Ravi Reddy Manumachu Assistant Professor · School of Computer Science · University College Dublin

I earned my B.Tech from the Indian Institute of Technology (I.I.T), Madras (1997) and my PhD in Computer Science (high-performance heterogeneous computing) at UCD (2005). My research centres on heterogeneous parallel computing and energy-efficient computing.

I develop models, algorithms, and tools to optimise applications for performance and energy on extreme-scale heterogeneous platforms — cloud, grid, and supercomputers.

I am co‑author of functional performance & energy models, a theoretical framework for energy predictive models, and analyses of energy proportionality in modern server processors. I have also built accurate linear energy predictive models, model‑based methods, and data‑partitioning algorithms for bi‑objective optimisation of performance and energy on heterogeneous platforms.

I am the author of Heterogeneous MPI (an extension of MPI for heterogeneous clusters) and Heterogeneous ScaLAPACK (a linear algebra package for heterogeneous clusters). My work appears in top journals and conferences on parallel & distributed computing and energy science.

I direct the M.Sc. in Computer Science through Negotiated Learning (ucdmscnl.com) and serve as Postdoctoral Coordinator in the School of Computer Science. I am also a member of the School’s Equality, Diversity, and Inclusion (EDI) committee, contributing to the ATHENA SWAN action plans.

With 11+ years of industrial experience at IONA Technologies, Ansys, and Siemens Research, I have worked on distributed middleware, high‑performance web services, parallel CFD, and IBM mainframe processing.

Academic Profiles
Ph.D Thesis
HMPI: A Message‑Passing Library for Heterogeneous Networks of Computers
School of Computer Science, University College Dublin, Dublin, pp. 456, 06/2005
Software & Toolkits
OpenH Portable parallel applications on heterogeneous hybrid servers (2024) OpenH
libedm Energy modelling of data movement & communications (2023) libedm
EPPACK Energy proportionality toolkit for servers & data centers (2022) EPPACK
PAREPOPT Parallel algorithms for energy & performance optimisation (2021) PAREPOPT
HCLLIMB Model‑based performance optimisation for multicore (2018) HCLLIMB
LIBHCLOOC Out‑of‑core accelerator kernels on hybrid platforms (2018) LIBHCLOOC
PARALEPH Parallel algorithms for energy & performance (homogeneous clusters) (2017) PARALEPH
HCLWattsUp System‑level energy measurements using power meters (2016) HCLWattsUp
HeteroScaLAPACK Linear algebra library for heterogeneous networks (2010) HeteroScaLAPACK
HeteroMPI MPI extension for high‑performance heterogeneous computing (2004) HeteroMPI