Merge pull request #30 from tongzhou80/main

Added new framework APPy, and its implementation for one benchmark go_fast

Author: alexnick83

framework_info/appy.json

@@ -0,0 +1,10 @@
+{
+    "framework": {
+        "simple_name": "appy",
+        "full_name": "APPy",
+        "prefix": "ap",
+        "postfix": "appy",
+        "class": "APPyFramework",
+        "arch": "gpu"
+    }
+}

npbench/benchmarks/go_fast/go_fast_appy.py

@@ -0,0 +1,13 @@
+# https://numba.readthedocs.io/en/stable/user/5minguide.html
+
+import torch
+import appy
+
+@appy.jit
+def go_fast(a):
+    trace = torch.zeros(1, dtype=a.dtype)
+    #pragma parallel for
+    for i in range(a.shape[0]):
+        #pragma atomic
+        trace[0] += torch.tanh(a[i, i])
+    return a + trace

npbench/infrastructure/__init__.py

@@ -10,4 +10,5 @@
 from .legate_framework import *
 from .numba_framework import *
 from .pythran_framework import *
+from .appy_framework import *
 from .jax_framework import *

npbench/infrastructure/appy_framework.py

@@ -0,0 +1,52 @@
+# Copyright 2021 ETH Zurich and the NPBench authors. All rights reserved.
+import pkg_resources
+
+from npbench.infrastructure import Benchmark, Framework
+from typing import Any, Callable, Dict
+
+
+class APPyFramework(Framework):
+    """ A class for reading and processing framework information. """
+
+    def __init__(self, fname: str):
+        """ Reads framework information.
+        :param fname: The framework name.
+        """
+
+        super().__init__(fname)
+
+    def version(self) -> str:
+        """ Return the framework version. """
+        return 0.1
+
+    # def copy_func(self) -> Callable:
+    #     """ Returns the copy-method that should be used 
+    #     for copying the benchmark arguments. """
+    #     import cupy
+    #     return cupy.asarray
+
+    def copy_func(self) -> Callable:
+        import torch
+        torch.set_default_device('cuda')
+        def inner(arr):
+            copy = torch.from_numpy(arr).to('cuda')
+            return copy
+        return inner
+
+    def imports(self) -> Dict[str, Any]:
+        import torch
+        import appy
+        return {'torch': torch}
+
+    def exec_str(self, bench: Benchmark, impl: Callable = None):
+        """ Generates the execution-string that should be used to call
+        the benchmark implementation.
+        :param bench: A benchmark.
+        :param impl: A benchmark implementation.
+        """
+
+        arg_str = self.arg_str(bench, impl)
+        # param_str = self.param_str(bench, impl)
+        main_exec_str = "__npb_result = __npb_impl({a})".format(a=arg_str)
+        sync_str = "torch.cuda.synchronize()"
+        return main_exec_str + "; " + sync_str