LLVM Code Optimisation for Automatic Differentiation: When Forward and Reverse Mode Lead in the Same Direction

Maximilian E. Schüle, Maximilian Springer, Alfons Kemper, Thomas Neumann

Publikation: Beitrag in Buch/Bericht/KonferenzbandKonferenzbeitragBegutachtung

2 Zitate (Scopus)

Abstract

Both forward and reverse mode automatic differentiation derive a model function as used for gradient descent automatically. Reverse mode calculates all derivatives in one run, whereas forward mode requires rerunning the algorithm with respect to every variable for which the derivative is needed. To allow for in-database machine learning, we have integrated automatic differentiation as an SQL operator inside the Umbra database system. To benchmark code-generation to GPU, we implement forward as well as reverse mode automatic differentiation. The inspection of the optimised LLVM code shows that nearly the same machine code is executed after the generated LLVM code has been optimised. Thus, both modes yield similar runtimes but different compilation times.

OriginalspracheEnglisch
TitelProceedings of the 6th Workshop on Data Management for End-To-End Machine Learning, DEEM 2022 - In conjunction with the 2022 ACM SIGMOD/PODS Conference
Herausgeber (Verlag)Association for Computing Machinery, Inc
ISBN (elektronisch)9781450393751
DOIs
PublikationsstatusVeröffentlicht - 12 Juni 2022
Veranstaltung6th Workshop on Data Management for End-To-End Machine Learning, DEEM 2022 - In conjunction with the 2022 ACM SIGMOD/PODS Conference - Virtual, Online, USA/Vereinigte Staaten
Dauer: 12 Juni 2022 → …

Publikationsreihe

NameProceedings of the 6th Workshop on Data Management for End-To-End Machine Learning, DEEM 2022 - In conjunction with the 2022 ACM SIGMOD/PODS Conference

Konferenz

Konferenz6th Workshop on Data Management for End-To-End Machine Learning, DEEM 2022 - In conjunction with the 2022 ACM SIGMOD/PODS Conference
Land/GebietUSA/Vereinigte Staaten
OrtVirtual, Online
Zeitraum12/06/22 → …

Fingerprint

Untersuchen Sie die Forschungsthemen von „LLVM Code Optimisation for Automatic Differentiation: When Forward and Reverse Mode Lead in the Same Direction“. Zusammen bilden sie einen einzigartigen Fingerprint.

Dieses zitieren