[llvm-exegesis] getNonRedundantWriteProcRes - perform basic toplogical sorting (PR58500)

getNonRedundantWriteProcRes was assuming that tblgen topologically sorted the cpu ModelProcResources[] arrays so that resource units were declared before the resource groups that used them, but unfortunately that doesn't appear to be true - in most cases it was just getting lucky based off the alphanumeric sorting that was being performed and the choice of the resource pipe names in most scheduler models (Intel models in particular).

This patch adds an explicit sort, based off llvm-mca's initializeUsedResources, that sorts by resource mask - I'm hoping this basic sorting is enough, I don't think overlapping groups or Super resources are a problem.

I'd like to take this further in the future and start sharing more code between llvm-mca and llvm-exegesis - while triaging this bug I saw how similar both approaches are, but are just dissimilar enough that any refactor isn't going to be trivial :(

Working with @courbet on a follow up unit test

Fixes #58500
Fixes #37045

Differential Revision: https://reviews.llvm.org/D136351

diff --git a/llvm/tools/llvm-exegesis/lib/CMakeLists.txt b/llvm/tools/llvm-exegesis/lib/CMakeLists.txt
index 922162f..c92b6ab 100644 (file)
--- a/llvm/tools/llvm-exegesis/lib/CMakeLists.txt
+++ b/llvm/tools/llvm-exegesis/lib/CMakeLists.txt
@@ -26,6 +26,7 @@ set(LLVM_LINK_COMPONENTS
   ExecutionEngine
   GlobalISel
   MC
+  MCA
   MCDisassembler
   MCJIT
   MCParser

diff --git a/llvm/tools/llvm-exegesis/lib/SchedClassResolution.cpp b/llvm/tools/llvm-exegesis/lib/SchedClassResolution.cpp
index 03386cf..024e511 100644 (file)
--- a/llvm/tools/llvm-exegesis/lib/SchedClassResolution.cpp
+++ b/llvm/tools/llvm-exegesis/lib/SchedClassResolution.cpp
@@ -10,6 +10,7 @@
 #include "BenchmarkResult.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/MC/MCAsmInfo.h"
+#include "llvm/MCA/Support.h"
 #include "llvm/Support/FormatVariadic.h"
 #include <limits>
 #include <unordered_set>
@@ -45,7 +46,7 @@ namespace exegesis {
 //
 // Note that in this case, P016 does not contribute any cycles, so it would
 // be removed by this function.
-// FIXME: Move this to MCSubtargetInfo and use it in llvm-mca.
+// FIXME: Merge this with the equivalent in llvm-mca.
 static SmallVector<MCWriteProcResEntry, 8>
 getNonRedundantWriteProcRes(const MCSchedClassDesc &SCDesc,
                             const MCSubtargetInfo &STI) {
@@ -53,12 +54,33 @@ getNonRedundantWriteProcRes(const MCSchedClassDesc &SCDesc,
   const auto &SM = STI.getSchedModel();
   const unsigned NumProcRes = SM.getNumProcResourceKinds();
 
-  // This assumes that the ProcResDescs are sorted in topological order, which
-  // is guaranteed by the tablegen backend.
-  SmallVector<float, 32> ProcResUnitUsage(NumProcRes);
+  // Collect resource masks.
+  SmallVector<uint64_t> ProcResourceMasks(NumProcRes);
+  mca::computeProcResourceMasks(SM, ProcResourceMasks);
+
+  // Sort entries by smaller resources for (basic) topological ordering.
+  using ResourceMaskAndEntry = std::pair<uint64_t, const MCWriteProcResEntry *>;
+  SmallVector<ResourceMaskAndEntry, 8> ResourceMaskAndEntries;
   for (const auto *WPR = STI.getWriteProcResBegin(&SCDesc),
                   *const WPREnd = STI.getWriteProcResEnd(&SCDesc);
        WPR != WPREnd; ++WPR) {
+    uint64_t Mask = ProcResourceMasks[WPR->ProcResourceIdx];
+    ResourceMaskAndEntries.push_back({Mask, WPR});
+  }
+  sort(ResourceMaskAndEntries,
+       [](const ResourceMaskAndEntry &A, const ResourceMaskAndEntry &B) {
+         unsigned popcntA = countPopulation(A.first);
+         unsigned popcntB = countPopulation(B.first);
+         if (popcntA < popcntB)
+           return true;
+         if (popcntA > popcntB)
+           return false;
+         return A.first < B.first;
+       });
+
+  SmallVector<float, 32> ProcResUnitUsage(NumProcRes);
+  for (const ResourceMaskAndEntry &Entry : ResourceMaskAndEntries) {
+    const MCWriteProcResEntry *WPR = Entry.second;
     const MCProcResourceDesc *const ProcResDesc =
         SM.getProcResource(WPR->ProcResourceIdx);
     if (ProcResDesc->SubUnitsIdxBegin == nullptr) {