llvm · paperchalice · Apr 30, 2025 · Apr 28, 2025 · Apr 28, 2025 · Apr 28, 2025
diff --git a/llvm/include/llvm/CodeGen/MachineUniformityAnalysis.h b/llvm/include/llvm/CodeGen/MachineUniformityAnalysis.h
@@ -17,6 +17,7 @@
 #include "llvm/ADT/GenericUniformityInfo.h"
 #include "llvm/CodeGen/MachineCycleAnalysis.h"
 #include "llvm/CodeGen/MachineDominators.h"
+#include "llvm/CodeGen/MachinePassManager.h"
 #include "llvm/CodeGen/MachineSSAContext.h"
 
 namespace llvm {
@@ -51,6 +52,27 @@ class MachineUniformityAnalysisPass : public MachineFunctionPass {
   // TODO: verify analysis
 };
 
+class MachineUniformityAnalysis
+    : public AnalysisInfoMixin<MachineUniformityAnalysis> {
+  friend AnalysisInfoMixin<MachineUniformityAnalysis>;
+  static AnalysisKey Key;
+
+public:
+  using Result = MachineUniformityInfo;
+  Result run(MachineFunction &MF, MachineFunctionAnalysisManager &MFAM);
+};
+
+class MachineUniformityPrinterPass
+    : public PassInfoMixin<MachineUniformityAnalysis> {
+  raw_ostream &OS;
+
+public:
+  explicit MachineUniformityPrinterPass(raw_ostream &OS) : OS(OS) {}
+  PreservedAnalyses run(MachineFunction &MF,
+                        MachineFunctionAnalysisManager &MFAM);
+  static bool isRequired() { return true; }
+};
+
 } // namespace llvm
 
 #endif // LLVM_CODEGEN_MACHINEUNIFORMITYANALYSIS_H
diff --git a/llvm/include/llvm/Passes/MachinePassRegistry.def b/llvm/include/llvm/Passes/MachinePassRegistry.def
@@ -114,6 +114,7 @@ MACHINE_FUNCTION_ANALYSIS("machine-opt-remark-emitter",
 MACHINE_FUNCTION_ANALYSIS("machine-post-dom-tree",
                           MachinePostDominatorTreeAnalysis())
 MACHINE_FUNCTION_ANALYSIS("machine-trace-metrics", MachineTraceMetricsAnalysis())
+MACHINE_FUNCTION_ANALYSIS("machine-uniformity", MachineUniformityAnalysis())
 MACHINE_FUNCTION_ANALYSIS("pass-instrumentation", PassInstrumentationAnalysis(PIC))
 MACHINE_FUNCTION_ANALYSIS("regalloc-evict", RegAllocEvictionAdvisorAnalysis())
 MACHINE_FUNCTION_ANALYSIS("regalloc-priority", RegAllocPriorityAdvisorAnalysis())
@@ -178,6 +179,8 @@ MACHINE_FUNCTION_PASS("print<machine-dom-tree>",
 MACHINE_FUNCTION_PASS("print<machine-loops>", MachineLoopPrinterPass(errs()))
 MACHINE_FUNCTION_PASS("print<machine-post-dom-tree>",
                       MachinePostDominatorTreePrinterPass(errs()))
+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("reg-usage-collector", RegUsageInfoCollectorPass())
@@ -302,11 +305,9 @@ DUMMY_MACHINE_FUNCTION_PASS("lrshrink", LiveRangeShrinkPass)
 DUMMY_MACHINE_FUNCTION_PASS("machine-combiner", MachineCombinerPass)
 DUMMY_MACHINE_FUNCTION_PASS("static-data-splitter", StaticDataSplitter)
 DUMMY_MACHINE_FUNCTION_PASS("machine-function-splitter", MachineFunctionSplitterPass)
-DUMMY_MACHINE_FUNCTION_PASS("machine-uniformity", MachineUniformityInfoWrapperPass)
 DUMMY_MACHINE_FUNCTION_PASS("machineinstr-printer", MachineFunctionPrinterPass)
 DUMMY_MACHINE_FUNCTION_PASS("mirfs-discriminators", MIRAddFSDiscriminatorsPass)
 DUMMY_MACHINE_FUNCTION_PASS("postra-machine-sink", PostRAMachineSinkingPass)
-DUMMY_MACHINE_FUNCTION_PASS("print-machine-uniformity", MachineUniformityInfoPrinterPass)
 DUMMY_MACHINE_FUNCTION_PASS("processimpdefs", ProcessImplicitDefsPass)
 DUMMY_MACHINE_FUNCTION_PASS("prologepilog", PrologEpilogInserterPass)
 DUMMY_MACHINE_FUNCTION_PASS("prologepilog-code", PrologEpilogCodeInserterPass)

diff --git a/llvm/lib/CodeGen/MachineUniformityAnalysis.cpp b/llvm/lib/CodeGen/MachineUniformityAnalysis.cpp
@@ -8,6 +8,7 @@
 
 #include "llvm/CodeGen/MachineUniformityAnalysis.h"
 #include "llvm/ADT/GenericUniformityImpl.h"
+#include "llvm/Analysis/TargetTransformInfo.h"
 #include "llvm/CodeGen/MachineCycleAnalysis.h"
 #include "llvm/CodeGen/MachineDominators.h"
 #include "llvm/CodeGen/MachineRegisterInfo.h"
@@ -177,6 +178,32 @@ class MachineUniformityInfoPrinterPass : public MachineFunctionPass {
 
 } // namespace
 
+AnalysisKey MachineUniformityAnalysis::Key;
+
+MachineUniformityAnalysis::Result
+MachineUniformityAnalysis::run(MachineFunction &MF,
+                               MachineFunctionAnalysisManager &MFAM) {
+  auto &DomTree = MFAM.getResult<MachineDominatorTreeAnalysis>(MF);
+  auto &CI = MFAM.getResult<MachineCycleAnalysis>(MF);
+  auto &FAM = MFAM.getResult<FunctionAnalysisManagerMachineFunctionProxy>(MF)
+                  .getManager();
+  auto &F = MF.getFunction();
+  auto &TTI = FAM.getResult<TargetIRAnalysis>(F);
+  return computeMachineUniformityInfo(MF, CI, DomTree,
+                                      TTI.hasBranchDivergence(&F));
+}
+
+PreservedAnalyses
+MachineUniformityPrinterPass::run(MachineFunction &MF,
+                                  MachineFunctionAnalysisManager &MFAM) {
+  auto &MUI = MFAM.getResult<MachineUniformityAnalysis>(MF);
+  OS << "MachineUniformityInfo for function: ";
+  MF.getFunction().printAsOperand(OS, /*PrintType=*/false);
+  OS << '\n';
+  MUI.print(OS);
+  return PreservedAnalyses::all();
+}
+
 char MachineUniformityAnalysisPass::ID = 0;
 
 MachineUniformityAnalysisPass::MachineUniformityAnalysisPass()

diff --git a/llvm/lib/Passes/PassBuilder.cpp b/llvm/lib/Passes/PassBuilder.cpp
@@ -133,6 +133,7 @@
 #include "llvm/CodeGen/MachineScheduler.h"
 #include "llvm/CodeGen/MachineSink.h"
 #include "llvm/CodeGen/MachineTraceMetrics.h"
+#include "llvm/CodeGen/MachineUniformityAnalysis.h"
 #include "llvm/CodeGen/MachineVerifier.h"
 #include "llvm/CodeGen/OptimizePHIs.h"
 #include "llvm/CodeGen/PHIElimination.h"

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/always-uniform-gmir.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/always-uniform-gmir.mir
@@ -1,5 +1,6 @@
 # NOTE: This file is Generic MIR translation of test/Analysis/UniformityAnalysis/AMDGPU/always_uniform.ll test file
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 ---
 name:            readfirstlane
 body:             |

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/always-uniform.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/always-uniform.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -mcpu=gfx900 -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -mcpu=gfx900 -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 # readlane, readfirstlane is always uniform
 

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/atomics-gmir.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/atomics-gmir.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 ---
 name:            test1

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/atomics.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/atomics.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 ---
 name:            test1

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/control-flow-intrinsics.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/control-flow-intrinsics.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -mcpu=gfx900 -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -mcpu=gfx900 -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 ---
 name:            f1

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/hidden-diverge-gmir.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/hidden-diverge-gmir.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 # CHECK-LABEL: MachineUniformityInfo for function: hidden_diverge
 # CHECK-LABEL: BLOCK bb.0
 # CHECK:     DIVERGENT: %{{[0-9]*}}: %{{[0-9]*}}:_(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.workitem.id.x)

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/hidden-loop-diverge.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/hidden-loop-diverge.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 # CHECK-LABEL: MachineUniformityInfo for function: hidden_loop_diverge
 
 # CHECK-LABEL: BLOCK bb.0

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/irreducible/branch-outside-gmir.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/irreducible/branch-outside-gmir.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 # CHECK-LABEL: MachineUniformityInfo for function: basic
 # CHECK-NEXT: CYCLES ASSSUMED DIVERGENT:

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/irreducible/diverged-entry-basic-gmir.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/irreducible/diverged-entry-basic-gmir.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 # CHECK-LABEL: MachineUniformityInfo for function: divergent_cycle_1
 # CHECK-NEXT: CYCLES ASSSUMED DIVERGENT:
 # CHECK-NEXT: depth=1: entries(bb.3 bb.1) bb.4 bb.2

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/irreducible/exit-divergence-gmir.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/irreducible/exit-divergence-gmir.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 # CHECK-LABEL: MachineUniformityInfo for function: basic
 # CHECK-NOT: CYCLES ASSSUMED DIVERGENT:
 # CHECK: CYCLES WITH DIVERGENT EXIT:

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/irreducible/irreducible-2-gmir.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/irreducible/irreducible-2-gmir.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 #        bb0(div)
 #       /    \

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/join-loopexit-gmir.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/join-loopexit-gmir.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 # CHECK-LABEL: MachineUniformityInfo for function: test
 
 # CHECK-LABEL: BLOCK bb.0

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/loads-gmir.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/loads-gmir.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 ---
 name:            loads

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/never-uniform.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/never-uniform.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -mcpu=gfx900 -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -mcpu=gfx900 -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 # loads from flat non uniform
 ---
 name:            flatloads

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/temporal-divergence.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/temporal-divergence.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 ---
 # CHECK-LABEL: MachineUniformityInfo for function: temporal_diverge

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/uses-value-from-cycle.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/MIR/uses-value-from-cycle.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -mcpu=gfx1030 -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -mcpu=gfx1030 -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 ---
 name:            f1
 tracksRegLiveness: true

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/deprecated/hidden-diverge.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/deprecated/hidden-diverge.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 # This test was generated using SelectionDAG, where the compilation flow does
 # not match the assumptions made in MachineUA. For now, this test mostly serves

diff --git a/llvm/test/Analysis/UniformityAnalysis/AMDGPU/deprecated/irreducible-1.mir b/llvm/test/Analysis/UniformityAnalysis/AMDGPU/deprecated/irreducible-1.mir
@@ -1,4 +1,5 @@
 # RUN: llc -mtriple=amdgcn-- -run-pass=print-machine-uniformity -o - %s 2>&1 | FileCheck %s
+# RUN: llc -mtriple=amdgcn-- -passes='print<machine-uniformity>' -filetype=null %s 2>&1 | FileCheck %s
 
 # This test was generated using SelectionDAG, where the compilation flow does
 # not match the assumptions made in MachineUA. For now, this test mostly serves