merrymercy commented on a change in pull request #5962: URL: https://github.com/apache/incubator-tvm/pull/5962#discussion_r449995770
########## File path: python/tvm/ansor/measure.py ########## @@ -0,0 +1,386 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +"""Distributed measurement infrastructure to measure the runtime costs of tensor programs + +These functions are responsible for building the tvm module, uploading it to +remote devices, recording the running time costs, and checking the correctness of the output. + +We implement these in python to utilize python's multiprocessing and error handling +""" + +import os +import time +import shutil +import traceback +import tempfile +import multiprocessing + +import tvm._ffi +from tvm.runtime import Object, module, ndarray +from tvm.driver import build_module +from tvm.ir import transform +from tvm.contrib import tar, ndk + +from . import _ffi_api +from .utils import get_const_tuple, NoDaemonPool, call_func_with_timeout + +# The maximum length of error message +MAX_ERROR_MSG_LEN = 512 + +# Global variables used in build function +GLOBAL_BUILD_ARGUMENTS = None + +@tvm._ffi.register_object("ansor.MeasureCallback") +class MeasureCallback(Object): + """ Base class for measurement callback function. """ + + +@tvm._ffi.register_object("ansor.MeasureInput") +class MeasureInput(Object): + """ Store the input of a measurement. + + Parameters + ---------- + task : SearchTask + The target SearchTask. + state : State + The current State to be measured. + """ + def __init__(self, task, state): + self.__init_handle_by_constructor__(_ffi_api.MeasureInput, task, state.state_object) + + +@tvm._ffi.register_object("ansor.BuildResult") +class BuildResult(Object): Review comment: This follows the name convention in autotvm. We separate the measurement into two steps: build and run. So we have the flow `MeasureInput -> BuildResult -> MeasureResult` A builder takes `MeasureInput` and retusn `BuildResult`. A runner then takes `BuildResult` and returns `MeasureResult` ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
