ESST / Modules / AES Cryptography

AES Cryptography

Hardware Accelerated Encryption & XTS Mode Stress Testing

Overview

The AES Cryptography module implements hardware-accelerated Advanced Encryption Standard (AES) operations using Intel's AES-NI instruction set. This module provides both AES-128 and AES-256 encryption/decryption with optimized XTS mode for maximum CPU utilization and cryptographic throughput testing.

Cryptographic Algorithms

AES-128/256 Block Cipher

Hardware-accelerated implementation using Intel AES-NI instructions. Supports both 128-bit and 256-bit key lengths with full round transformation including SubBytes, ShiftRows, MixColumns, and AddRoundKey operations performed in silicon.

; AES-256 encryption rounds vpxor xmm0, xmm0, [rdx] ; AddRoundKey vaesenc xmm0, xmm0, [rdx+16] ; Round 1-13 vaesenclast xmm0, xmm0, [rdx+224] ; Final round

XTS Mode Parallel Processing

Ultra-intensive parallel AES encryption/decryption in XTS mode, processing 4 blocks simultaneously with tweak operations. Designed for maximum CPU stress through extensive register utilization and unrolled cryptographic rounds.

; Parallel 4-block processing vpxor xmm11, xmm11, xmm15 ; Apply tweak vaesenc xmm11, xmm11, xmm0 ; Parallel encryption vaesenc xmm12, xmm12, xmm0 ; 4 blocks at once

Cryptographic Features

🔐
AES-NI Acceleration
Hardware-level cryptographic processing using dedicated silicon
🔑
Dynamic Key Expansion
Optimized AES-256 key schedule generation with RCON handling
⚡
XTS Mode
Tweakable block cipher for sector-level encryption stress testing

Hardware Stress Targets

🛡️

AES-NI Execution Units

Dedicated cryptographic hardware units are saturated with continuous AES encryption/decryption operations at maximum throughput.

🧮

SIMD Register File

Extensive use of XMM registers for parallel block processing stresses the 128-bit SIMD register allocation and renaming mechanisms.

🔄

Memory Bandwidth

High-throughput key schedule access and block data movement tests memory subsystem performance under cryptographic workloads.

⚙️

Instruction Pipeline

Unrolled cryptographic rounds create complex instruction dependencies that challenge out-of-order execution and pipeline efficiency.

Performance Characteristics

Cryptographic Throughput

14 AES-256 Rounds
4x Parallel Blocks
240B Key Schedule Size
100% AES-NI Utilization

Technical Implementation

Hardware Optimization Features:

  • Intel AES-NI instruction utilization for maximum hardware acceleration
  • Parallel 4-block processing to saturate cryptographic execution units
  • Unrolled round transformations eliminate loop overhead
  • Pre-loaded round keys minimize memory access latency
  • XTS tweak operations for additional computational complexity

Assembly Optimizations:

  • VAESENC/VAESDEC instructions for single-cycle round operations
  • VMOVDQU for efficient 128-bit block data movement
  • Register-heavy key scheduling to minimize cache pressure
  • Conditional processing for variable block counts
  • VAESKEYGENASSIST for hardware-accelerated key expansion

Cryptographic Standards:

  • FIPS 197 compliant AES implementation
  • IEEE P1619 XTS mode for tweakable encryption
  • Support for both 128-bit and 256-bit key lengths
  • Proper RCON handling in key schedule generation
⚠️ Hardware Requirements & Performance Impact

This module requires Intel AES-NI or AMD equivalent cryptographic acceleration. Will consume 100% of available cryptographic execution units and may cause thermal throttling under sustained loads. Ensure adequate cooling and verify AES-NI support before execution.

← Back to Modules