Abstract | ||
---|---|---|
The simulation of algorithms from quantum computing is currently the most affordable solution for development of new applications. Due to the high computational cost of such simulation, solutions towards novel features that increase the performance are always desired. This work proposes an extension for the D-GM's simulation framework, establishing the support for GPU-aware distributed quantum simulation. The project explores the concepts of heterogeneous computing, merging distributed and GPU computing in a single programming environment. Our results comprehend the distributed simulation of systems comprised by Hadamard transformations up to 21 qubits. Detailed analysis and a performance comparison between PyCUDA and JCUDA frameworks for our application are discussed. This work is a significant step towards the ultimate goal of our project, which is the hybrid simulation of quantum algorithms, i.e., exploring multi-core CPUs and GPUs distributed along a cluster, achieving scalability when larger systems are simulated. |
Year | DOI | Venue |
---|---|---|
2014 | 10.1145/2554850.2554892 | SAC |
Keywords | Field | DocType |
gpu computing,distributed systems,applications,quantum process,quantum computing,distributed quantum simulation,other architecture styles | Unconventional computing,Quantum process,Computer science,Quantum computer,Symmetric multiprocessor system,Distributed algorithm,Quantum algorithm,General-purpose computing on graphics processing units,Scalability,Distributed computing | Conference |
Citations | PageRank | References |
4 | 0.55 | 5 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Anderson Avila | 1 | 7 | 3.05 |
Adriano Maron | 2 | 16 | 2.50 |
Renata Reiser | 3 | 72 | 15.53 |
Maurício L. Pilla | 4 | 21 | 11.50 |
Adenauer C. Yamin | 5 | 62 | 20.67 |