Dr. Dipl.-Ing. Raphael Mosaner, BSc.

Oracle Labs
Altenberger Straße 69
4040 Linz, Austria

Building: Computer Science Building (Science Park 3) Room: 203
Phone: + 43-732-2468-4347
Fax: + 43-732-2468-4345
Email: raphael.m.mosaner@oracle.com

Research

In my PhD, I am researching how machine learning can be employed to improve compiler optimizations.

Research Interests

Interpreters, compilers and virtual machines
Machine learning
Compiler optimizations

Publications

Mosaner, R.; Barany, G.; Leopoldseder, D.; Mössenböck, H.:
Improving Vectorization Heuristics in a Dynamic Compiler with Machine Learning Models.
Workshop on Virtual Machines and Language Implementations (VMIL'22), at SPLASH'22, Dec.5-10, 2022, Auckland, New Zealand, pp.36-47
Mosaner, R.; Leopoldseder, D.; Kisling, W.; Stadler, L.; Mössenböck, H.:
Machine-Learning-Based Self-optimizing Compiler Heuristics.
19th Intl. Conf. on Managed Programming Languages and Runtimes (MPLR'22), Brussels, Sept. 14-15, 2022, pp.98-111
Mosaner, R.; Leopoldseder, D.; Kisling, W.; Stadler, L.; Mössenböck, H.:
Compilation Forking: A Fast and Flexible Way of Generating Data for Compiler-Internal Machine Learning Tasks.
The Programming Journal, vol. 7, issue 1, article 3, June 2022, pp.1-29
Mosaner, R.; Leopoldseder, D.; Stadler, L.; Mössenböck, H.:
Using Machine Learning to Predict the Code Size Impact of Duplication Heuristics in a Dynamic Compiler.
WIP paper, 18th Intl. Conf. on Managed Programming Languages and Runtimes (MPLR'21), Sept. 29-30, 2021, Münster, Germany, pp.127-135.
Mosaner, R.:
Machine Learning to Ease Understanding of Data Driven Compiler Optimizations.
Doctoral Symposium at SPLASH'20, November 15-20, 2020, Online, pp.4-6
Mosaner, R.; Leopoldseder, D.; Rigger, M.; Schatz, R.; Mössenböck, H.:
Supporting On-Stack Replacement in Unstructured Languages by Loop Reconstruction and Extraction.
Intl. Conf. on Managed Programming Languages and Runtimes (MPLR'19), 20-25 October, 2019, Athens, Greece.

Talks

Apart from presenting publications at conferences, I have given talks on my work on other occasions as well.

GraalCGO 2021: Improving Compiler Optimizations by Employing Machine Learning
VMM 2020: Let Data Drive Compiler Optimizations

Projects

These are the projects I am currently working on or worked on in the past.

GraalVM, a universal virtual machine for running applications written in different programming languages.
Sulong a high-performance LLVM bitcode runtime built on the GraalVM by Oracle Labs.

Awards

Winner of the SPLASH 2020 Student Research Competition in the Graduate category
Würdigungspreis des Bundesministeriums für Bildung, Wissenschaft und Forschung für die besten Diplom- und Masterabschlüsse 2019

Teaching

These are the courses that I am currently giving or have given in the past.