Optimize Automatic Differentiation Performance in C++ - Steve Bronder - CppCon 2025 @CppCon
Optimize Automatic Differentiation Performance in C++ - Steve Bronder - CppCon 2025  @CppCon
Uploaded December 2025 | Updated September 2026, 3 weeks ago
cppcon.org
---

Optimize Automatic Differentiation Performance in C++ - Steve Bronder - CppCon 2025
---

Reverse‑mode automatic differentiation (AD) powers everything from back‑propagation that trains trillion‑parameter large language models to the Stan programming language's Bayesian inference engines. Performance tricks like arena allocators, expression‑templates, SIMD friendly data structures, and template meta-programming finds its way inside C++ AD libraries. Milliseconds saved per gradient compound can turn into hours of wall‑time wins.

This session dissects the engineering behind those performance wins, showcasing improvements across different C++ AD libraries over time. Attendees will see how contemporary C++ AD techniques can make AD so fast that there is rarely a need to add hand-written derivatives to your program. The talk assumes familiarity with modern C++ but no prior exposure to automatic differentiation.

---

Slides: github.com/CppCon/CppCon2025/blob/main/Presentations/From_Bayesian_Inference_to_LLMs.pdf

Work at Hudson River Trading (HRT): tinyurl.com/safxfctf
---

Steve Bronder

Steve Bronder is a software engineer at the Flatiron Institute in New York City who focuses on high‑performance C++ libraries for statistical modeling and automatic differentiation. He is a a core maintainer of the Stan Math project and contributes to open‑source tooling that accelerates scientific computing. Previously, he worked as a data scientist at Capital One. Steve holds an M.A. in Quantitative Methods in the Social Sciences from Columbia University. In his spare time he enjoys tabletop games and running.
---

CppCon is the annual, week-long face-to-face gathering for the entire C++ community. The conference is organized by the C++ community for the community. You will enjoy inspirational talks and a friendly atmosphere designed to help attendees learn from each other, meet interesting people, and generally have a stimulating experience. Taking place this year in Aurora, Colorado, near the Denver airport, and including multiple diverse tracks, the conference will appeal to anyone from C++ novices to experts.
Annual CppCon Conference - cppcon.org
linkedin.com/company/cppcon
https://x.com/cppcon
facebook.com/CppConference
reddit.com/r/cppcon
https://mastodon.social/@CppCon
---

Videos Filmed & Edited by Bash Films: BashFilms.com
YouTube Channel Managed by Digital Medium Ltd: events.digital-medium.co.uk
---
#cpp #cplusplus #cppcon #cppprogramming #cplusplusprogramming #softwaredevelopment #softwareengineering #coding #code #computerscience #technology #technews #programming #programmer
Optimize Automatic Differentiation Performance in C++ - Steve Bronder - CppCon 2025C++ Sender Patterns to Wrangle C++ Concurrency in Embedded Devices - Michael Caisse - CppCon 2024Debugging C++ Coroutines in GDB - Zartaj Majeed - CppCon 2025Back to Basics: Lifetime Management in Cpp - Phil Nash - CppCon 2024Lightning Talk: Chplx - Bridging Chapel and C++ for Enhanced Asynchronous Many-Task ProgrammingUsing Modern C++ to Build XOffsetDatastructure - Zero-Encoding & Zero-Decoding Serialization LibraryA Case-study in Rewriting a Legacy Gui Library for Real-time Audio Software in Modern C++ (Reprise)Upgrading Sea of Thieves From C++14 to C++20 Wasnt Easy Heres Why - Keith Stockdale - CppCon 2025Lightning Talk: Can `set_value` Actually Throw? - Robert Leahy - CppCon 2025Lightning Talk: Saturday Is Coming Faster - Convert year_month_day to Weekday Faster - Part 2C++ Data Structures That Make Video Games Go Round - Al-Afiq Yeong - CppCon 2024Whats New for C++ in VS Code: CMake Improvements and GitHub Copilot Agents - Alexandra Kemper
CppCon |

Optimize Automatic Differentiation Performance in C++ - Steve Bronder - CppCon 2025

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER