Search Results

Documents authored by Brunthaler, Stefan

Document
DOI: 10.4230/LIPIcs.ECOOP.2024.6
Cross Module Quickening - The Curious Case of C Extensions

Authors: Felix Berlakovich and Stefan Brunthaler

Published in: LIPIcs, Volume 313, 38th European Conference on Object-Oriented Programming (ECOOP 2024)

Abstract
Dynamic programming languages such as Python offer expressive power and programmer productivity at the expense of performance. Although the topic of optimizing Python has received considerable attention over the years, a key obstacle remains elusive: C extensions. Time and again, optimized run-time environments, such as JIT compilers and optimizing interpreters, fall short of optimizing across C extensions, as they cannot reason about the native code hiding underneath. To bridge this gap, we present an analysis of C extensions for Python. The analysis data indicates that C extensions come in different varieties. One such variety is to merely speed up a single thing, such as reading a file and processing it directly in C. Another variety offers broad access through an API, resulting in a domain-specific language realized by function calls. While the former variety of C extensions offer little optimization potential for optimizing run-times, we find that the latter variety does offer considerable optimization potential. This optimization potential rests on dynamic locality that C extensions cannot readily tap. We introduce a new, interpreter-based optimization leveraging this untapped optimization potential called Cross-Module Quickening. The key idea is that C extensions can use an optimization interface to register highly-optimized operations on C extension-specific datatypes. A quickening interpreter uses these information to continuously specialize programs with C extensions. To quantify the attainable performance potential of going beyond C extensions, we demonstrate a concrete instantiation of Cross-Module Quickening for the CPython interpreter and the popular NumPy C extension. We evaluate our implementation with the NPBench benchmark suite and report performance improvements by a factor of up to 2.84.
Cite as

Felix Berlakovich and Stefan Brunthaler. Cross Module Quickening - The Curious Case of C Extensions. In 38th European Conference on Object-Oriented Programming (ECOOP 2024). Leibniz International Proceedings in Informatics (LIPIcs), Volume 313, pp. 6:1-6:29, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2024)

Copy BibTex To Clipboard

@InProceedings{berlakovich_et_al:LIPIcs.ECOOP.2024.6,
  author =	{Berlakovich, Felix and Brunthaler, Stefan},
  title =	{{Cross Module Quickening - The Curious Case of C Extensions}},
  booktitle =	{38th European Conference on Object-Oriented Programming (ECOOP 2024)},
  pages =	{6:1--6:29},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-341-6},
  ISSN =	{1868-8969},
  year =	{2024},
  volume =	{313},
  editor =	{Aldrich, Jonathan and Salvaneschi, Guido},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.ECOOP.2024.6},
  URN =		{urn:nbn:de:0030-drops-208557},
  doi =		{10.4230/LIPIcs.ECOOP.2024.6},
  annote =	{Keywords: interpreter, optimizations, C extensions, Python}
}
Document
DOI: 10.4230/LIPIcs.ECOOP.2024.7
HOBBIT: Hashed OBject Based InTegrity

Authors: Matthias Bernad and Stefan Brunthaler

Published in: LIPIcs, Volume 313, 38th European Conference on Object-Oriented Programming (ECOOP 2024)

Abstract
C vulnerabilities usually hold verbatim for C++ programs. The counterfeit-object-oriented programming attack demonstrated that this relation is asymmetric, i.e., it only applies to C++. The problem pinpointed by this COOP attack is that C++ does not validate the integrity of its objects. By injecting malicious objects with manipulated virtual function table pointers, attackers can hijack control-flow of programs. The software security community addressed the COOP-problem in the years following its discovery, but together with the emergence of transient-execution attacks, such as Spectre, researchers also shifted their attention. We present Hobbit, a software-only solution to prevent COOP attacks by validating object integrity for virtual function pointer tables. Hobbit does not require any hardware specific features, scales to multi-million lines of C++ source code, and our LLVM-based implementation offers a configurable performance impact between 121.63% and 2.80% on compute-intensive SPEC CPU C++ benchmarks. Hobbit’s security analysis indicates strong resistance to brute forcing attacks and demonstrates additional benefits of using execute-only memory.
Cite as

Matthias Bernad and Stefan Brunthaler. HOBBIT: Hashed OBject Based InTegrity. In 38th European Conference on Object-Oriented Programming (ECOOP 2024). Leibniz International Proceedings in Informatics (LIPIcs), Volume 313, pp. 7:1-7:25, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2024)

Copy BibTex To Clipboard

@InProceedings{bernad_et_al:LIPIcs.ECOOP.2024.7,
  author =	{Bernad, Matthias and Brunthaler, Stefan},
  title =	{{HOBBIT: Hashed OBject Based InTegrity}},
  booktitle =	{38th European Conference on Object-Oriented Programming (ECOOP 2024)},
  pages =	{7:1--7:25},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-341-6},
  ISSN =	{1868-8969},
  year =	{2024},
  volume =	{313},
  editor =	{Aldrich, Jonathan and Salvaneschi, Guido},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.ECOOP.2024.7},
  URN =		{urn:nbn:de:0030-drops-208566},
  doi =		{10.4230/LIPIcs.ECOOP.2024.7},
  annote =	{Keywords: software security, code-reuse attacks, language-based security, counterfeit-object-oriented programming, object integrity, compiler security}
}
Document
DOI: 10.4230/LIPIcs.ECOOP.2018.16
Accelerating Dynamically-Typed Languages on Heterogeneous Platforms Using Guards Optimization

Authors: Mohaned Qunaibit, Stefan Brunthaler, Yeoul Na, Stijn Volckaert, and Michael Franz

Published in: LIPIcs, Volume 109, 32nd European Conference on Object-Oriented Programming (ECOOP 2018)

Abstract
Scientific applications are ideal candidates for the "heterogeneous computing" paradigm, in which parts of a computation are "offloaded" to available accelerator hardware such as GPUs. However, when such applications are written in dynamic languages such as Python or R, as they increasingly are, things become less straightforward. The same flexibility that makes these languages so appealing to programmers also significantly complicates the problem of automatically and transparently partitioning a program's execution between a CPU and available accelerator hardware without having to rely on programmer annotations. A common way of handling the features of dynamic languages is by introducing speculation in conjunction with guards to ascertain the validity of assumptions made in the speculative computation. Unfortunately, a single guard violation during the execution of "offloaded" code may result in a huge performance penalty and necessitate the complete re-execution of the offloaded computation. In the case of dynamic languages, this problem is compounded by the fact that a full compiler analysis is not always possible ahead of time. This paper presents MegaGuards, a new approach for speculatively executing dynamic languages on heterogeneous platforms in a fully automatic and transparent manner. Our method translates each target loop into a single static region devoid of any dynamic type features. The dynamic parts are instead handled by a construct that we call a mega guard which checks all the speculative assumptions ahead of its corresponding static region. Notably, the advantage of MegaGuards is not limited to heterogeneous computing; because it removes guards from compute-intensive loops, the approach also improves sequential performance. We have implemented MegaGuards along with an automatic loop parallelization backend in ZipPy, a Python Virtual Machine. The results of a careful and detailed evaluation reveal very significant speedups of an order of magnitude on average with a maximum speedup of up to two orders of magnitudes when compared to the original ZipPy performance as a baseline. These results demonstrate the potential for applying heterogeneous computing to dynamic languages.
Cite as

Mohaned Qunaibit, Stefan Brunthaler, Yeoul Na, Stijn Volckaert, and Michael Franz. Accelerating Dynamically-Typed Languages on Heterogeneous Platforms Using Guards Optimization. In 32nd European Conference on Object-Oriented Programming (ECOOP 2018). Leibniz International Proceedings in Informatics (LIPIcs), Volume 109, pp. 16:1-16:29, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2018)

Copy BibTex To Clipboard

@InProceedings{qunaibit_et_al:LIPIcs.ECOOP.2018.16,
  author =	{Qunaibit, Mohaned and Brunthaler, Stefan and Na, Yeoul and Volckaert, Stijn and Franz, Michael},
  title =	{{Accelerating Dynamically-Typed Languages on Heterogeneous Platforms Using Guards Optimization}},
  booktitle =	{32nd European Conference on Object-Oriented Programming (ECOOP 2018)},
  pages =	{16:1--16:29},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-079-8},
  ISSN =	{1868-8969},
  year =	{2018},
  volume =	{109},
  editor =	{Millstein, Todd},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.ECOOP.2018.16},
  URN =		{urn:nbn:de:0030-drops-92217},
  doi =		{10.4230/LIPIcs.ECOOP.2018.16},
  annote =	{Keywords: Type Specialization, Guards Optimization, Automatic Heterogeneous Computing, Automatic Parallelism}
}
Questions / Remarks / Feedback
X

Feedback for Dagstuhl Publishing


Thanks for your feedback!

Feedback submitted

Could not send message

Please try again later or send an E-mail
© 2023-2024 Schloss Dagstuhl – LZI GmbH Imprint Privacy Contact