Closed Shicheng-Guo closed 5 months ago
Hey @Shicheng-Guo,
Regeneron contributed an implementation of Regenie rewritten in Python using Pandas user defined functions. The performance is roughly equivalent in terms of core hours, but parallelizing with Spark allows the processing time to be significantly faster as you can leverage more than one node
Glow itself a broad framework that has applications beyond distributing the Regenie algorithm. To learn more please start here, https://glow.readthedocs.io/en/latest/introduction.html
Anyone can share the difference/performance between GLOW and REGENIE ? Thanks