IRIS-HEP Fellow: Manasvi Goyal
Fellowship dates: May – Jul, 2022
Home Institution: Delhi Technological University
Project: Accelerating Awkward Array BuildersAwkward Array is a popular library for nested, variable-sized data, including arbitrary-length lists, records, mixed types, and missing data, to manipulate JSON-like data using NumPy-like idioms. This project concentrates on improving the performance of the builders of Awkward Arrays by exploiting different techniques including Just-in-time compilation (JIT) to improve the performance of interpreted programs. Run-time generation enables detailed specialization for higher performance, lazy loading for reduced memory use in large scale projects. The tasks include making the LayoutBuilder take advantage of a JIT compiler to become as fast as specialised output and to optimise GrowableBuffer and Forth's OutputBuffers. This project will prove to be useful in preventing memory leaks, optimised allocation of memory and providing a significant gain in speeds.
More information: My project proposal
Jim Pivarski (Princeton University)
Ianna Osborne (CERN)