# 🚀 Performance Benchmarks: TQL-11 (AVX2)
Tequel-rs (TQL-11) is engineered to bridge the gap between high-entropy (384-bit) and high-throughput processing. Below is the comparative analysis between standard hashing algorithms and Tequel-rs v1.0.0.
## 📊 Throughput Comparison (Parallel / Rayon)
| SHA-256 | 256 | `410` | 1.0x (Baseline) |
| SHA-3 | 256 | `280` | 0.68x |
| **TEQUEL (TQL-11)** | **384** | **`1,250`** | **3.05x** |
| BLAKE3 | 256 | `2,100` | 5.12x |
---
## 🔬 Core Efficiency (x86_64)
The TQL-11 engine is designed to saturate the memory bus on modern CPUs by utilizing **256-bit YMM registers** and **2x instruction unrolling**.
* **Multi-Core Throughput:** ~1.2 GiB/s (Scalable with `rayon`)
* **Single-Core Throughput:** ~23 MiB/s
* **Cycles per Byte (CpB):** ~3.1 (Optimized for AVX2)
* **Register Pinning:** Direct utilization of 16x YMM registers to minimize cache misses.
## 🛠️ Testing Environment
Benchmarks were performed on a standardized industrial-grade node:
- **Architecture:** x86_64 with Native AVX2 Support.
- **Memory:** Zero-allocation, stack-only processing during the core mix.
- **Compiler:** `rustc 1.75+` with `--release` and `target-cpu=native`.
> **Note:** Performance is memory-bound during large-scale parallelization. Results may vary based on CPU cache size and RAM frequency.