Repositories

Project maintainer and lead contributor. Developed CPU/GPU Ewald summation algorithms for the Stokes and Laplace kernels in different periodic regimes.

Project collaborator and top 10 contributor. Implemented functionality for mathematical special functions, automatic just-in-time (JIT) recompilation, CMake JIT compilation, building and installing on MacOS and AMD GPUs, documentation and continuous integration.

Implemented a new matrix-free MTTKRP kernel for dense tensors: MTTKRP_Dense_Perm_Kernel, the basis for the publication A Performance Portable Matrix Free Dense MTTKRP in GenTen.

An interactive debugger for performance portable Python HPC kernels. Helped to develop benchmarks and examples, collaborated on architecture design and features.