Research & Consulting
Montmory PRNG & structural diagnostics for Monte Carlo and HPC.
ENINCA Consulting focuses on pseudo-random generator design, structural diagnostics for Monte Carlo algorithms, and high-performance simulation (CPU, GPU Metal, CUDA).
The goal is to better understand how random streams behave inside real-world simulation workflows, and how to reduce compute costs without sacrificing accuracy.
Montmory_TACM PRNG
Montmory_TACM is an experimental 64-bit pseudo-random generator designed for deterministic Monte Carlo and massively parallel workloads.
It has been empirically validated using the full BigCrush battery (TestU01, 160/160 tests) on:
- CPU (Apple M1, C implementation)
- GPU Apple Metal
- GPU NVIDIA CUDA (A10)
Montmory_TACM is a research PRNG; it is not intended for cryptographic or security-sensitive use.
Structural Monte Carlo diagnostics
Beyond pass/fail testing (e.g. TestU01), ENINCA is developing a lightweight diagnostic module to evaluate the structural behaviour of PRNGs and Monte Carlo pipelines.
These diagnostics aim to provide interpretable indicators that help:
- compare different generators in a given workflow,
- understand how many samples are effectively informative,
- estimate the cost of improving precision in large-scale simulations.
Early results suggest that some PRNGs require ~15–20% more samples than others to reach the same precision, even when they all pass the same batteries.
Domains & use cases
This work is relevant to any context where large Monte Carlo campaigns are used:
- Financial risk (VaR, ES, XVA)
- Uncertainty quantification (UQ) and reliability
- Aeronautics and energy simulation
- HPC Monte Carlo pipelines on CPU/GPU clusters
About & contact
ENINCA Consulting is led by Pascal Montmory, working at the intersection of pseudo-random generation, Monte Carlo theory and high-performance computing.
Detailed technical notes and benchmarks can be shared under NDA for research or industrial evaluation.