mingfeima

(Internal Tranining Material)

Usually the first step in performance optimization is to do profiling, e.g. to identify performance hotspots of a workload. This gist tells basic knowledge of performance profiling on PyTorch, you will get:

• How to find the bottleneck operator?
• How to trace source file of a particular operator?
• How do I indentify threading issues? (oversubscription)
• How do I tell a specific operator is running efficiently or not?

This tutorial takes one of my recent projects - pssp-transformer as an example to guide you through path of PyTorch CPU peformance optimization. Focus will be on Part 1 & Part 2.

danijar

# Working example for my blog post at:
# https://danijar.github.io/structuring-your-tensorflow-models
import functools
import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data


def doublewrap(function):
    """
    A decorator decorator, allowing to use the decorator to be used without

endolith

% From http://phaseportrait.blogspot.com/2008/06/sinc-interpolation-in-matlab.html

% Ideally "resamples" x vector from s to u by sinc interpolation
function y = sinc_interp(x,s,u)
    % Interpolates x sampled sampled at "s" instants
    % Output y is sampled at "u" instants ("u" for "upsampled")
    % (EXPECTS x, s, and u to be ROW VECTORS!!)

    % Find the period of the undersampled signal
    T = s(2)-s(1);