Skip to content

DevDiff Stress Testing & Performance Benchmark Suite ​

DevDiff is designed to operate efficiently on high-velocity developer workstations and continuous integration (CI/CD) pipelines. To guarantee deterministic performance, bounded memory usage, and stability across large monorepos, DevDiff runs automated benchmark suites against extreme repository conditions.


Benchmark Architecture & Execution Flow ​


Running the Automated Stress Testing Suite ​

To run the local stress-testing suite against your local workstation environment:

bash
# Give execution permissions (macOS / Linux / WSL / Git Bash)
chmod +x test/stress/full-suite.sh

# Execute the complete automated benchmark suite
./test/stress/full-suite.sh

Environment Benchmark Flag Options ​

bash
# Run with detailed memory profiling enabled
STRESS_PROFILE_MEMORY=true ./test/stress/full-suite.sh

# Run specific concurrency target (e.g. 10 parallel agents)
STRESS_CONCURRENCY=10 ./test/stress/full-suite.sh

Comprehensive Benchmark Matrix ​

Test IDBenchmark NamePrimary Focus AreaTarget / SLA Pass CriteriaVerified Benchmark Results
T1Single File DiffProcess Spawn & AST ParsingExecution in < 200ms118ms
T2Memory Index LookupSub-50ms Persistent MemoryResponse in < 50ms32ms
T31,000 Files Large DiffProgressive Chunking ResiliencyAuto-chunking without context overflowPassed (100% chunks processed)
T45,000 Files Memory CapHeap Memory & Garbage CollectionMemory peak strictly < 512MBPeak RAM 214MB
T55x Swarm ConcurrencyParallel Thread Lock-Free Read/Write5 simultaneous agent queries without lock collisionPassed (0 file lock errors)
T6Edge Cases & SymlinksFile System ResiliencyGracefully skip binaries, broken symlinks, & null bytesPassed (0 crashes)
T7Prompt Injection GuardAdversarial LLM Security100% block rate on prompt/shell injectionsPassed (0 jailbreaks)
T8High-Entropy RedactionSecret Leak Prevention100% masking of API keys & JWT tokensPassed (0 credential leaks)

Detailed Stress Testing Guides ​