Supercharge Your Performance with the Optimus Cirrus Platform

Building 21 million lines of Scala and Java code on a clean machine in less than 2 minutes: can it be done? Yes! The Optimus Cirrus platform is capable of automatically caching and parallelizing tasks across thousands of compute nodes comprising over a petabyte of RAM.

Building 21 million lines of Scala and Java code on a clean machine in less than 2 minutes: can it be done? Yes! The Optimus Cirrus platform is capable of automatically caching and parallelizing tasks across thousands of compute nodes comprising over a petabyte of RAM, and is used in production every day to support a significant part of Morgan Stanley’s business. The platform also underpins its own build tool which is responsible for well-scoped, cacheable compilation in a distributed environment while keeping its developers happy and productive. In this talk we’ll show you how Optimus Cirrus derives its power from its foundations in Scala and functional programming principles. We’ll demonstrate how its build tool reduced our compile times from over 1 hour down to as little as a few seconds, and discuss compelling technical challenges we encountered that are interesting to any developer creating applications that need to perform well.

DARJA’S NOTEThis is Daniela Sfregola and Valerie Saunders Duncan’s talk, I am submitting it because they can’t use PaperCall. They sent me an email with all the details and I copied it for the Reviewers to go over.

Autors’ notes: This talk will be co-presented. If the talk is accepted, we will need to update the speakers’ bios.

Daniela is a Vice President in Morgan Stanley’s Risk Infrastructure division working on the Optimus Cirrus platform. She has 10+ years of experience in Software Engineering and she is the author of the book “Get Programming With Scala”, published by Manning.

Valerie has been a software developer at Morgan Stanley in Montreal and London for 6 years. She works on developer tools for a large Scala monorepo, with a focus on streamlining the code contribution processes.