diff --git a/src/infiniop/ops/paged_caching/bang/paged_caching_bang.mlu b/src/infiniop/ops/paged_caching/bang/paged_caching_bang.mlu
index 0f7597fa5..bbe57a216 100644
--- a/src/infiniop/ops/paged_caching/bang/paged_caching_bang.mlu
+++ b/src/infiniop/ops/paged_caching/bang/paged_caching_bang.mlu
@@ -5,6 +5,8 @@
 
 namespace {
 
+__nram__ char paged_caching_nram_buffer[NRAM_MAX_SIZE];
+
 template <typename Tdata>
 __mlu_global__ void pagedCachingKernel(
     Tdata *k_cache,
@@ -48,8 +50,12 @@ __mlu_global__ void pagedCachingKernel(
                      + head_idx * v_cache_head_stride
                      + block_offset * v_cache_slot_stride;
 
-        __memcpy(k_dst, k_src, head_size * sizeof(Tdata), GDRAM2GDRAM);
-        __memcpy(v_dst, v_src, head_size * sizeof(Tdata), GDRAM2GDRAM);
+        char *nram_base = reinterpret_cast<char *>(((reinterpret_cast<size_t>(paged_caching_nram_buffer) + ALIGN_SIZE - 1) / ALIGN_SIZE) * ALIGN_SIZE);
+        Tdata *tmp = reinterpret_cast<Tdata *>(nram_base);
+        __memcpy(tmp, k_src, head_size * sizeof(Tdata), GDRAM2NRAM);
+        __memcpy(tmp + head_size, v_src, head_size * sizeof(Tdata), GDRAM2NRAM);
+        __memcpy(k_dst, tmp, head_size * sizeof(Tdata), NRAM2GDRAM);
+        __memcpy(v_dst, tmp + head_size, head_size * sizeof(Tdata), NRAM2GDRAM);
     }
 }