Customer Cases
Pricing

How to Enhance Your Performance Testing with PerfDog Custom Data Extension

Discover how to integrate PerfDog Custom Data Extension into your project for more accurate and convenient performance testing and analysis.
 

Introduction

To address performance issues more effectively, PerfDog now supports real-time synchronization of custom performance-related data. This feature enables data display, storage, and comparison, allowing you to add labels and comments directly to your code. With PerfDog Custom Data Extension, you can synchronize detailed data, such as vertex or texture, memory usage, DrawCall, Triangle, network conditions, and more. Additionally, you can store logical data, such as scene, coordinates, orientation, number of monsters or characters, and particle effects, to facilitate scene reproduction and analysis.

 

Follow the steps below to integrate PerfDog Custom Data Extension into your project.

Integrating PerfDogExtension

1. Add PerfDog source files to your project

Download the perfdog_extension.h and perfdog_extension.cpp source files provided by PerfDog from Github and add them to your project.

2. Enable PerfDog custom data expansion

Call the int EnableSendToPerfDog() interface after your game starts to enable PerfDog custom data expansion. This interface creates a socket and listens on port 53000.

3. Send custom data

PerfDog Custom Data Extension supports sending floating-point numbers, integers, and strings, with one key corresponding to one to three values. Use the following functions to send custom data:

a. Floating-point numbers

void PostValueF(const std::string& category, const std::string& key, float a);
void PostValueF(const std::string& category, const std::string& key, float a, float b);
void PostValueF(const std::string& category, const std::string& key, float a, float b, float c);

b. Integers

void PostValueI(const std::string& category, const std::string& key, int a);
void PostValueI(const std::string& category, const std::string& key, int a, int b);
void PostValueI(const std::string& category, const std::string& key, int a, int b, int c);

 

c. Strings

void PostValueS(const std::string& category, const std::string& key, const std::string& value);

 

d. Set scene label

void setLabel(const std::string& name);

 

e. Add tag

void addNote(const std::string& name);

 

After receiving the data, the client will organize all data with the same category into a table, with the title as the category and the line name as the key. If there are multiple values, the line names will be key_a, key_b, key_c.

 

4. Disable PerfDogExtension

To disable the PerfDogExtension, comment the code in the frame in perfdog_extension.h.

 

Known Issues

● On macOS, iOS devices cannot obtain custom data in Wi-Fi mode.

Performance Testing

Test environment: Pixel 3, small core locked at 1766MHz, test program bound to small core.

Test results: When PostValueI is called 20,000 times per second, the CPU usage is 1% to 2%.

Troubleshooting

● Android: Use adb logcat to check the phone log and search for "PerfDogExtension". For example, the error log for an occupied port:

 

dres2022-07-27 17:15:29.495 8680-8682/? E/PerfDogExtension: bind:Ads already in use

 

● iOS: Use the console on macOS to view phone logs and search for "PerfDogExtension".

In Conclusion

Integrating PerfDog Custom Data Extension into your project can significantly improve your performance testing and analysis capabilities. By following the steps outlined in this guide, you can easily enable real-time synchronization of custom performance-related data, allowing for more accurate and convenient testing. With support for various data types, scene labels, and tags, PerfDog Custom Data Extension is a powerful tool for developers seeking to optimize their applications and games. Don't let performance issues hinder your project's success – harness the power of PerfDog Custom Data Extension today.

PD网络测试推广
Latest Posts
133 LLM Evaluation Metrics: A Complete Guide for 2026 | Performance, Quality & Cost Learn how to evaluate Large Language Models with 33 essential metrics covering latency, output quality, safety, and cost. Includes a practical learning roadmap for AI engineers and testers.
2CAP & BASE Theory: Distributed System High Availability & Chaos Engineering Learn the CAP and BASE theories for distributed systems, including Consistency, Availability, Partition Tolerance, and practical chaos engineering testing strategies for Kubernetes and MySQL architectures.
3LLM-Powered Test Case Generation & Optimization: Full QA Practical Guide Master LLM-powered test case generation & full lifecycle optimization. Learn standardized workflows, edge case design, enterprise implementation & common pitfalls for modern QA teams.
4How to Build a Complete Performance Testing Knowledge System Learn how to build a systematic performance testing knowledge system, master core terminology, pressure models, system architecture, monitoring, troubleshooting and practical testing skills.
5Prompt Engineering for Intelligent Testing: LLM Optimization & Cases Master 6 core prompt optimization techniques for AI-powered intelligent testing. Explore real enterprise cases, common pitfalls and best practices to stabilize LLM outputs for software testing.