Mini-Workshop on Differentiable Programming for High-Performance, Data-Intensive Computations [7 Apr at 1700CET/0800PDT.]
Passing this on for anyone that may be interested...
---------- Forwarded message ---------
From: Vassil Vassilev <vvasilev@...>
Date: Thu, Apr 1, 2021 at 11:09 AM
Subject: Mini-Workshop on Differentiable Programming for High-Performance, Data-Intensive Computations [7 Apr at 1700CET/0800PDT.]
To: <rootdev@...>, cling-dev <cling-dev@...>
We would like to invite you to the first mini-workshop about
"Differentiable Programming for
High-Performance, Data-Intensive Computations" starting on
7th of April at 1700 CET/0800 PDT. You can join directly by
clicking on the zoom link.
We will have 4 talks focusing on automatic differentiation (AD) tools and approaches to improve the performance of algorithms by orders of magnitude. William Moses, MIT, will talk about Enzyme -- a low-level post-optimization AD implementation with LLVM; Angel Angelov, ChaosGroup, will talk about the benefits of AD in the computer graphics engines and shading languages in particular; Vassil Vassilev, Princeton, will talk about Clad -- a C++ language-level AD implementation, CUDA, and its integration with ROOT; Benjamin Huth, Universität Regensburg, will talk about using AD based on autodiff in the context of common track reconstruction software ACTS.
Please feel free to
forward the invitation to everybody who might be interested.
David & Vassil