GitPedia

Dperf

dperf: High-Performance Network Load Testing Tool Based on DPDK

From baidu·Updated June 14, 2026·View on GitHub·

**dperf** is a high-performance network traffic generator and load testing tool based on **DPDK**. The project is written primarily in C, distributed under the Apache License 2.0 license, first published in 2021. It has gained significant community traction with 5,565 stars and 556 forks on GitHub. Key topics include: benchmark, dpdk, l4lb, linux-networking, load-testing.

Latest release: v1.9.0
July 23, 2025View Changelog →

dperf Apache V2 License Mentioned in ChatGPT <a href="https://hellogithub.com/repository/67958cc5d1f44a6a84f3544e3c007e5f" target="_blank"><img src="https://abroad.hellogithub.com/v1/widgets/recommend.svg?rid=67958cc5d1f44a6a84f3544e3c007e5f&claim_uid=Thc9mJByaKSbdng&theme=small" alt="Featured|HelloGitHub" /></a>

dperf is a high-performance network traffic generator and load testing tool based on DPDK.

Advantages

  • High Performance

    Built on DPDK, dperf can generate massive traffic using a single x86 server — achieving tens of millions of HTTP Connections Per Second (CPS), hundreds of Gbps throughput, and billions of concurrent connections.

  • Comprehensive Statistics

    Provides detailed real-time metrics and identifies every packet drop or error.

  • Versatile Use Cases

    • Load and stability testing for Layer 4 Load Balancers and gateways
    • Network performance benchmarking for cloud servers
    • Evaluation of NIC and CPU packet processing capabilities
    • Acts as a high-performance HTTP server or client for testing scenarios

Performance

All results were measured using the following hardware and configuration:

  • CPU: Intel Xeon Gold 5418Y × 2
  • Memory: 128 GB × 2
  • NIC: Intel E810-C Dual-Port 100GbE × 2
  • OS: Linux 5.10.0

HTTP Connections Per Second (CPS)

Client CoresServer CoresHTTP CPS (Million)Client CPU Usage (%)Server CPU Usage (%)
1147471
2287472
44167370
88327068
1616647068

HTTP Throughput (Gbps)

Client CoresServer CoresRX ThroughputTX ThroughputClient CPU Usage (%)Server CPU Usage (%)
1198.3 Gbps98.3 Gbps7880
22196.7 Gbps196.7 Gbps7882

Concurrent Connections

Client CoresServer CoresConnections (Billion)Client CPU Usage (%)Server CPU Usage (%)Memory Usage (GB)
111484860
2224848120
4444848240

TCP Packets Per Second (PPS)

Client CoresRX PPS (Mpps)TX PPS (Mpps)Client CPU Usage (%)
116.816.899
6105.2105.299
12204.6204.699

Real-Time Statistics

dperf prints real-time statistics every second, including CPS, TPS, PPS, packet drops, socket errors, and HTTP status counts.
Example output:

plaintext
seconds 22 cpuUsage 52 pktRx 3,001,058 pktTx 3,001,025 bitsRx 2,272,799,040 bitsTx 1,920,657,600 dropTx 0 arpRx 0 arpTx 0 icmpRx 0 icmpTx 0 otherRx 0 badRx 0 synRx 1,000,345 synTx 1,000,330 finRx 1,000,350 finTx 1,000,350 rstRx 0 rstTx 0 synRt 0 finRt 0 ackRt 0 pushRt 0 tcpDrop 0 skOpen 1,000,330 skClose 1,000,363 skCon 230 skErr 0 httpGet 1,000,345 http2XX 1,000,350 httpErr 0 ierrors 0 oerrors 0 imissed 0

Documentation

See the official website at https://dperf.org/.

Contributing

We welcome contributions! Please see the CONTRIBUTING file for details.

Patent

  • Title: Testing Method and Apparatus for Network Devices
  • Inventor: Jianzhang Peng
  • Patent Number: CN114205274B
  • Issue Date: June 11, 2024

Acknowledgment

We gratefully acknowledge xnetin for providing the high-performance testing platform used in our benchmarking experiments.

Author

Jianzhang Peng holds a Ph.D. in Computer Science from the University of Science and Technology of China (USTC). He previously worked as a Principal Engineer at Baidu, where he contributed to the development of high-performance L4 load balancer systems. He initiated and developed the dperf project during his time at Baidu, and continues to maintain it as an open-source contributor. His current focus is on low-latency network protocol stacks for quantitative trading systems.

License

dperf is licensed under the Apache License 2.0.

Contributors

Showing top 11 contributors by commit count.

View all contributors on GitHub →

This article is auto-generated from baidu/dperf via the GitHub API.Last fetched: 6/14/2026