[Bugfix] Avoid uninitialized usage of azp_val when AZP is false. (vllm-project#24335)

mohankku · xuebwang-amd · commit 76f3b833cd6a · 2025-10-24T09:19:15.000Z
Signed-off-by: Mohan Kumar Kumar &lt;mohan.cbein@gmail.com&gt;
Signed-off-by: mohankku &lt;mohan.cbein@gmail.com&gt;
Signed-off-by: xuebwang-amd &lt;xuebwang@amd.com&gt;
diff --git a/csrc/cpu/dnnl_kernels.cpp b/csrc/cpu/dnnl_kernels.cpp
@@ -145,7 +145,8 @@ void dynamic_scaled_int8_quant_impl(const scalar_t* input, int8_t* output,
       }
     }
 
-    float scale_val, azp_val;
+    float scale_val;
+    float azp_val = 0.0f;
     if constexpr (AZP) {
       float max_scalar = max_value.reduce_max();
       float min_scalar = min_value.reduce_min();

Original file line number	Diff line number	Diff line change
`@@ -145,7 +145,8 @@ void dynamic_scaled_int8_quant_impl(const scalar_t* input, int8_t* output,`
`145`	`145`	`}`
`146`	`146`	`}`
`147`	`147`
`148`		`- float scale_val, azp_val;`
	`148`	`+ float scale_val;`
	`149`	`+ float azp_val = 0.0f;`
`149`	`150`	`if constexpr (AZP) {`
`150`	`151`	`float max_scalar = max_value.reduce_max();`
`151`	`152`	`float min_scalar = min_value.reduce_min();`