inserted a check for argument types in the interface

kernel_tuner/interface.py

@@ -275,6 +275,7 @@ def tune_kernel(kernel_name, kernel_string, problem_size, arguments,
     max_threads = dev.max_threads
 
     #move data to GPU
+    _check_argument_list(arguments)
     gpu_args = dev.ready_argument_list(arguments)
 
     #compute cartesian product of all tunable parameters
@@ -519,3 +520,7 @@ def _check_kernel_correctness(dev, func, gpu_args, threads, grid, answer, instan
         raise Exception("Error " + instance_string + " failed correctness check")
     return correct
 
+def _check_argument_list(args):
+    for (i, arg) in enumerate(args):
+        if not isinstance(arg, (numpy.ndarray, numpy.generic)):
+            raise TypeError("Argument at position " + str(i) + " of type: " + str(type(arg)) + " should be of type numpy.ndarray or numpy scalar")

test/test_interface.py

@@ -117,3 +117,22 @@ def test_check_kernel_correctness(dev_interface):
             assert args[1] == 'gpu_args'
     assert dev.memcpy_dtoh.called == 1
     assert test
+
+def test_check_argument_list1():
+    args = [numpy.int32(5), 'blah', numpy.array([1, 2, 3])]
+    try:
+        kernel_tuner._check_argument_list(args)
+        print("Expected a TypeError to be raised")
+        assert False
+    except TypeError as e:
+        print(str(e))
+        assert "at position 1" in str(e)
+    except:
+        print("Expected a TypeError to be raised")
+        assert False
+
+def test_check_argument_list2():
+    args = [numpy.int32(5), numpy.float64(4.6), numpy.array([1, 2, 3])]
+    kernel_tuner._check_argument_list(args)
+    #test that no exception is raised
+    assert True