unittest: don't pin CleanNaNInfs to one implementation's output

The contract is only that a bad lane becomes something that yields zero
when multiplied by zero, by whatever route is cheapest. SSE2 clamps to
+-FLT_MAX, NEON, LSX and the scalar fallback zero the lane - both fine.
Check that property, and that good lanes are untouched.

Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>

unittest/TestCrossSIMD.cpp

@@ -391,9 +391,15 @@ static bool TestVec4F32NaNInf() {
 	EXPECT_TRUE(std::isinf(result[2]));
 	EXPECT_TRUE(std::isinf(result[3]));
 
+	// The contract is just "a value that yields zero when multiplied by zero" - whatever is
+	// cheapest to get there. So the implementations legitimately differ on what a bad lane
+	// becomes: SSE2 clamps to +-FLT_MAX, NEON, LSX and the scalar fallback zero it. Check the
+	// property rather than the value, and that good lanes are left alone.
 	v.CleanNaNInfs().Store(result);
-	static const float known_clean[4] = { 1.0f, 0.0f, 0.0f, 0.0f };
-	if (!CompareFloats(result, known_clean, 4, __LINE__)) return false;
+	EXPECT_EQ_FLOAT(result[0], 1.0f);
+	for (int i = 0; i < 4; i++) {
+		EXPECT_EQ_FLOAT(result[i] * 0.0f, 0.0f);
+	}
 
 	return true;
 }