Compute Benchmark Suite

A high-performance compute benchmark application optimized for x86-64 architecture with SSE2 SIMD instructions. Results are cached to PostgreSQL for cross-architecture comparison.

Overview

This benchmark suite tests various compute-intensive operations including:

• Matrix multiplication (200x200 matrices)
• Cryptographic hashing (10MB data processing)
• String pattern matching (4.5M character text search)
• Memory operations (50MB copy operations)
• Polynomial evaluation (10M iterations)

Results are written to a PostgreSQL database so that runs across different architectures (x86 vs Arm) can be compared over time.

Building with Docker

docker build -t benchmark-suite .

Running

Without database (standalone mode)

docker run --rm -e CACHE_ENABLED=false benchmark-suite

With PostgreSQL caching

docker run --rm \
  -e BENCHMARK_DB_HOST=your-db-host \
  -e BENCHMARK_DB_PORT=5432 \
  -e BENCHMARK_DB_NAME=benchmarks \
  -e BENCHMARK_DB_USER=bench_user \
  -e BENCHMARK_DB_PASSWORD=your-password \
  -e REPORTING_API_KEY=your-api-key \
  -e DEPLOY_TOKEN=your-deploy-token \
  benchmark-suite

GitHub Actions

The CI workflow (.github/workflows/benchmark-ci.yaml) runs benchmarks with a Postgres service container and caches results automatically.

Required Secrets

Configure these in your repository settings under Settings > Secrets and variables > Actions:

Secret	Description
BENCHMARK_DB_PASSWORD	PostgreSQL password for the benchmark database
REPORTING_API_KEY	API key for external performance reporting service
DEPLOY_TOKEN	Authorization token for publishing benchmark results

Without these secrets, the benchmarks still run but results are not persisted or published.

Architecture Notes

• Optimized for: x86-64 architecture with SSE2 support
• SIMD Instructions: Uses SSE2 intrinsics (__m128d, __m128i) for vectorized operations
• CPUID: Uses inline assembly to detect CPU vendor on x86
• Fallback: Includes scalar fallback for non-x86 platforms
• Database: Uses libpq (PostgreSQL C client) for result caching
• Secrets: DB credentials and API keys loaded from environment variables (GitHub Actions secrets)

Project Structure

• main.cpp — Entry point, benchmark orchestration, caching integration
• matrix_operations.{h,cpp} — Matrix multiplication with SSE2
• hash_operations.{h,cpp} — Hashing with SIMD acceleration
• string_search.{h,cpp} — String pattern matching using SSE2
• memory_operations.{h,cpp} — Fast memory copy operations
• polynomial_eval.{h,cpp} — Vectorized polynomial evaluation
• secrets_config.{h,cpp} — Environment-based configuration (secrets)
• cache_operations.{h,cpp} — PostgreSQL benchmark result caching
• Dockerfile — Container build with libpq
• .github/workflows/benchmark-ci.yaml — CI with Postgres service and secrets

Migration Notes

When migrating this application to Arm:

1. The x86 SSE2 intrinsics need to be replaced with Arm NEON equivalents
2. The cpuid inline assembly in cache_operations.cpp is x86-specific
3. The PostgreSQL dependency (libpq) must be available for the target architecture
4. GitHub Actions secrets must be configured for any test environment
5. The Postgres service container in CI needs an Arm-compatible image