This month we spotlight Lokman Abbas Turki, lecturer and researcher at Sorbonne University in Paris, France.
Overview
The article features Lokman Abbas Turki, a researcher at Sorbonne University, who applies high performance computing (HPC) to complex mathematical finance problems and cryptography. It highlights his research focus on probability, Monte Carlo methods, and the use of NVIDIA GPUs for accelerating computations.
What You'll Learn
1
How to apply Monte Carlo methods in mathematical finance using GPUs
2
Why high performance computing is crucial for cryptography resilience
3
When to use parallel architectures for computational problems
Prerequisites & Requirements
- Basic understanding of probability and high performance computing
- Familiarity with CUDA and parallel programming(optional)
Key Questions Answered
What is Lokman Abbas Turki's research focus?
Lokman Abbas Turki's research focuses on applying probability and high performance computing to complex mathematical finance problems and enhancing the resilience of asymmetric cryptosystems against side-channel attacks. His work often utilizes Monte Carlo methods, which are well-suited for parallel computing environments.
How has Lokman Abbas Turki utilized NVIDIA technology in his research?
He has implemented efficient CUDA parallelization of his algorithms on NVIDIA GPUs, achieving speedups exceeding 20 times compared to vectorization using AVX on CPUs. For highly parallel operations, speedups can exceed 200 times, significantly enhancing computational efficiency.
What challenges does Turki's research address in cryptography?
Turki's research addresses challenges in asymmetric cryptography, particularly regarding the resilience of cryptosystems against sophisticated template attacks. His work involves studying the impact of randomization in cryptographic computations to enhance security against side-channel leaks.
What are some technological breakthroughs achieved by Turki?
Turki has achieved breakthroughs in simulating and controlling bias in high-dimensional problems related to parabolic PDEs and has proposed a novel training method for deep learning that accelerates convergence for complex financial simulations, such as Credit Valuation Adjustment.
Key Statistics & Figures
Speedup factor of CUDA parallelization
Exceeds 20 times
Compared to vectorization using AVX on CPUs, with some operations achieving over 200 times speedup.
Technologies & Tools
Backend
Cuda
Used for parallelizing algorithms on NVIDIA GPUs to enhance computational performance.
Hardware
Nvidia Gpus
Provide the necessary computing power for high performance computing tasks in Turki's research.
Key Actionable Insights
1Leverage high performance computing to enhance the efficiency of financial simulations.Utilizing GPUs can significantly reduce computation time for complex financial models, making it feasible to conduct large-scale simulations that were previously impractical.
2Adopt stochastic methods based on Monte Carlo for scalable solutions.These methods are more adaptable to parallel architectures and can be integrated with machine learning techniques, providing a robust approach to solving complex problems in finance and cryptography.
3Focus on developing skills in both probability/statistics and low-level programming.Mastering these areas is essential for researchers aiming to work on scalable solutions that leverage modern computing power effectively.
Common Pitfalls
1
Neglecting the importance of scalability in research projects.
Many researchers focus on theoretical aspects without considering how their methods will perform with increasing data sizes or computing power. This can lead to solutions that are not practical for real-world applications.
Related Concepts
High Performance Computing
Monte Carlo Methods
Asymmetric Cryptography
Deep Learning In Finance