diff --git a/ggml-cuda.cu b/ggml-cuda.cu
index fbf12044d..5714320b8 100644
--- a/ggml-cuda.cu
+++ b/ggml-cuda.cu
@@ -340,7 +340,7 @@ void ggml_init_cublas(void) {
     }
 }
 
-cudaError_t ggml_cuda_cpy_tensor2D(void * dst, const struct ggml_tensor * src, uint64_t i3, uint64_t i2, cudaStream_t stream) {
+cudaError_t ggml_cuda_h2d_tensor_2d(void * dst, const struct ggml_tensor * src, uint64_t i3, uint64_t i2, cudaStream_t stream) {
     const uint64_t ne0 = src->ne[0];
     const uint64_t ne1 = src->ne[1];
     const uint64_t nb0 = src->nb[0];
diff --git a/ggml-cuda.h b/ggml-cuda.h
index 2db34609f..06e2a2886 100644
--- a/ggml-cuda.h
+++ b/ggml-cuda.h
@@ -40,7 +40,7 @@ void dequantize_row_q5_0_cuda(const void * vx, float * y, int k, cudaStream_t st
 void dequantize_row_q5_1_cuda(const void * vx, float * y, int k, cudaStream_t stream);
 void dequantize_row_q8_0_cuda(const void * vx, float * y, int k, cudaStream_t stream);
 
-cudaError_t ggml_cuda_cpy_tensor2D(void * dst, const struct ggml_tensor * src, uint64_t i3, uint64_t i2, cudaStream_t stream);
+cudaError_t ggml_cuda_h2d_tensor_2d(void * dst, const struct ggml_tensor * src, uint64_t i3, uint64_t i2, cudaStream_t stream);
 
 #ifdef  __cplusplus
 }
diff --git a/ggml.c b/ggml.c
index c1bfe5c8b..00932a7fe 100644
--- a/ggml.c
+++ b/ggml.c
@@ -8238,8 +8238,8 @@ static void ggml_compute_forward_mul_mat_f32(
 
 #if defined(GGML_USE_CUBLAS)
                 // copy data to device
-                CUDA_CHECK(ggml_cuda_cpy_tensor2D(d_X, src0, i03, i02, g_cudaStream));
-                CUDA_CHECK(ggml_cuda_cpy_tensor2D(d_Y, src1, i03, i02, g_cudaStream));
+                CUDA_CHECK(ggml_cuda_h2d_tensor_2d(d_X, src0, i03, i02, g_cudaStream));
+                CUDA_CHECK(ggml_cuda_h2d_tensor_2d(d_Y, src1, i03, i02, g_cudaStream));
 
                 // compute
                 CUBLAS_CHECK(
@@ -8461,7 +8461,7 @@ static void ggml_compute_forward_mul_mat_f16_f32(
                 float * d = (float *) ((char *) dst->data + i02*nb2 + i03*nb3);
 
                 // copy data to device
-                CUDA_CHECK(ggml_cuda_cpy_tensor2D(d_X, src0, i03, i02, g_cudaStream));
+                CUDA_CHECK(ggml_cuda_h2d_tensor_2d(d_X, src0, i03, i02, g_cudaStream));
                 CUDA_CHECK(cudaMemcpyAsync(d_Y, y, sizeof(ggml_fp16_t) * y_ne, cudaMemcpyHostToDevice, g_cudaStream));
 
                 // compute
@@ -8713,7 +8713,7 @@ static void ggml_compute_forward_mul_mat_q_f32(
 
 #if defined(GGML_USE_CUBLAS)
                 // copy and dequantize on device
-                CUDA_CHECK(ggml_cuda_cpy_tensor2D(d_Q, src0, i03, i02, g_cudaStream));
+                CUDA_CHECK(ggml_cuda_h2d_tensor_2d(d_Q, src0, i03, i02, g_cudaStream));
 
                 dequantize_row_q_cuda(d_Q, d_X, ne01 * ne00, g_cudaStream);
                 CUDA_CHECK(cudaGetLastError());
@@ -8732,7 +8732,7 @@ static void ggml_compute_forward_mul_mat_q_f32(
 
 #if defined(GGML_USE_CUBLAS)
                 // copy data to device
-                CUDA_CHECK(ggml_cuda_cpy_tensor2D(d_Y, src1, i03, i02, g_cudaStream));
+                CUDA_CHECK(ggml_cuda_h2d_tensor_2d(d_Y, src1, i03, i02, g_cudaStream));
 
                 // compute
                 CUBLAS_CHECK(