessImplicitDefsLegacy::getAnalysisUsage(AnalysisUsage &AU) const {
AU.setPreservesCFG();
AU.addPreserved<AAResultsWrapperPass>();
MachineFunctionPass::getAnalysisUsage(AU);
@@ -133,9 +139,24 @@ void ProcessImplicitDefs::processImplicitDef(MachineInstr *MI) {
LLVM_DEBUG(dbgs() << "Keeping physreg: " << *MI);
}
+bool ProcessImplicitDefsLegacy::runOnMachineFunction(MachineFunction &MF) {
+ return ProcessImplicitDefs().run(MF);
+}
+
+PreservedAnalyses
+ProcessImplicitDefsPass::run(MachineFunction &MF,
+ MachineFunctionAnalysisManager &MFAM) {
+ if (!ProcessImplicitDefs().run(MF))
+ return PreservedAnalyses::all();
+
+ return getMachineFunctionPassPreservedAnalyses()
+ .preserveSet<CFGAnalyses>()
+ .preserve<AAManager>();
+}
+
/// processImplicitDefs - Process IMPLICIT_DEF instructions and turn them into
/// <undef> operands.
-bool ProcessImplicitDefs::runOnMachineFunction(MachineFunction &MF) {
+bool ProcessImplicitDefs::run(MachineFunction &MF) {
LLVM_DEBUG(dbgs() << "********** PROCESS IMPLICIT DEFS **********\n"
<< "********** Function: " << MF.getName() << '\n');
diff --git a/llvm/lib/Passes/PassBuilder.cpp b/llvm/lib/Passes/PassBuilder.cpp
index a9d192a7fad55..29cbaf62c2d88 100644
--- a/llvm/lib/Passes/PassBuilder.cpp
+++ b/llvm/lib/Passes/PassBuilder.cpp
@@ -145,6 +145,7 @@
#include "llvm/CodeGen/PostRAHazardRecognizer.h"
#include "llvm/CodeGen/PostRASchedulerList.h"
#include "llvm/CodeGen/PreISelIntrinsicLowering.h"
+#include "llvm/CodeGen/ProcessImplicitDefs.h"
#include "llvm/CodeGen/RegAllocEvictionAdvisor.h"
#include "llvm/CodeGen/RegAllocFast.h"
#include "llvm/CodeGen/RegAllocGreedyPass.h"
diff --git a/llvm/test/CodeGen/AMDGPU/llc-pipeline-npm.ll b/llvm/test/CodeGen/AMDGPU/llc-pipeline-npm.ll
index 634a7fb6eb8e9..9c1d3dcf080ee 100644
--- a/llvm/test/CodeGen/AMDGPU/llc-pipeline-npm.ll
+++ b/llvm/test/CodeGen/AMDGPU/llc-pipeline-npm.ll
@@ -9,9 +9,9 @@
; GCN-O0: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(atomic-expand,verify,gc-lowering,lower-constant-intrinsics,unreachableblockelim,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(lower-switch,lower-invoke,unreachableblockelim,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-control-flow,amdgpu-rewrite-undef-for-phi,lcssa,require<uniformity>,callbr-prepare,safe-stack,stack-protector,verify)),cgscc(function(machine-function(amdgpu-isel,si-fix-sgpr-copies,si-i1-copies,finalize-isel,localstackalloc))),require<reg-usage>,cgscc(function(machine-function(reg-usage-propagation,phi-node-elimination,two-address-instruction,regallocfast,si-fix-vgpr-copies,remove-redundant-debug-values,fixup-statepoint-caller-saved,prolog-epilog,post-ra-pseudos,fentry-insert,xray-instrumentation,patchable-function,si-memory-legalizer,si-insert-waitcnts,si-late-branch-lowering,post-RA-hazard-rec,amdgpu-wait-sgpr-hazards,branch-relaxation,reg-usage-collector,remove-loads-into-fake-uses,live-debug-values,machine-sanmd,stack-frame-layout,verify),invalidate<machine-function-info>))
-; GCN-O2: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,function(amdgpu-image-intrinsic-opt),expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(infer-address-spaces,amdgpu-atomic-optimizer,atomic-expand,amdgpu-promote-alloca,separate-const-offset-from-gep<>,slsr,early-cse<>,nary-reassociate,early-cse<>,amdgpu-codegenprepare,verify,loop-mssa(loop-reduce),mergeicmps,expand-memcmp,gc-lowering,lower-constant-intrinsics,unreachableblockelim,consthoist,replace-with-veclib,partially-inline-libcalls,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,early-cse<>,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(codegenprepare,load-store-vectorizer,lower-switch,lower-invoke,unreachableblockelim,flatten-cfg,sink,amdgpu-late-codegenprepare,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-control-flow,amdgpu-rewrite-undef-for-phi,lcssa)),amdgpu-perf-hint,cgscc(function(require<uniformity>,callbr-prepare,safe-stack,stack-protector,verify)),cgscc(function(machine-function(amdgpu-isel,si-fix-sgpr-copies,si-i1-copies,finalize-isel,early-tailduplication,opt-phis,stack-coloring,localstackalloc,dead-mi-elimination,early-machinelicm,machine-cse,machine-sink,peephole-opt,dead-mi-elimination,si-fold-operands,gcn-dpp-combine,si-load-store-opt,si-peephole-sdwa,early-machinelicm,machine-cse,si-fold-operands,dead-mi-elimination,si-shrink-instructions))),require<reg-usage>,cgscc(function(machine-function(reg-usage-propagation,detect-dead-lanes,dead-mi-elimination,init-undef,ProcessImplicitDefsPass,unreachable-mbb-elimination,require<live-vars>,si-opt-vgpr-liverange,require<machine-loops>,phi-node-elimination,si-lower-control-flow,two-address-instruction,register-coalescer,rename-independent-subregs,amdgpu-rewrite-partial-reg-uses,machine-scheduler,amdgpu-pre-ra-optimizations,si-wqm,si-optimize-exec-masking-pre-ra,si-form-memory-clauses,amdgpu-pre-ra-long-branch-reg,greedy<sgpr>,virt-reg-rewriter<no-clear-vregs>,stack-slot-coloring,si-lower-sgpr-spills,si-pre-allocate-wwm-regs,greedy<wwm>,si-lower-wwm-copies,virt-reg-rewriter<no-clear-vregs>,amdgpu-reserve-wwm-regs,greedy<vgpr>,amdgpu-nsa-reassign,virt-reg-rewriter,amdgpu-mark-last-scratch-load,machine-cp,machinelicm,si-fix-vgpr-copies,si-optimize-exec-masking,remove-redundant-debug-values,fixup-statepoint-caller-saved,postra-machine-sink,shrink-wrap,prolog-epilog,branch-folder,tailduplication,machine-latecleanup,machine-cp,post-ra-pseudos,postmisched,block-placement,fentry-insert,xray-instrumentation,patchable-function,gcn-create-vopd,si-memory-legalizer,si-insert-waitcnts,si-late-branch-lowering,si-pre-emit-peephole,post-RA-hazard-rec,amdgpu-wait-sgpr-hazards,amdgpu-insert-delay-alu,branch-relaxation,reg-usage-collector,remove-loads-into-fake-uses,live-debug-values,machine-sanmd,stack-frame-layout,verify),invalidate<machine-function-info>))
+; GCN-O2: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,function(amdgpu-image-intrinsic-opt),expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(infer-address-spaces,amdgpu-atomic-optimizer,atomic-expand,amdgpu-promote-alloca,separate-const-offset-from-gep<>,slsr,early-cse<>,nary-reassociate,early-cse<>,amdgpu-codegenprepare,verify,loop-mssa(loop-reduce),mergeicmps,expand-memcmp,gc-lowering,lower-constant-intrinsics,unreachableblockelim,consthoist,replace-with-veclib,partially-inline-libcalls,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,early-cse<>,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(codegenprepare,load-store-vectorizer,lower-switch,lower-invoke,unreachableblockelim,flatten-cfg,sink,amdgpu-late-codegenprepare,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-control-flow,amdgpu-rewrite-undef-for-phi,lcssa)),amdgpu-perf-hint,cgscc(function(require<uniformity>,callbr-prepare,safe-stack,stack-protector,verify)),cgscc(function(machine-function(amdgpu-isel,si-fix-sgpr-copies,si-i1-copies,finalize-isel,early-tailduplication,opt-phis,stack-coloring,localstackalloc,dead-mi-elimination,early-machinelicm,machine-cse,machine-sink,peephole-opt,dead-mi-elimination,si-fold-operands,gcn-dpp-combine,si-load-store-opt,si-peephole-sdwa,early-machinelicm,machine-cse,si-fold-operands,dead-mi-elimination,si-shrink-instructions))),require<reg-usage>,cgscc(function(machine-function(reg-usage-propagation,detect-dead-lanes,dead-mi-elimination,init-undef,process-imp-defs,unreachable-mbb-elimination,require<live-vars>,si-opt-vgpr-liverange,require<machine-loops>,phi-node-elimination,si-lower-control-flow,two-address-instruction,register-coalescer,rename-independent-subregs,amdgpu-rewrite-partial-reg-uses,machine-scheduler,amdgpu-pre-ra-optimizations,si-wqm,si-optimize-exec-masking-pre-ra,si-form-memory-clauses,amdgpu-pre-ra-long-branch-reg,greedy<sgpr>,virt-reg-rewriter<no-clear-vregs>,stack-slot-coloring,si-lower-sgpr-spills,si-pre-allocate-wwm-regs,greedy<wwm>,si-lower-wwm-copies,virt-reg-rewriter<no-clear-vregs>,amdgpu-reserve-wwm-regs,greedy<vgpr>,amdgpu-nsa-reassign,virt-reg-rewriter,amdgpu-mark-last-scratch-load,machine-cp,machinelicm,si-fix-vgpr-copies,si-optimize-exec-masking,remove-redundant-debug-values,fixup-statepoint-caller-saved,postra-machine-sink,shrink-wrap,prolog-epilog,branch-folder,tailduplication,machine-latecleanup,machine-cp,post-ra-pseudos,postmisched,block-placement,fentry-insert,xray-instrumentation,patchable-function,gcn-create-vopd,si-memory-legalizer,si-insert-waitcnts,si-late-branch-lowering,si-pre-emit-peephole,post-RA-hazard-rec,amdgpu-wait-sgpr-hazards,amdgpu-insert-delay-alu,branch-relaxation,reg-usage-collector,remove-loads-into-fake-uses,live-debug-values,machine-sanmd,stack-frame-layout,verify),invalidate<machine-function-info>))
-; GCN-O3: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,function(amdgpu-image-intrinsic-opt),expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(infer-address-spaces,amdgpu-atomic-optimizer,atomic-expand,amdgpu-promote-alloca,separate-const-offset-from-gep<>,slsr,gvn<>,nary-reassociate,early-cse<>,amdgpu-codegenprepare,verify,loop-mssa(loop-reduce),mergeicmps,expand-memcmp,gc-lowering,lower-constant-intrinsics,unreachableblockelim,consthoist,replace-with-veclib,partially-inline-libcalls,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,gvn<>,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(codegenprepare,load-store-vectorizer,lower-switch,lower-invoke,unreachableblockelim,flatten-cfg,sink,amdgpu-late-codegenprepare,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-control-flow,amdgpu-rewrite-undef-for-phi,lcssa)),amdgpu-perf-hint,cgscc(function(require<uniformity>,callbr-prepare,safe-stack,stack-protector,verify)),cgscc(function(machine-function(amdgpu-isel,si-fix-sgpr-copies,si-i1-copies,finalize-isel,early-tailduplication,opt-phis,stack-coloring,localstackalloc,dead-mi-elimination,early-machinelicm,machine-cse,machine-sink,peephole-opt,dead-mi-elimination,si-fold-operands,gcn-dpp-combine,si-load-store-opt,si-peephole-sdwa,early-machinelicm,machine-cse,si-fold-operands,dead-mi-elimination,si-shrink-instructions))),require<reg-usage>,cgscc(function(machine-function(reg-usage-propagation,detect-dead-lanes,dead-mi-elimination,init-undef,ProcessImplicitDefsPass,unreachable-mbb-elimination,require<live-vars>,si-opt-vgpr-liverange,require<machine-loops>,phi-node-elimination,si-lower-control-flow,two-address-instruction,register-coalescer,rename-independent-subregs,amdgpu-rewrite-partial-reg-uses,machine-scheduler,amdgpu-pre-ra-optimizations,si-wqm,si-optimize-exec-masking-pre-ra,si-form-memory-clauses,amdgpu-pre-ra-long-branch-reg,greedy<sgpr>,virt-reg-rewriter<no-clear-vregs>,stack-slot-coloring,si-lower-sgpr-spills,si-pre-allocate-wwm-regs,greedy<wwm>,si-lower-wwm-copies,virt-reg-rewriter<no-clear-vregs>,amdgpu-reserve-wwm-regs,greedy<vgpr>,amdgpu-nsa-reassign,virt-reg-rewriter,amdgpu-mark-last-scratch-load,machine-cp,machinelicm,si-fix-vgpr-copies,si-optimize-exec-masking,remove-redundant-debug-values,fixup-statepoint-caller-saved,postra-machine-sink,shrink-wrap,prolog-epilog,branch-folder,tailduplication,machine-latecleanup,machine-cp,post-ra-pseudos,postmisched,block-placement,fentry-insert,xray-instrumentation,patchable-function,gcn-create-vopd,si-memory-legalizer,si-insert-waitcnts,si-late-branch-lowering,si-pre-emit-peephole,post-RA-hazard-rec,amdgpu-wait-sgpr-hazards,amdgpu-insert-delay-alu,branch-relaxation,reg-usage-collector,remove-loads-into-fake-uses,live-debug-values,machine-sanmd,stack-frame-layout,verify),invalidate<machine-function-info>))
+; GCN-O3: require<MachineModuleAnalysis>,require<profile-summary>,require<collector-metadata>,pre-isel-intrinsic-lowering,function(expand-large-div-rem,expand-fp),amdgpu-remove-incompatible-functions,amdgpu-printf-runtime-binding,amdgpu-lower-ctor-dtor,function(amdgpu-image-intrinsic-opt),expand-variadics,amdgpu-always-inline,always-inline,amdgpu-export-kernel-runtime-handles,amdgpu-sw-lower-lds,amdgpu-lower-module-lds,function(infer-address-spaces,amdgpu-atomic-optimizer,atomic-expand,amdgpu-promote-alloca,separate-const-offset-from-gep<>,slsr,gvn<>,nary-reassociate,early-cse<>,amdgpu-codegenprepare,verify,loop-mssa(loop-reduce),mergeicmps,expand-memcmp,gc-lowering,lower-constant-intrinsics,unreachableblockelim,consthoist,replace-with-veclib,partially-inline-libcalls,ee-instrument<post-inline>,scalarize-masked-mem-intrin,expand-reductions,gvn<>,amdgpu-lower-kernel-arguments),amdgpu-lower-buffer-fat-pointers,cgscc(function(codegenprepare,load-store-vectorizer,lower-switch,lower-invoke,unreachableblockelim,flatten-cfg,sink,amdgpu-late-codegenprepare,amdgpu-unify-divergent-exit-nodes,fix-irreducible,unify-loop-exits,StructurizeCFGPass,amdgpu-annotate-uniform,si-annotate-contr...
[truncated]
|
arsenm
approved these changes
May 7, 2025
optimisan
force-pushed
the
users/optimisan/pb/register-f-passes
branch
from
9b03fcd to
875c983
Compare
optimisan
force-pushed
the
users/optimisan/port/process-imp-defs
branch
from
4c6166a to
0977795
Compare
This was referenced May 12, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
AMDGPU/llc-pipeline-npm.ll test update is because the dummy class name is now changed to its CL name.