This directory contains performance testing utilities and stress tests for the LibTTAK subsystems.

Core Benchmarks

1. TTL Cache Multi-threaded Benchmark

Located in ttl-cache-multithread-bench/.

• Focus: Lock-free shared memory throughput under heavy churn.
• Strategy: Uses Choi Seok-jeong's OLS principle to achieve extreme Ops/s by trading off RSS.
• Peak Result: 29.2M Ops/s (Lock-Free Shards).

2. General Benchmark Hub (<tt>ttak_bench.c</tt>)

• Focus: Micro-benchmarking core primitives (allocators, atomics, math).

Architectural Philosophy

We prioritize predictable latency and massive throughput over minimal memory footprint.

• Traditional C libraries often sacrifice Ops/s to save a few kilobytes of RAM.
• LibTTAK reverses this: we use segmented memory structures and deterministic isolation to ensure that synchronization never becomes a CPU bottleneck.

Analysis Tools

Use the visualize_memory_calls.py script in the root scripts/ directory to analyze allocation patterns generated by these benchmarks.