Dr Philip Blakely

Associate Teaching Professor

Philip Blakely studied for a BA, MMath, and PhD in Mathematics at the University of Cambridge. He received his PhD from the University of Cambridge in 2010, in Numerical Relativity, under the supervision of Professor Nikolaos Nikiforakis.

He became a post-doc in the Laboratory for Scientific Computing in 2009, working on Adaptive Mesh Refinement techniques and interface-capturing methods.

In 2023 he was appointed Assistant Teaching Professor where he coordinates HPC teaching for the MPhil in Scientific Computing, and supervises students for the research projects, as well as lecturing in Research Computing, C++, and CUDA, and is Academy Director for the HPC Autumn Academy.

Research

Dr Blakely’s PhD thesis concerned Bondi-Hoyle-Lyttleton accretion onto black holes. He developed a fully three-dimensional code, based on the Overture library [http://www.overtureframework.org/], which allowed him to investigate the effects of the spin of the black hole, as well as the effect of a non-uniform fluid density upstream of the black hole.

Since then, Philip has worked on developing an AMR framework which has been used in research within LSC. He has worked directly with industrial partners and sponsors, including Orica Mining Services and Quaise Energy, as well as supporting LSC researchers working with other industrial partners.

For Orica Mining Services, Philip developed software to model discrete fractured particles and their motion under the effective of explosives. This helps to predict the final fractured rock distribution, and hence optimize for the energy required to subsequently break down the rock.

For Quaise Energy, he has developed code to accurately model low-speed gas flow through heated rocks, aiming to optimize the hardware configuration required to evaporate rock using Millimetre wave beams, to extract geothermal energy from rock.

Philip’s current research interests are the efficient implementation of various fundamental numerical methods on a range of computational hardware. In particular, he is working on parallelising AMR algorithms, in conjunction with multi-material evolution based on Ghost-Fluid methods. He is also interested in implicit-explicit (IMEX) schemes, and comparison of efficiency and accuracy of a range of multiphysics numerical methods. In particular, he is interested in C++ and CUDA and how they can be used most effectively in developing efficient, robust, and maintainable large-scale simulation codes.