Project Stage 2: Profiling (x86_64)

-
Introduction
As the profiling I have done on the previous blog was using Aarch64 architecture, I will do profiling in x86_64 architecture and describe the result. I recommend to take a look at the previous blogs related to this project: Project Stage 1 - Compining and Benchmarking Butteraugli and Project Stage 2 - Profiling Butteraugli

 The Project I Selected: Butteraugli
The Butteraugli is the project Google created. It estimates the psychovisual similarity of two images. The detailed information about this project is described on the Google Opensource website about the Butteraugli. The source code for this project is located here: GitHub - Butteraugli.

 How to use gprof and perf
The detailed process of how to use gprof and perf for Butteraugli profiling is described in the Project Stage 2 - Profiling Butteraugli. 

 Result of the prof in x86_64 architecture
Image size: 5mb

 Image size: 10mb

 Result of the perf in x86_64 architecture
Image size: 5mb

 Image size: 10mb

 Conclusion
I have checked the major functions takes the most time in this project are butteraugli::Convolution(butteraugli::Image<float> const&, std::vector<float>, std::vector<float, std::allocator<float>> const&, float) and std::vector<float, std::allocator<float> >::operator[](unsigned long) in both the Aarch64 and x86_64 architectures. Therefore, If I optimize those functions, this project can be faster in both Aarch64 and x86_64 architecture.

 Reflection
As I have already done profiling in Aarch64 architecture, profiling in x86_64 architecture is not difficult. I have fun profiling this project. I will focus on two functions take the most time for this project, optimize them, and blog the result on the stage 3 post.

Comments

Post a Comment

Popular posts from this blog

Open Source Project: Lab2 (Pull Request)

-
Image
 Introduction In this lab, I practiced contributing and submitting pull requests to other repositories I don't own. Following the lab instruction , the things I plan to practice are the following: forking and cloning other projects creating branches to work on new features and fix bugs working on code I didn't write, trying to maintain the original style and not break things creating pull requests collaborating with other developers on GitHub updating my pull requests to include fixes for review comments The Project I picked for pull request ( https://github.com/lixiaoqity/testLink/pull/7 ) I looked around the classmates' open source project and picked the project named testLink. The GitHub repository link is the following:  https://github.com/lixiaoqity/testLink/pull/7 .  I picked this project because it also uses the NodeJS which I picked for my project. It was really interesting to see how code can be written differently for the same subject. When I look over the project
Read more

Release 4.0 - Part 3: Release

-
In this post, I am going to describe the journey of the release 4.0. Release 4.0 is the 3 weeks project. In the first week, I chose the project to work on. As the issue I chose required knowledge I do not have any idea (Gatsby and Nextjs), I need to learn these. In the second week, I learned the Gatsby and Nextjs to read the document on the Next official website. Also, I checked the pull request other people already created for the project. And this week, I started writing codes to resolve the issue. It was hard to begin to migrate the pages from Gatsby to the Nextjs. I needed to google a lot to find the proper way to migrate the code for the project. The components I was assigned are Error page, SEO, and SearchHelp components. The SEO component is relatively easy because the Next official website indicated how to do it most. However, the Error page is the most difficult part for me. When I googled how to make the Nextjs Errorpage, most people recommend to use the JavaScript extension
Read more

Lab 8: Automated Testing and Continuous Integration

-
When I developed a program, it was time-consuming to test all the functions I created when I add a feature. I needed to run each feature, check whether it gives expected output or not, and check all the output. It is possible I forgot to test a feature and later on, I realize I did not test the feature. I wonder if there any ways I can do it automatically. Now, I got the way how to do that. There is automated testing for it! To begin with it, I started searching the testing framework to apply my command-line tool named url-fi. There are some testing frameworks for Node.js such as Jest, Mocha, and AVA. The testing framework I chose is  Jest  which is a testing framework developed by Facebook. I've installed Jest running 'npm install --save-dev jest' and test whether it works properly or not.  This document  is the way I followed to install the Jest testing framework. After I've installed the Jest testing framework, I worked on separating functions. As my tool contains al
Read more