[llvm-branch-commits] [llvm] [AArch64] Tweak the cost-model of partial reductions to mitigate regressions from #181706 (PR #181707)

[Florian Hahn via llvm-branch-commits]

================
@@ -5964,7 +5969,7 @@ InstructionCost AArch64TTIImpl::getPartialReductionCost(
       return Cost;
     // Floating-point types aren't promoted, so expanding the partial reduction
     // is more expensive.
-    return Cost + 20;
+    return Cost * 8;
----------------
fhahn wrote:

why scale by 8?

https://github.com/llvm/llvm-project/pull/181707