diff --git a/CMakeLists.txt b/CMakeLists.txt
index a827eda96..4c585c2d7 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -96,8 +96,8 @@ option(LLAMA_LLAMAFILE                       "llama: use llamafile SGEMM"
 set(LLAMA_BLAS_VENDOR "Generic" CACHE STRING "llama: BLAS library vendor")
 option(LLAMA_CUDA                            "llama: use CUDA"                                  OFF)
 option(LLAMA_CUBLAS                          "llama: use CUDA (deprecated, use LLAMA_CUDA)"     OFF)
-option(LLAMA_FORCE_DMMV                      "llama: use dmmv instead of mmvq CUDA kernels"     OFF)
-option(LLAMA_FORCE_MMQ                       "llama: use mmq kernels instead of cuBLAS"         OFF)
+option(LLAMA_FORCE_DMMV                      "llama: use dmmv instead of mmvq kernels on GPU"   OFF)
+option(LLAMA_FORCE_MMQ                       "llama: use mmq kernels instead of Math Lib"       OFF)
 set(LLAMA_CUDA_DMMV_X      "32" CACHE STRING "llama: x stride for dmmv CUDA kernels")
 set(LLAMA_CUDA_MMV_Y        "1" CACHE STRING "llama: y block size for mmv CUDA kernels")
 option(LLAMA_CUDA_F16                        "llama: use 16 bit floats for some calculations"   OFF)
@@ -628,10 +628,10 @@ if (LLAMA_SYCL)
         add_compile_definitions(GGML_SYCL_F16)
     endif()
 
-    if (LLAMA_SYCL_FORCE_DMMV)
+    if (LLAMA_FORCE_DMMV)
         add_compile_definitions(GGML_SYCL_FORCE_DMMV)
     endif()
-    if (LLAMA_SYCL_FORCE_MMQ)
+    if (LLAMA_FORCE_MMQ)
         add_compile_definitions(GGML_SYCL_FORCE_MMQ)
     endif()
 
diff --git a/ggml-sycl.cpp b/ggml-sycl.cpp
index 623202d9a..47dbfcde3 100644
--- a/ggml-sycl.cpp
+++ b/ggml-sycl.cpp
@@ -2978,10 +2978,9 @@ static int g_work_group_size = 0;
 
 #define GGML_SYCL_MAX_NODES 8192 //TODO: adapt to hardwares
 
-
-//define for XMX in Intel GPU
-//TODO: currently, it's not used for XMX really.
-#define SYCL_USE_XMX
+#if !defined(GGML_SYCL_FORCE_MMQ)
+    #define SYCL_USE_XMX
+#endif
 
 // max batch size to use MMQ kernels when tensor cores are available
 #define MMQ_MAX_BATCH_SIZE 32
@@ -15228,10 +15227,6 @@ static void ggml_sycl_mul_mat(const ggml_tensor * src0, const ggml_tensor * src1
         }
     }
 
-#if !defined(GGML_SYCL_FORCE_MMQ)
-    #define SYCL_USE_XMX
-#endif
-
 #ifdef SYCL_USE_XMX
     bool use_xmx = true;
 #else