fix sgl-kernel build failure on AMD (#3352)

sgl-kernel/src/sgl-kernel/include/utils.h

@@ -16,7 +16,9 @@ limitations under the License.
 #pragma once
 
 #include <cuda_runtime.h>
+#ifndef USE_ROCM
 #include <pytorch_extension_utils.h>
+#endif
 #include <torch/extension.h>
 
 #include <sstream>
@@ -63,6 +65,7 @@ inline int getSMVersion() {
   return sm_major * 10 + sm_minor;
 }
 
+#ifndef USE_ROCM
 #define DISPATCH_PYTORCH_DTYPE_TO_CTYPE_FLOAT_FP16(pytorch_dtype, c_type, ...)           \
   [&]() -> bool {                                                                        \
     switch (pytorch_dtype) {                                                             \
@@ -79,6 +82,7 @@ inline int getSMVersion() {
         return false;                                                                    \
     }                                                                                    \
   }()
+#endif
 
 #define DISPATCH_CASE_INTEGRAL_TYPES(...)              \
   AT_DISPATCH_CASE(at::ScalarType::Byte, __VA_ARGS__)  \