add micro-benchmark for PEP 492 coroutines

Demos/benchmarks/coroutines.py

+#!/usr/bin/python3
+# cython: language_level=3
+# micro benchmarks for coroutines
+
+COUNT = 100000
+
+import cython
+
+
+async def done(n):
+    return n
+
+
+@cython.locals(N=cython.Py_ssize_t)
+async def count_to(N):
+    count = 0
+    for i in range(N):
+        count += await done(i)
+    return count
+
+
+async def await_all(*coroutines):
+    count = 0
+    for coro in coroutines:
+        count += await coro
+    return count
+
+
+@cython.locals(N=cython.Py_ssize_t)
+def bm_await_nested(N):
+    return await_all(
+        count_to(N),
+        await_all(
+            count_to(N),
+            await_all(*[count_to(COUNT // i) for i in range(1, N+1)]),
+            count_to(N)),
+        count_to(N))
+
+
+def await_one(coro):
+    a = coro.__await__()
+    try:
+        while True:
+            await_one(next(a))
+    except StopIteration as exc:
+        result = exc.args[0] if exc.args else None
+    else:
+        result = 0
+    return result
+
+
+def time(fn, *args):
+    from time import time
+    begin = time()
+    result = await_one(fn(*args))
+    end = time()
+    return result, end-begin
+
+
+def benchmark(N):
+    times = []
+    for _ in range(N):
+        result, t = time(bm_await_nested, 1000)
+        times.append(t)
+        assert result == 8221043302, result
+    return times
+
+
+main = benchmark
+
+if __name__ == "__main__":
+    import optparse
+    parser = optparse.OptionParser(
+        usage="%prog [options]",
+        description="Micro benchmarks for generators.")
+
+    import util
+    util.add_standard_options_to(parser)
+    options, args = parser.parse_args()
+
+    util.run_benchmark(options, options.num_runs, benchmark)