_5d4b8c.png)
_02926c.png)
Hello everyone,
After a rigorous and fair benchmarking competition between the three major Ragnarok server emulatorsâHorizon, Hercules, and rAthenaâI'm excited to announce the final results and crown the winner!
Test Methodology:
- All servers were set up with identical world states: the npc folder was removed (renamed to npc-removed) to ensure no NPC scripts or events could affect performance. Only the core source code differed.
- The test measured the average response time for walk (move) packets and the number of packets required to achieve 2,000 successful walk responses.
- This approach isolates the networking and core logic performance of each emulator.
Results:
Server | Avg. Walk Response Time (s) | Walk Packets Sent | Walk Packets Received |
|---|---|---|---|
Horizon | 0.016543 | 7,330 | 2,000 |
Hercules | 0.017796 | 34,496 | 2,000 |
rAthena | 0.033758 | 25,093 | 2,000 |
Technical Details:
- Horizon demonstrated the fastest and most reliable walk packet handling, with the lowest average response time and the fewest packets needed for 2,000 successful responses.
- Hercules performed well in response time but required nearly five times as many packets as Horizon, indicating more packet loss or retries.
- rAthena had the slowest response time and also required a high number of packets, showing both higher latency and less efficient packet processing.
Conclusion:
> The final winner of the competition is: Horizon
Congratulations to the Horizon team! for their outstanding performance in both speed and reliability under identical, script-free conditions!
If you have questions about the methodology or want to see the raw data, feel free to ask.
A patch for this benchmark has been added to this post, if you wish to check out the results on your own and run the code you are free to do so.
â Protos
Recommended Comments