Benchmarks and Performance

Benchmarks & Performance Guide 📊

This document provides complete benchmark numbers, methodology, and reproduction instructions for VortexMQ v1.0.0.

---

🏁 Summary Benchmark Results

Tested on Apple Silicon (M4 / ARM64, 10 Cores, macOS Sequoia):

Benchmark Function	Operations/sec	Latency	Memory per Op	Allocations
BenchmarkRingBuffer_PushPop	167.1 Million ops/sec	5.98 ns/op	0 B/op	0 allocs/op
BenchmarkBroker_Publish	2.33 Million msgs/sec	428.5 ns/op	279 B/op	6 allocs/op
BenchmarkBroker_ProduceConsume	1.54 Million msgs/sec	647.3 ns/op	285 B/op	7 allocs/op
BenchmarkWAL_Write	642,000 writes/sec	1.55 μs/op	689 B/op	4 allocs/op

---

🔬 Benchmark Analysis & Deep-Dive

1. In-Memory Ring Buffer (5.98 ns/op, 0 B/op)

• What it measures: Direct memory queue ingestion and extraction using our bounded circular ring buffer.
• Why it is so fast:
  • Power-of-two capacity bitwise mask: eliminates arithmetic division instructions.
  • Zero heap allocations: reusable slice pointers ensure the Go garbage collector is never triggered.

2. Multi-Goroutine Parallel Publish (2.33 Million msgs/sec)

• What it measures: Concurrent publisher goroutines hammering the Broker's topic management, payload parsing, and queue ingestion simultaneously.
• Why it scales:
  • Sharded lock striping across topics ensures goroutines publishing to different topics experience zero lock contention.

3. Full Producer-Consumer Pipeline (1.54 Million msgs/sec)

• What it measures: Continuous, simultaneous publish and consume pipelines under heavy multi-client load.

4. Segmented Write-Ahead Log (642,000 commits/sec)

• What it measures: Raw disk serialization with 4-byte length prefix, 4-byte IEEE CRC32 checksum calculation, opcode tagging, and sequential append writes.

---

🛠️ How to Reproduce Benchmarks Locally

You can run the full benchmark suite on your own machine using the Go test runner:

# 1. Clone the repository
git clone https://github.com/GargAnshu9468/vortexmq.git
cd vortexmq

# 2. Run all benchmarks with memory allocation reporting
go test -benchmem -bench=. ./benchmarks/...

Expected Output Format:

goos: darwin
goarch: arm64
pkg: github.com/GargAnshu9468/vortexmq/benchmarks
cpu: Apple M4
BenchmarkRingBuffer_PushPop-10          196265694         5.984 ns/op          0 B/op          0 allocs/op
BenchmarkBroker_Publish-10                2816172       428.5 ns/op          279 B/op          6 allocs/op
BenchmarkBroker_ProduceConsume-10         1909074       647.3 ns/op          285 B/op          7 allocs/op
BenchmarkWAL_Write-10                      765975      1557 ns/op            689 B/op          4 allocs/op
PASS

---

🎛️ Profiling CPU and Memory with pprof

To generate flame graphs and memory allocation profiles:

# Generate CPU profile
go test -bench=BenchmarkBroker_Publish -cpuprofile=cpu.pprof ./benchmarks/...

# Inspect CPU bottlenecks with interactive pprof
go tool pprof -http=:8080 cpu.pprof

# Generate Heap Memory profile
go test -bench=BenchmarkRingBuffer_PushPop -memprofile=mem.pprof ./benchmarks/...
go tool pprof -http=:8080 mem.pprof