Embedded Files
Algorithms & Languages for Programming Hybrid Architectures
Algorithms & Languages for Programming Hybrid Architectures
Algorithmic Challenges: AI, Complex Systems, Cybersecurity - Nick's Class, NP hard, NP complete, e.g., Trap-door, Post-Quantum Crypto, & Non-Linear, Large Degree, High Resolution and Dimension challenges (discrete and real number numerical and semi-numerical).
Architecture: Quantum, Neuromorphic (Analog+Digital), & GP-GPU.
Hybrid HPC: Shared memory, Distributed memory, & Co-processing (GP-GPU and QPU).
Group Research: Literature Review Papers Read
Publications: SPIE - Hybrid FPGA vs. GP-GPU; IBM - HPC On-Demand, Big Data, HPC Education; Intel: SSD for Scalable; ASEE - Parallel Programming, HPC Quantum+Parallel; Info: Cluster @Home.
Quantum vs. Parallel Advantage
Quantum vs. Parallel Advantage
Goals: Demonstrate clear quantum advantage compared to parallel advantage for SP (semi-prime) factoring and determine the scaling point at which quantum advantage overtakes parallel of any scale.
Method: Investigation of Grover's and Shor's quantum algorithms and their parallel equivalents for SP factoring. Development of reversible quantum gate programs that scale to any number of qubits (bits for the parallel equivalent). For the parallel comparison we have OpenMP & CUDA scalable code and also use MPI cado-nfs.
Hypothesis: The cross-over where there is clear quantum advantage occurs between SPs of size 66 to 128+ bits (with prime components of 33 to 64+ bits) and currently available quantum computers of 33 qubits or more should allow for demonstration of quantum advantage with time measurement (SU=Tseq/Tpar OR SU=Tseq/Tquant) based on theorized time complexity of each (see analysis, models).
Cross-overs where Quantum is faster than parallel: 16-core Parallel - 26-bit, A100 - 36-bit, ORNL Frontier - 48-bit
Significance: The hypothesized SP quantum advantage can be extended to any RSA challenge size, thus showing that RSA and trap-door function encryption is generally not safe from quantum factoring cryptanalysis as QIST scales (as is theorized). Parallel not only has less time complexity reduction than quantum, but has scaling inefficiency (Amdahl's law), however, it is not clear that this is true for all problems/algorithms and the size of the class of algorithms that benefit from quantum advantage is unknown. Planned publications - Technical: https://qce.quantum.ieee.org, ICPP 25 (ICCP WIki), SPIE Defense+Commercial 2026, arXiv e-Print archive; Educational: QIST 25 poster, ASEE national 2026 & ASEE PSW regional.
SP Challenge References: Schneier Blog on 829-bit digial vs. Quantum, China State Key Lab 2048-bit RSA claim (48-bit demo with 10 qubits), GNFS code, GIMPS, Mersenne, Largest (2024).
More Quantum Conferences, Journals and Talks: Dwave Qubits, POPL PLanQC, SC, QIPC, ACM TQC, IEEE HPC Quantum, IEEE TQE, Dr. Martonosi NSF CISE on Quantum.
Programming: Quantum Gate Programming with CUDA-Q simulation to target Quantinuum/Qiskit and Parallel MPI + CUDA scaling for infinitely scalable most efficient parallel comparison.
Programming Contests:
Eth Denver: https://www.ethdenver.com/
SC Cluster Competition: https://sc25.supercomputing.org/students/student-cluster-competition/
Topic: Complex Systems
Topic: Complex Systems
Programming: R, Python
Algorithms: graphs, non-linear systems, agents, networks, self-organization and emergent properties, optimization, and chaotic dynamics. Includes non-numerical, semi-numerical, and numerical. Examples are QUBO, sub-set sum, graphs and mazes, etc.
More info: CSCI 651 Applied Graph Theory, Santa Fe Institute, Complexity Explained; Clay Math Institute: https://www.claymath.org/millennium-problems/
Topic: Quantum Computing (QIST)
Topic: Quantum Computing (QIST)
Programming: CUDA-Q, Q# - Overview, Qiskit - Overview
Algorithms: Shor's, Simon, Grover's, QUBO, Machine Learning, and Simulation
QIST Introduction: Babu, Hafiz Md Hasan. Quantum Computing: A pathway to quantum logic design. IOP Publishing, 2023, Lala, Parag. Quantum Computing. McGraw Hill Professional, 2019, and CMU Intro.
More info: CSCI 420: Applied Quantum Computing for Computer Scientists, Application of Quantum Gates, IBM Quantum Computer, Sandia Quantum Performance Library, Quantum Jobs Board, QED-C Jobs, Chicago Quantum Exchange, IBM Quantum Scaling
World Quantum Day - https://worldquantumday.org/ , Top Quantum Research Universities, NVIDIA QPU, and Intel Tunnel Falls, Research
Topic: Cybersecurity (Cryptanalysis)
Topic: Cybersecurity (Cryptanalysis)
Programming: Python, and C/C++.
Algorithms: Encryption and hashing: AES, RSA, SHA, & cryptanalysis.
Cybersecurity: CINS 448 Computer Security, CINS 548 Advanced Computer Security
Cryptanalysis: Math 317 Cryptography
Example Ciphers: Caesar, Vigenere, Block cipher, TwoFish, Applied Cryptography, etc.
Privacy Organizations and Resources: EFF, PGP, RSA (RSA info)
Hybrid Computer Architectures
Hybrid Computer Architectures
Quantum Cloud + Parallel
Quantum Cloud + Parallel
Programming: Python
Algorithms: Fundamental, QUBO, simulation (> Poly)
More info: CSCI 420: Applied Quantum Computing for Computer Scientists, IBM Q-Cloud, Google, Amazon, NASA Ames, Qiskit Seminars
Neuromorphic
Neuromorphic
Programming: Deep Learning with Tensorflow, PyTorch, or MATLAB, Open Neuromorphic
Algorithms: R-CNN, SSD, YOLO, etc. (ML)
More info: CSCI 612 Applied Computer Vision, CSCI 581 Machine Learning, EU Brains, Intel Neuromorphic, Google Cloud TPU, Edge AI + Vision Alliance, IEEE Spectrum AI Chips , Intel Neuromorphic Research, Intel Loihi2, IBM True North
Page updated
Report abuse