#include <vector>
using namespace llvm;
+#define DEBUG_TYPE "divergence"
+
namespace {
class DivergencePropagator {
PDT, DivergentValues);
DP.populateWithSourcesOfDivergence();
DP.propagate();
+ LLVM_DEBUG(
+ dbgs() << "\nAfter divergence analysis on " << F.getName() << ":\n";
+ print(dbgs(), F.getParent())
+ );
return false;
}
// Dumps all divergent values in F, arguments and then instructions.
for (auto &Arg : F->args()) {
- if (DivergentValues.count(&Arg))
- OS << "DIVERGENT: " << Arg << "\n";
+ OS << (DivergentValues.count(&Arg) ? "DIVERGENT: " : " ");
+ OS << Arg << "\n";
}
// Iterate instructions using instructions() to ensure a deterministic order.
- for (auto &I : instructions(F)) {
- if (DivergentValues.count(&I))
- OS << "DIVERGENT:" << I << "\n";
+ for (auto BI = F->begin(), BE = F->end(); BI != BE; ++BI) {
+ auto &BB = *BI;
+ OS << "\n " << BB.getName() << ":\n";
+ for (auto &I : BB.instructionsWithoutDebug()) {
+ OS << (DivergentValues.count(&I) ? "DIVERGENT: " : " ");
+ OS << I << "\n";
+ }
}
+ OS << "\n";
}
; CHECK-LABEL: 'test1':
; CHECK-NEXT: DIVERGENT: i32 %bound
+; CHECK: {{^ *}}%counter =
; CHECK-NEXT: DIVERGENT: %break = icmp sge i32 %counter, %bound
; CHECK-NEXT: DIVERGENT: br i1 %break, label %footer, label %body
-; CHECK-NEXT: DIVERGENT: br i1 %break, label %end, label %header
+; CHECK: {{^ *}}%counter.next =
+; CHECK: {{^ *}}%counter.footer =
+; CHECK: DIVERGENT: br i1 %break, label %end, label %header
; Note: %counter is not divergent!
define amdgpu_ps void @test1(i32 %bound) {
entry: