diff --git a/src/graph/backend/dnnl/kernels/batch_norm.cpp b/src/graph/backend/dnnl/kernels/batch_norm.cpp
index f94de6e99d9..eb9aadc54ef 100644
--- a/src/graph/backend/dnnl/kernels/batch_norm.cpp
+++ b/src/graph/backend/dnnl/kernels/batch_norm.cpp
@@ -135,9 +135,10 @@ status_t batch_norm_fwd_t::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -204,9 +205,10 @@ status_t batch_norm_fwd_t::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -279,9 +281,10 @@ status_t batch_norm_fwd_t::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/conv_base.cpp b/src/graph/backend/dnnl/kernels/conv_base.cpp
index c082b50cdfc..256e1426471 100644
--- a/src/graph/backend/dnnl/kernels/conv_base.cpp
+++ b/src/graph/backend/dnnl/kernels/conv_base.cpp
@@ -63,9 +63,10 @@ status_t conv_base_t::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -132,9 +133,10 @@ status_t conv_base_t::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -207,9 +209,10 @@ status_t conv_base_t::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/eltwise.cpp b/src/graph/backend/dnnl/kernels/eltwise.cpp
index 0d46b49a255..e8abe4a368c 100644
--- a/src/graph/backend/dnnl/kernels/eltwise.cpp
+++ b/src/graph/backend/dnnl/kernels/eltwise.cpp
@@ -138,9 +138,10 @@ status_t eltwise_fwd_t<quantized>::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -208,9 +209,10 @@ status_t eltwise_fwd_t<quantized>::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -284,9 +286,10 @@ status_t eltwise_fwd_t<quantized>::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/group_norm.cpp b/src/graph/backend/dnnl/kernels/group_norm.cpp
index 116a6c36d0d..2b11861c1e4 100644
--- a/src/graph/backend/dnnl/kernels/group_norm.cpp
+++ b/src/graph/backend/dnnl/kernels/group_norm.cpp
@@ -143,9 +143,10 @@ status_t group_norm_fwd_t::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -212,9 +213,10 @@ status_t group_norm_fwd_t::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -287,9 +289,10 @@ status_t group_norm_fwd_t::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/kernel_base.cpp b/src/graph/backend/dnnl/kernels/kernel_base.cpp
index 8a47a558925..798ddaebaa3 100644
--- a/src/graph/backend/dnnl/kernels/kernel_base.cpp
+++ b/src/graph/backend/dnnl/kernels/kernel_base.cpp
@@ -43,6 +43,24 @@ bool kernel_base_t::enabled_constant_cache() const {
     return enabled;
 }
 
+size_t kernel_base_t::encode_constant_cache_key(
+        const std::vector<tensor_t> &inputs, size_t cache_key) const {
+    // Encode the constant memory address into cache key for differentiation
+    size_t encoded_cache_key = cache_key;
+    for (const auto &in : inputs) {
+        if (in.get_logical_tensor().property
+                == dnnl_graph_tensor_property_t::
+                        dnnl_graph_tensor_property_constant) {
+            auto data_handle = in.get_data_handle();
+            if (data_handle != nullptr) {
+                encoded_cache_key = hash_combine(encoded_cache_key,
+                        reinterpret_cast<uintptr_t>(data_handle));
+            }
+        }
+    }
+    return encoded_cache_key;
+}
+
 const std::vector<inplace_pair_t> &kernel_base_t::get_inplace_pairs() const {
     return inplace_pairs_;
 };
diff --git a/src/graph/backend/dnnl/kernels/kernel_base.hpp b/src/graph/backend/dnnl/kernels/kernel_base.hpp
index 4c3a3849a61..1146ff7f445 100644
--- a/src/graph/backend/dnnl/kernels/kernel_base.hpp
+++ b/src/graph/backend/dnnl/kernels/kernel_base.hpp
@@ -103,6 +103,9 @@ struct kernel_base_t {
 
     bool enabled_constant_cache() const;
 
+    size_t encode_constant_cache_key(
+            const std::vector<tensor_t> &inputs, size_t cache_key) const;
+
     const std::vector<inplace_pair_t> &get_inplace_pairs() const;
 
 protected:
diff --git a/src/graph/backend/dnnl/kernels/large_partition.cpp b/src/graph/backend/dnnl/kernels/large_partition.cpp
index 16b821d4c88..a8d8393ac42 100644
--- a/src/graph/backend/dnnl/kernels/large_partition.cpp
+++ b/src/graph/backend/dnnl/kernels/large_partition.cpp
@@ -248,9 +248,10 @@ status_t larger_partition_kernel_t::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -316,9 +317,10 @@ status_t larger_partition_kernel_t::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -389,9 +391,10 @@ status_t larger_partition_kernel_t::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/layer_norm.cpp b/src/graph/backend/dnnl/kernels/layer_norm.cpp
index 9e3c028c02a..f7a4d084987 100644
--- a/src/graph/backend/dnnl/kernels/layer_norm.cpp
+++ b/src/graph/backend/dnnl/kernels/layer_norm.cpp
@@ -140,9 +140,10 @@ status_t layer_norm_fwd_t::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -209,9 +210,10 @@ status_t layer_norm_fwd_t::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -284,9 +286,10 @@ status_t layer_norm_fwd_t::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/matmul.cpp b/src/graph/backend/dnnl/kernels/matmul.cpp
index ccac714dfc6..80146cdbc96 100644
--- a/src/graph/backend/dnnl/kernels/matmul.cpp
+++ b/src/graph/backend/dnnl/kernels/matmul.cpp
@@ -193,9 +193,10 @@ status_t matmul_t<quantized>::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -263,9 +264,10 @@ status_t matmul_t<quantized>::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -339,9 +341,10 @@ status_t matmul_t<quantized>::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/pool.cpp b/src/graph/backend/dnnl/kernels/pool.cpp
index aa881571cf1..9703ffe0495 100644
--- a/src/graph/backend/dnnl/kernels/pool.cpp
+++ b/src/graph/backend/dnnl/kernels/pool.cpp
@@ -170,9 +170,10 @@ status_t pooling_fwd_t<quantized>::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -240,9 +241,10 @@ status_t pooling_fwd_t<quantized>::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -316,9 +318,10 @@ status_t pooling_fwd_t<quantized>::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/quantize.cpp b/src/graph/backend/dnnl/kernels/quantize.cpp
index c76c3592d4b..31388ef5d10 100644
--- a/src/graph/backend/dnnl/kernels/quantize.cpp
+++ b/src/graph/backend/dnnl/kernels/quantize.cpp
@@ -134,9 +134,10 @@ status_t quantize_dequantize_t::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -208,9 +209,10 @@ status_t quantize_dequantize_t::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -283,9 +285,10 @@ status_t quantize_dequantize_t::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/reorder.cpp b/src/graph/backend/dnnl/kernels/reorder.cpp
index 9144ddd2556..b4f38f8376c 100644
--- a/src/graph/backend/dnnl/kernels/reorder.cpp
+++ b/src/graph/backend/dnnl/kernels/reorder.cpp
@@ -153,9 +153,10 @@ status_t reorder_t<quantized>::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -223,9 +224,10 @@ status_t reorder_t<quantized>::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -299,9 +301,10 @@ status_t reorder_t<quantized>::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/select.cpp b/src/graph/backend/dnnl/kernels/select.cpp
index bc6fa38d1f0..3cbb75f2aff 100644
--- a/src/graph/backend/dnnl/kernels/select.cpp
+++ b/src/graph/backend/dnnl/kernels/select.cpp
@@ -134,9 +134,10 @@ status_t select_t::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -203,9 +204,10 @@ status_t select_t::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -278,9 +280,10 @@ status_t select_t::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
diff --git a/src/graph/backend/dnnl/kernels/softmax.cpp b/src/graph/backend/dnnl/kernels/softmax.cpp
index d8d871b5e99..1474f6abac9 100644
--- a/src/graph/backend/dnnl/kernels/softmax.cpp
+++ b/src/graph/backend/dnnl/kernels/softmax.cpp
@@ -137,9 +137,10 @@ status_t softmax_fwd_t::execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -206,9 +207,10 @@ status_t softmax_fwd_t::sycl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();
@@ -281,9 +283,10 @@ status_t softmax_fwd_t::ocl_execute_impl(const stream_t *g_stream,
 
     constant_cache_t::cached_t c_buffer;
     if (enabled_constant_cache()) {
+        size_t encoded_key = encode_constant_cache_key(inputs, constant_key_);
         std::promise<constant_cache_t::cached_t> c_promise;
         constant_cache_t::value_t cached_value
-                = dnnl_constant_cache_get_or_add(p_engine_, constant_key_,
+                = dnnl_constant_cache_get_or_add(p_engine_, encoded_key,
                         memory_planner_.total_internal_persistent_size(),
                         c_promise.get_future());
         bool is_from_cache = cached_value.valid();