Merge pull request #8 from alendit/jitclass_benchmark

Add jitclass method benchmark

benchmarks/bench_jitclass.py

@@ -0,0 +1,85 @@
+"""
+Benchmarks for jitclass method dispatching and call overhead
+"""
+
+import numpy as numpy
+
+
+def setup():
+    from numba import njit, jitclass, uint32
+
+    @jitclass([('val', uint32)])
+    class Box:
+        def __init__(self, val):
+            self.val = val
+
+        def inc(self):
+            self.val += 1
+
+        @property
+        def value(self):
+            return self.val
+
+        @value.setter
+        def value(self, new_value):
+            self.val = new_value
+
+    @njit
+    def method_call(N):
+        b = Box(0)
+        for i in range(N):
+            b.inc()
+        return b
+
+    @njit
+    def property_access(N):
+        acc = 0
+        b = Box(10)
+        for i in range(N):
+            acc += b.value
+
+        return acc
+
+    @njit
+    def property_setting(N):
+        b = Box(0)
+        for i in range(N):
+            b.value = i
+
+        return b
+
+    @njit
+    def constructor_call(N):
+        l = []
+        for i in range(N):
+            l.append(Box(i))
+
+        return l
+
+    globals().update(locals())
+
+
+class JitClassDispatch:
+
+    N = 10 ** 8
+    funcs = [
+            'constructor_call',
+            'method_call',
+            'property_access',
+            'property_setting',
+        ]
+
+    def setup(self):
+        for fn in self.funcs:
+            globals()[fn](1)
+
+    @classmethod
+    def generate_benchmarks(cls):
+        for fn in cls.funcs:
+            fname = "time_{}".format(fn)
+            def f(self):
+                globals()[fn](self.N)
+            f.__name__ = fname
+            setattr(cls, fname, f)
+
+JitClassDispatch.generate_benchmarks()