As i coded on OuroborosDB i noticed that i need a very fast queue for a rather unique architectural design decision.
I try to build the network module in such a way that i can test the behavior completely deterministic while “simulating” entire clusters in a single process.
So i build a test prototype of my global network queue with go’s channels and noticed that it was a major performance bottleneck, after writing 2 different ring buffer queue implementations it became clear that some queues behave completely different under different congestion levels and core counts - some so unpredictable that i just did not wanted to use them in my project.
This prompted me to take a relatively large chunk out of my free time and write a suite to benchmark different queue implementations i build under different conditions and score them based on their performance and predictability.
The result of this work is GoQueueBench
These are the results of the benchmark suite:
| Implementation | Overall Score | Throughput Light Load | Throughput Heavy Load | Throughput Average | Stability Ratio | Homogeneity Factor | Uncertainty | Total Tests |
|---|---|---|---|---|---|---|---|---|
| VortexQueue | 11341466 | 6926449 | 5502925 | 8776309 | 1.15 | 0.87 | 0.25 | 681 |
| LightningQueue | 9631771 | 6638213 | 4627690 | 6036728 | 0.99 | 0.95 | 0.31 | 681 |
| FastMPMCQueue | 9384067 | 6870924 | 4598620 | 6070151 | 0.96 | 0.93 | 0.28 | 681 |
| OptimizedMPMCQueue | 9105262 | 6436385 | 4379823 | 5838555 | 0.97 | 0.94 | 0.32 | 681 |
| OptimizedMPMCQueueSharded | 8130197 | 6369891 | 3834140 | 6781865 | 0.84 | 0.88 | 0.39 | 681 |
| MultiHeadQueue | 7391203 | 4363332 | 3492068 | 5558849 | 1.12 | 0.91 | 0.36 | 681 |
| BasicMPMCQueue | 5599252 | 4370889 | 2669612 | 3667715 | 0.89 | 0.93 | 0.30 | 681 |
| Golang Buffered Channel | 5312485 | 6667828 | 2760985 | 4312720 | 0.54 | 0.82 | 0.66 | 681 |
| FastMPMCQueueTicket | 3229780 | 7705164 | 1203924 | 5803821 | 0.21 | 0.64 | 1.19 | 681 |
Please note that i build the package so that all queue adhere to the same interface and can be swapped out easily.