[llvm] 34c6992 - [Arm] Control forced unrolling of small loops (#170127)

[via llvm-commits]

Author: Vladi Krapp
Date: 2025-12-02T08:39:26Z
New Revision: 34c699246d9d2ad0e09306d4faed6e8d7ec87aa5

URL: https://github.com/llvm/llvm-project/commit/34c699246d9d2ad0e09306d4faed6e8d7ec87aa5
DIFF: https://github.com/llvm/llvm-project/commit/34c699246d9d2ad0e09306d4faed6e8d7ec87aa5.diff

LOG: [Arm] Control forced unrolling of small loops (#170127)

* Add flag to control cost threshold for forced unrolling of loops.
  Existing value preserved as default.

Added: 
    

Modified: 
    llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp b/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
index fdb0ec40cb41f..cecc9544ed835 100644
--- a/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
+++ b/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
@@ -66,6 +66,11 @@ extern cl::opt<bool> EnableMaskedGatherScatters;
 
 extern cl::opt<unsigned> MVEMaxSupportedInterleaveFactor;
 
+static cl::opt<int> ArmForceUnrollThreshold(
+    "arm-force-unroll-threshold", cl::init(12), cl::Hidden,
+    cl::desc(
+        "Threshold for forced unrolling of small loops in Arm architecture"));
+
 /// Convert a vector load intrinsic into a simple llvm load instruction.
 /// This is beneficial when the underlying object being addressed comes
 /// from a constant, since we get constant-folding for free.
@@ -2731,7 +2736,7 @@ void ARMTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE,
 
   // Force unrolling small loops can be very useful because of the branch
   // taken cost of the backedge.
-  if (Cost < 12)
+  if (Cost < ArmForceUnrollThreshold)
     UP.Force = true;
 }