[llvm] [AMDGPU][True16] reduce compile-time heuristic for coalescer pass (PR #205256)

[Guo Chen via llvm-commits]

https://github.com/broxigarchen updated https://github.com/llvm/llvm-project/pull/205256

>From 2bb6e7a797ec720c0f5e2c2d1765082eb494d6bf Mon Sep 17 00:00:00 2001
From: guochen2 <guochen2 at amd.com>
Date: Tue, 23 Jun 2026 00:59:14 -0400
Subject: [PATCH] reduce heristic

---
 llvm/lib/CodeGen/RegisterCoalescer.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/llvm/lib/CodeGen/RegisterCoalescer.cpp b/llvm/lib/CodeGen/RegisterCoalescer.cpp
index eb43eddfa081e..4cb291008c37e 100644
--- a/llvm/lib/CodeGen/RegisterCoalescer.cpp
+++ b/llvm/lib/CodeGen/RegisterCoalescer.cpp
@@ -119,7 +119,7 @@ static cl::opt<unsigned> LargeIntervalFreqThreshold(
     cl::desc("For a large interval, if it is coalesced with other live "
              "intervals many times more than the threshold, stop its "
              "coalescing to control the compile time. "),
-    cl::init(256));
+    cl::init(2));
 
 namespace {