Doug
May the lightning talks be with you
Host:
MongoDB HQ
Date:
Wed May 29, 2019
Nace Partners: pizza, beer, and drinks.
🎤 Raw SIMD in Rust, some simple examples and benchmarks
Rust has had stable support for many SIMD intrinsics on x86 since version 1.27, with support for further instruction set extensions (AVX512) and other architectures (ARM NEON) coming in the future. We'll look at some simple examples of how to use these intrinsics, and benchmarks to see how much they speed up our code. We'll also use godbolt.org to look at how different implementations generate different assembly code.
Jack has been working on side projects in Rust for several years, particularly the Bao project (https://github.com/oconnor663/bao). Previously he was an engineer at Keybase.
🎤 Tower and Friends: Lego(s) for distributed systems
Tower provides a simple, but powerful paradigm for building (asynchronous) services. Tower makes it easy to create building blocks which can be reused across an application, providing functionality such as request retrying, buffering, and load balancing.
I'll give a brief introduction to the Service trait API, useful existing libraries, and give some example how we use it in building our distributed app at Oso.
Sam is CTO and cofounder of Oso (osohq.com), where he is trying to make security less terrible for dev/ops teams. Sam has been using Rust since about the time it turned 1.0, and fell in love with it while building proof of concepts for various research papers during his PhD.
🎤 Datalog and Rust
Brief description of the talk Datalog is a small logic programming language, often useful for expressing program analyses. In particular, Rust's next-gen borrow checker, Polonius, is framed as a Datalog query over facts produced by the Rust compiler about a program in question. At the same time, several interesting next-gen Datalog engines are implemented in Rust. There is the potential for great synergy!
I'll briefly introduce Datalog, sketch how parts of Rust's borrow checker can be encoded as Datalog, and then name-check some Datalog engines written in Rust and their neat features.
Frank is the chief scientist at and a cofounder of Materialize, Inc., a NYC start-up doing interactive realtime SQL in Rust. He has been using Rust since 2014, when his C#-based gig went away, and has written the datafrog lightweight Datalog .. "engine", let's say, among several other more substantial projects.