[llvm] [GlobalMerge][NFC] Skip sorting by profitability when it is not needed (PR #124146)

Michael Maitland via llvm-commits llvm-commits at lists.llvm.org
Thu Jan 23 08:57:03 PST 2025


https://github.com/michaelmaitland created https://github.com/llvm/llvm-project/pull/124146

We were previously sorting by profitability even if we were choosing to merge all globals together, which is not impacted by UsedGlobalSet order.

We can also remove iteration of UsedGlobalSets in reverse order in both cases. In the first csae, the order does not matter. In the second case, we just sort by the order we need instead of sorting in the opposite direction and calling reverse.

This change should only be an improvement on compile time. I have not measured it, but I think it would never make things worse.

>From 3676606ec61b8f7f7c45d75d1747c9dd47b4610c Mon Sep 17 00:00:00 2001
From: Michael Maitland <michaeltmaitland at gmail.com>
Date: Thu, 23 Jan 2025 08:51:14 -0800
Subject: [PATCH] [GlobalMerge][NFC] Skip sorting by profitability when it is
 not needed

We were previously sorting by profitability even if we were choosing to merge
all globals together, which is not impacted by UsedGlobalSet order.

We can also remove iteration of UsedGlobalSets in reverse order in both cases.
In the first csae, the order does not matter. In the second case, we just sort
by the order we need instead of sorting in the opposite direction and calling
reverse.

This change should only be an improvement on compile time. I have not measured
it, but I think it would never make things worse.
---
 llvm/lib/CodeGen/GlobalMerge.cpp | 26 ++++++++++++--------------
 1 file changed, 12 insertions(+), 14 deletions(-)

diff --git a/llvm/lib/CodeGen/GlobalMerge.cpp b/llvm/lib/CodeGen/GlobalMerge.cpp
index 48d4d7848d84a7..65b6d6dc30d5c2 100644
--- a/llvm/lib/CodeGen/GlobalMerge.cpp
+++ b/llvm/lib/CodeGen/GlobalMerge.cpp
@@ -422,24 +422,12 @@ bool GlobalMergeImpl::doMerge(SmallVectorImpl<GlobalVariable *> &Globals,
     }
   }
 
-  // Now we found a bunch of sets of globals used together.  We accumulated
-  // the number of times we encountered the sets (i.e., the number of blocks
-  // that use that exact set of globals).
-  //
-  // Multiply that by the size of the set to give us a crude profitability
-  // metric.
-  llvm::stable_sort(UsedGlobalSets,
-                    [](const UsedGlobalSet &UGS1, const UsedGlobalSet &UGS2) {
-                      return UGS1.Globals.count() * UGS1.UsageCount <
-                             UGS2.Globals.count() * UGS2.UsageCount;
-                    });
-
   // We can choose to merge all globals together, but ignore globals never used
   // with another global.  This catches the obviously non-profitable cases of
   // having a single global, but is aggressive enough for any other case.
   if (GlobalMergeIgnoreSingleUse) {
     BitVector AllGlobals(Globals.size());
-    for (const UsedGlobalSet &UGS : llvm::reverse(UsedGlobalSets)) {
+    for (const UsedGlobalSet &UGS : UsedGlobalSets) {
       if (UGS.UsageCount == 0)
         continue;
       if (UGS.Globals.count() > 1)
@@ -448,6 +436,16 @@ bool GlobalMergeImpl::doMerge(SmallVectorImpl<GlobalVariable *> &Globals,
     return doMerge(Globals, AllGlobals, M, isConst, AddrSpace);
   }
 
+  // Now we found a bunch of sets of globals used together. We accumulated
+  // the number of times we encountered the sets (i.e., the number of functions 
+  // that use that exact set of globals). Multiply that by the size of the set
+  // to give us a crude profitability metric.
+  llvm::stable_sort(UsedGlobalSets,
+                    [](const UsedGlobalSet &UGS1, const UsedGlobalSet &UGS2) {
+                      return UGS1.Globals.count() * UGS1.UsageCount >=
+                             UGS2.Globals.count() * UGS2.UsageCount;
+                    });
+
   // Starting from the sets with the best (=biggest) profitability, find a
   // good combination.
   // The ideal (and expensive) solution can only be found by trying all
@@ -457,7 +455,7 @@ bool GlobalMergeImpl::doMerge(SmallVectorImpl<GlobalVariable *> &Globals,
   BitVector PickedGlobals(Globals.size());
   bool Changed = false;
 
-  for (const UsedGlobalSet &UGS : llvm::reverse(UsedGlobalSets)) {
+  for (const UsedGlobalSet &UGS : UsedGlobalSets) {
     if (UGS.UsageCount == 0)
       continue;
     if (PickedGlobals.anyCommon(UGS.Globals))



More information about the llvm-commits mailing list