High-performance password hash cracking with NVIDIA CUDA GPU acceleration. Cracking de hashes haute performance avec acceleration GPU NVIDIA CUDA.
Features · Installation · Usage · Benchmarks · API
LEGAL DISCLAIMER / AVERTISSEMENT LEGAL
This tool is designed exclusively for authorized penetration testing and educational purposes. Unauthorized use against systems you do not own or have explicit written permission to test is illegal and unethical. The author assumes no liability for misuse of this software. Always obtain proper authorization before testing.
Cet outil est concu exclusivement pour les tests d'intrusion autorises et a des fins educatives. L'utilisation non autorisee contre des systemes que vous ne possedez pas ou pour lesquels vous n'avez pas d'autorisation ecrite explicite est illegale et contraire a l'ethique. L'auteur n'assume aucune responsabilite en cas de mauvaise utilisation de ce logiciel.
┌─────────────────────────────────────────────────────────────────┐
│ HashCracker-GPU Engine │
├─────────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────┐ ┌──────────────┐ ┌───────────────────────┐ │
│ │ CLI Layer │ │ Reporter │ │ Pydantic Models │ │
│ │ (argparse) │ │ JSON/CSV/TXT │ │ HashTarget, Session │ │
│ └──────┬───────┘ └──────┬───────┘ └───────────┬───────────┘ │
│ │ │ │ │
│ ┌──────▼─────────────────▼───────────────────────▼───────────┐ │
│ │ Cracker Engine │ │
│ │ Dictionary | Brute Force | Combination | Mask Attacks │ │
│ │ Policy Analysis | Multi-hash | File Processing │ │
│ └──────┬─────────────────────────────────────────┬───────────┘ │
│ │ │ │
│ ┌──────▼───────────┐ ┌──────────▼──────────┐ │
│ │ Dictionary Mgr │ │ GPU Hash Engine │ │
│ │ Wordlist Loading │ │ CUDA Tensor Ops │ │
│ │ 50+ Mutation Rules│ │ Batch Processing │ │
│ │ Mask Generation │ │ Multi-GPU Support │ │
│ │ Combo Generator │ │ VRAM Management │ │
│ └──────────────────┘ └─────────────────────┘ │
│ │
│ Supported Algorithms: MD5 | SHA1 | SHA256 | SHA512 | NTLM | bcrypt │
└─────────────────────────────────────────────────────────────────┘
- Multi-Algorithm Support: MD5, SHA1, SHA256, SHA512, NTLM (MD4), bcrypt
- GPU Acceleration: NVIDIA CUDA via PyTorch tensor operations for massively parallel hash comparison
- 4 Attack Modes:
- Dictionary Attack: Wordlist-based with 50+ mutation rules (l33t speak, case toggling, digit appending, etc.)
- Brute Force: Configurable charset and length range with keyspace estimation
- Combination Attack: Cross-product of multiple wordlists
- Mask Attack: Pattern-based generation (
?u?l?l?l?d?d=Abcd12)
- 50+ Mutation Rules: l33t speak, case manipulation, digit/special appending, truncation, keyboard walks
- Smart Reporting: JSON/CSV/Text export with password policy analysis and security recommendations
- Multi-GPU Support: Distribute workload across multiple NVIDIA GPUs
- VRAM-Aware: Automatic batch size adjustment based on available GPU memory
- Production-Ready: Pydantic models, comprehensive logging, graceful shutdown
- Support Multi-Algorithmes: MD5, SHA1, SHA256, SHA512, NTLM (MD4), bcrypt
- Acceleration GPU: NVIDIA CUDA via operations tensorielles PyTorch pour comparaison massivement parallele
- 4 Modes d'Attaque:
- Attaque Dictionnaire: Basee sur wordlist avec 50+ regles de mutation (l33t speak, alternance casse, ajout chiffres, etc.)
- Force Brute: Charset et longueur configurables avec estimation du keyspace
- Attaque Combinaison: Produit cartesien de plusieurs wordlists
- Attaque Masque: Generation par pattern (
?u?l?l?l?d?d=Abcd12)
- 50+ Regles de Mutation: l33t speak, manipulation de casse, ajout chiffres/speciaux, troncature, keyboard walks
- Rapports Intelligents: Export JSON/CSV/Texte avec analyse de politique de mots de passe et recommandations
- Support Multi-GPU: Repartition de charge sur plusieurs GPU NVIDIA
- Gestion VRAM: Ajustement automatique de la taille de batch selon la memoire GPU
- Production-Ready: Modeles Pydantic, logging complet, arret propre
# Clone the repository
git clone https://github.com/ayinedjimi/HashCracker-GPU.git
cd HashCracker-GPU
# Install dependencies
pip install -r requirements.txt
# Install in development mode
pip install -e .
# Verify GPU detection
hashcracker-gpu devices- Python 3.9+
- PyTorch 2.0+ (with CUDA for GPU acceleration)
- NVIDIA GPU with CUDA support (optional, CPU fallback available)
# MD5 with dictionary attack
hashcracker-gpu crack 5f4dcc3b5aa765d61d8327deb882cf99 \
-t md5 -w /path/to/rockyou.txt
# SHA256 with mutations
hashcracker-gpu crack <hash> -t sha256 \
-w wordlist.txt -r "capitalize,append_digits,leet_simple"
# NTLM with brute force fallback
hashcracker-gpu crack a4f49c406510bdcab6824ee7c30fd852 \
-t ntlm -c "lower,digits" --max-length 6# Dictionary attack on hash file
hashcracker-gpu batch hashes.txt -t md5 \
-w rockyou.txt -a dictionary \
--output-json report.json --output-csv results.csv
# Brute force attack
hashcracker-gpu batch hashes.txt -t sha1 \
-a bruteforce -c "alphanum" --min-length 4 --max-length 8
# Mask attack
hashcracker-gpu batch hashes.txt -t md5 \
-a mask -m "?u?l?l?l?d?d,?l?l?l?l?l?d"# Benchmark all algorithms
hashcracker-gpu benchmark
# Benchmark specific algorithms
hashcracker-gpu benchmark --hash-types md5,sha256,ntlm -n 1000000hashcracker-gpu wordlist-info /path/to/rockyou.txt /path/to/custom.txtPerformance on NVIDIA RTX 3090 (24GB VRAM):
| Algorithm | Speed (H/s) | MH/s |
|---|---|---|
| MD5 | 45,200,000 | 45.2 |
| SHA1 | 32,800,000 | 32.8 |
| SHA256 | 18,500,000 | 18.5 |
| SHA512 | 8,200,000 | 8.2 |
| NTLM | 52,100,000 | 52.1 |
| bcrypt | 28,500 | 0.03 |
Note: bcrypt is intentionally slow by design. GPU acceleration primarily benefits the comparison phase for non-bcrypt algorithms.
from hashcracker_gpu.cracker import HashCracker
from hashcracker_gpu.models import HashTarget, HashType, MaskPattern
# Initialize cracker
cracker = HashCracker(batch_size=65536)
# Dictionary attack
targets = [HashTarget(hash_value="5f4dcc...", hash_type=HashType.MD5)]
session = cracker.dictionary_attack(targets, ["/path/to/wordlist.txt"])
# Mask attack
masks = [MaskPattern(pattern="?u?l?l?l?d?d")]
session = cracker.mask_attack(targets, masks)
# Generate report
from hashcracker_gpu.reporter import CrackReporter
reporter = CrackReporter(session)
reporter.export_json("report.json")
reporter.print_summary()- NTLM-HASH-CALCULATOR - NTLM hash computation utilities
- TcpSnooper - TCP traffic analysis
- NetFlowLiteCollector - NetFlow collection and analysis
Ayi NEDJIMI
- Website: ayinedjimi-consultants.fr
- GitHub: github.com/ayinedjimi
- HuggingFace: huggingface.co/AYI-NEDJIMI
This project is licensed under the MIT License - see the LICENSE file for details.
HashCracker-GPU - Built for security professionals, by Ayi NEDJIMI
For authorized penetration testing and educational purposes only.