LLVM / Clang Social Meetup - Toronto Area

We're pleased to collaborating with the University of Toronto Computer Science Student Union on hosting the next Toronto LLVM Meetup! Please join us for a pair of technical talks, followed by socializing with fellow LLVM enthusiasts!

Location
Bahen Centre of Information Technology
Room BA3200

Agenda
6:30 - 7:00 pm Doors open, socializing
7:00 – 7:30 pm Presentation "Getting started with LLVM (it’s not that scary!)" - by Ayokunle Amodu
7:30 - 8:00 pm Presentation "From OpenMP to Offload: Lowering Safe Rust to GPUs" - by Manuel Drehwald
8:00 - 8:30 pm pizza and additional socializing

Presentations:

Getting started with LLVM (it’s not that scary!)
LLVM can look intimidating to newcomers: a large codebase, many subprojects, and reviewers who seem to know everything. This talk walks through how one can go from a first patch to becoming a regular contributor, drawing on the speaker’s own path as a recent upstream contributor. Key topics include choosing a subproject that matches your interests, working in small incremental patches, and using AI tools with discipline, without losing sight of what you actually understand.

From OpenMP to Offload: Lowering Safe Rust to GPUs
Rust is a safe language which can prevent data races and memory-safety bugs at compile time. This makes it very easy to write parallel Rust code for CPUs, but once we start looking at how to use GPUs, things become a lot more complex. In this talk, we will see how we mapped safe Rust GPU code to LLVM's libomptarget, where its functionality is limiting us, and what we hope to gain from moving to LLVM's Offload library in the future.

Speakers

Ayokunle Amodu
Ayokunle Amodu is a recent graduate and compiler engineer based in Toronto. He has been contributing to LLVM since last year, with a particular enthusiasm for MLIR. He recently completed a fellowship with the Compiler Research Group, where his work addressed a long-standing scalability issue in Clang modules. He holds a BSc in Computing Science from the University of Alberta, where his journey began in the Compiler Design and Optimization Lab.

Manuel Drehwald
Manuel Drehwald is a PhD student at the University of Toronto, a Rust compiler team member, and currently an intern at NVIDIA Research. He is interested in making Rust more usable for projects in the areas of HPC, Scientific Computing, and ML. As part of that goal, he has been leading the efforts to integrate LLVM based automatic differentiation and GPU offload into the Rust compiler.