Publications

Links to my published journal and conference papers, preprints, dissertations, and theses.

Peer-reviewed publications

MATLAB Simulator of Level-Index Arithmetic. Proc. 31st IEEE Symp. Comput. Arithmetic. Málaga, Spain. Jul. 2024.

Monotonicity of Multi-Term Floating-Point Adders. IEEE Trans. Comput., 73:6. Feb. 2024.

CPFloat: A C library for emulating low-precision arithmetic (with M. Fasi). ACM Trans. Math. Software, 49:2. Jun. 2023. Blog post on this research.

Matrix Multiplication in Multiword Arithmetic: Error Analysis and Application to GPU Tensor Cores (with M. Fasi, N. J. Higham, F. Lopez, and T. Mary). SIAM J. Sci. Comput., 45:1. Feb. 2023.

Stochastic rounding: implementation, error analysis, and applications (with M. Croci, M. Fasi, N. J. Higham, and T. Mary). R. Soc. Open Sci., 9:3. Mar. 2022. Science News article on this research.

Anymatrix: An Extensible MATLAB Matrix Collection (with N. J. Higham). Numer. Algorithms., 90:3. Dec. 2021. Blog post on this research.

Stochastic Rounding: Algorithms and Hardware Accelerator. Proc. 2021 Int. Joint Conf. Neural Networks (IJCNN). Shenzhen, China. Sep. 2021.

Algorithms for Stochastically Rounded Elementary Arithmetic Operations in IEEE 754 Floating-Point Arithmetic (with M. Fasi). IEEE Trans. Emerg. Topics Comput., 9:3. Mar. 2021 (Special Section on “Emerging and Impacting Trends on Computer Arithmetic”). Blog post 1, 2 on this research.

Numerical Behavior of NVIDIA Tensor Cores (with M.Fasi, N. J. Higham, and S. Pranesh). PeerJ Comput. Sci., 7:e330. Feb. 2021. Blog post on this research.

Issues with rounding in the GCC implementation of the ISO 18037:2008 standard fixed-point arithmetic. Proc. 27th IEEE Symp. Comput. Arithmetic. Portland, OR, US. Aug. 2020. Blog post on this research.

Stochastic rounding and reduced-precision fixed-point arithmetic for solving neural ordinary differential equations (with M. Hopkins, D. R. Lester, and S. Furber). Phil. Trans. R. Soc. A, 378:2166. Jan. 2020.

sPyNNaker: A Software Package for Running PyNN Simulations on SpiNNaker (with O. Rhodes, P. A. Bogdan, C. Brenninkmeijer, S. Davidson, D. Fellows, A. Gait, D. R. Lester, L. Plana, A. G. D. Rowley, A. B. Stokes, and S. B. Furber.). Front. Neurosci., 12:816. Nov. 2018.

Approximate Fixed-Point Elementary Function Accelerator for the SpiNNaker-2 Neuromorphic Chip (with D. R. Lester, D. Shang, S. B. Furber, G. Liu, J. Garside, S. Scholze, S. Höppner, and A. Dixius). Proc. 25th IEEE Symp. Comput. Arithmetic. Amherst, MA, US. Sep. 2018.

Neuromodulated Synaptic Plasticity on the SpiNNaker Neuromorphic System (with G. P. García, J. C. Knight, and S. B. Furber). Front. Neurosci., 12:105. Feb. 2018.

Brewing the first ever automatic memory management utility for SpiNNaker: Real-time garbage collection for STDP simulations (with D. R. Lester). 2017 Int. Joint Conf. Neural Networks (IJCNN). Anchorage, Alaska, US. 14-19 May 2017.

Peer-reviewed abstracts

Anymatrix: an extensible MATLAB matrix collection (with N. J. Higham). Extended abstract. 7th IMA Conference on Numerical Linear Algebra and Optimization. Jun. 2022.

A Trick for an Accurate e^−|x| Function in Fixed-Point Arithmetics. Extended abstract. 19th Int. Symp. Scientific Computing, Comput. Arithmetic, Verified Numer. Computations (SCAN2020). Sep. 2021.

CPFloat: A C library for emulating low-precision arithmetic (with M. Fasi). Extended abstract. 19th Int. Symp. Scientific Computing, Comput. Arithmetic, Verified Numer. Computations (SCAN2020). Sep. 2021.

Chapters

SpiNNaker: A Spiking Neural Network Architecture (S. Furber, and P. Bogdan (editors)). Boston-Delft: Now Publishers. Mar. 2020 (M. Mikaitis contributed to Chapters 4 and 8).

Preprints or technical reports

Error analysis of matrix multiplication with narrow range floating-point arithmetic (with Theo Mary). hal-04671474. Aug. 2024.

Probabilistic error analysis of limited-precision stochastic rounding (with El-Mehdi El Arar, Massimiliano Fasi, and Silviu Ioan-Filip). hal-04665809. Jul. 2024.

List of potential typographical issues in IEEE Standard for Floating-Point Arithmetic (IEEE 754-2019). Note. Mar. 2024.

The SpiNNaker 2 Processing Element Architecture for Hybrid Digital Neuromorphic Computing (with S. Höppner, Y. Yan, B. Vogginger, C. Liu, F. Kelber, A. Dixius, S. Scholze, J. Partzsch, M. Stolba, F. Neumärker, G. Ellguth, S. Hartmann, S. Schiefer, T. Hocker, D. Walter, G. Liu, J. Garside, S. Furber, C. Mayr). arXiv:2103.08392 [cs.AR]. Aug. 2022.

Anymatrix: An Extensible MATLAB Matrix Collection. Users' Guide (with N. J. Higham). Tech. Report 2021.15, Manchester Institute for Mathematical Sciences, The University of Manchester, UK. Oct. 2021.

Dissertations or theses

Arithmetic Accelerators for a Digital Neuromorphic Processor. PhD Computer Science, University of Manchester. Jul. 2020.

Application-specific Real-Time garbage collection for neural simulations on SpiNNaker. BSc Computer Science, University of Manchester. May 2016.