diff --git a/src/cmd/compile/internal/ssa/schedule.go b/src/cmd/compile/internal/ssa/schedule.go
index 4cd60d7..c291e5c 100644
--- a/src/cmd/compile/internal/ssa/schedule.go
+++ b/src/cmd/compile/internal/ssa/schedule.go
@@ -25,8 +25,9 @@
 )
 
 type ValHeap struct {
-	a     []*Value
-	score []int8
+	a           []*Value
+	score       []int8
+	inBlockUses []bool
 }
 
 func (h ValHeap) Len() int      { return len(h.a) }
@@ -56,6 +57,12 @@
 	// Note: only scores are required for correct scheduling.
 	// Everything else is just heuristics.
 
+	ix := h.inBlockUses[x.ID]
+	iy := h.inBlockUses[y.ID]
+	if ix != iy {
+		return ix // values with in-block uses come earlier
+	}
+
 	if x.Pos != y.Pos { // Favor in-order line stepping
 		if x.Block == x.Block.Func.Entry && x.Pos.IsStmt() != y.Pos.IsStmt() {
 			// In the entry block, put statement-marked instructions earlier.
@@ -110,6 +117,23 @@
 	nextMem := f.Cache.allocValueSlice(f.NumValues())
 	defer f.Cache.freeValueSlice(nextMem)
 
+	// inBlockUses records whether a value is used in the block
+	// in which it lives. (block control values don't count as uses.)
+	inBlockUses := f.Cache.allocBoolSlice(f.NumValues())
+	defer f.Cache.freeBoolSlice(inBlockUses)
+	if f.Config.optimize {
+		for _, b := range f.Blocks {
+			for _, v := range b.Values {
+				for _, a := range v.Args {
+					if a.Block == b {
+						inBlockUses[a.ID] = true
+					}
+				}
+			}
+		}
+	}
+	priq.inBlockUses = inBlockUses
+
 	for _, b := range f.Blocks {
 		// Compute score. Larger numbers are scheduled closer to the end of the block.
 		for _, v := range b.Values {
diff --git a/test/codegen/issue58166.go b/test/codegen/issue58166.go
new file mode 100644
index 0000000..8be5aac
--- /dev/null
+++ b/test/codegen/issue58166.go
@@ -0,0 +1,23 @@
+// asmcheck
+
+// Copyright 2023 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package p
+
+func dgemmSerialNotNot(m, n, k int, a []float64, lda int, b []float64, ldb int, c []float64, ldc int, alpha float64) {
+	for i := 0; i < m; i++ {
+		ctmp := c[i*ldc : i*ldc+n]
+		for l, v := range a[i*lda : i*lda+k] {
+			tmp := alpha * v
+			if tmp != 0 {
+				x := b[l*ldb : l*ldb+n]
+				// amd64:"INCQ"
+				for i, v := range x {
+					ctmp[i] += tmp * v
+				}
+			}
+		}
+	}
+}