Skip to content

Latest commit

 

History

History
85 lines (65 loc) · 5 KB

2023.md

File metadata and controls

85 lines (65 loc) · 5 KB
layout title header subheader registration schedule calendar mapZoom mapCenter mapPin conference permalink
home
Enzyme Conference 2023
February 22-24, 2023 University of Colorado Boulder, CO, USA
February 22-24, 2023 (before PPoPP/CGO/CC/HPCA/SIAM CSE) <br> University of Colorado Boulder, CO, USA
12
-105.26145734222993, 40.0098165377417
-105.26282099999537, 40.00681043602383
EnzymeCon23
/2023/

Attendance

The registration fee is $160, which allows us to cover organization, student travel, lunch and coffee. The registration is now open.

Student travel grants are available, please see below.

Program

Full schedule with abstracts and author names is available here.

Wednesday, February 22

  • 12:15pm — Registration Starts
  • 1:15pm — Opening
  • 1:30pm — Session 1: Enzyme Applications
    • PSDR-Enzyme: Physics-Based Differentiable Rendering Using Enzyme
    • Automatic Differentiation in Solid Mechanics: Interpretation and Composition [slides]
    • Differentiating Large-Scale Finite Element Applications with MFEM
    • A Cross-Language Probabilistic Programming Protocol for Physics and Beyond
    • Numba-Enzyme: A Differentiable JIT-ed Python
  • 3:30pm — Cofee Break
  • 4:00pm — Enzyme Tutorial
  • 5:00pm — Hackathon and Spontaneous Roundtables

Thursday, February 23

  • 9:00am — Session 2: Lessons from AD Tools
    • A survey of Tapenade in contrast with Enzyme
    • NIFTy: The Why and How of Building AD from Scratch
    • AD with Integrals
  • 10:30am — Coffee Break
  • 11:00am — Roundtables
    • DJ4Earth: Oceans, ice sheets, adjoints, and AD
  • 12:00pm — Lunch
  • 1:30pm — Session 3: Enzyme in Modern Programming Languages
    • Compilation Augmentation Enables High-Performance Batch Differentiation
    • Differentiable molecular simulation with Molly.jl [slides]
    • Adjoint Checkpointing using Custom Differentiation Rules
    • Oxide-Enzyme - Integrating Enzyme into the Rust compiler
  • 3:30pm — Coffee Break
  • 4:00pm — Session 4: Enzyme Extensions
    • Hackable Autodiff: Extending Enzyme to MLIR for Reverse Mode Gradients
    • Autodiff semantics and the level of abstraction [slides]
    • Accelerating Black Hole Imaging with Enzyme
  • 5:30pm — Hackathon and Spontaneous Roundtables

Friday, February 24

  • 8:30am — Hackathon day

Description

Automatic differentiation is an established field that is seeing renewed interest thanks to its practical applications to machine learning, simulation for natural sciences, and the emerging differentiable programming paradigm. The Enzyme project brought automatic differentiation to the LLVM compiler infrastructure broadly used for both research and production, and has been successfully used to enable automatic differentiation across multiple programming languages (C++, Julia, Rust, etc.) and target architectures. This workshop aims at bringing together the contributors and users of Enzyme in particular, and people interested in automatic differentiation within the LLVM infrastructure in general, for two days of fast-paced high-bandwidth idea exchange and collaboration bootstrapping. The workshop format remains flexible and friendly to practitioners and researchers alike; we expect a mix of presentations, tutorials, roundtable discussions and live programming sessions.

Logistics

Denver airport (DEN) is the nearest airport to the venue. There is a bus that runs hourly between the airport and Boulder (Route AB1 bus) which takes roughly 90 minutes. Alternatively, the airport is a 30-60 minute drive to Boulder.

Several hotels are a short walk from the CU Boulder campus. The cheapest convenient lodging is the Millenium for roughly $100/night. Similarly nearby are also the Boulder Marriott and Hilton Boulder, at roughly $200/night.

Venue

EnzymeCon will take place at CU Boulder in room ECNT 312 in the Engineering Center North Tower.

Student Travel Grants

The Enzyme conference is committed to helping students attend and is pleased to offer limited funds to offset the cost of attendance. Travel grants will be made out of available funds and may not be able to cover all needs. Please submit your application through this Google Form.