4 Search Results for "Liu, Hai"

Document

DOI: 10.4230/LIPIcs.SEA.2024.27

Improved Cut Strategy for Tensor Network Contraction Orders

Authors: Christoph Staudt, Mark Blacher, Julien Klaus, Farin Lippmann, and Joachim Giesen

Published in: LIPIcs, Volume 301, 22nd International Symposium on Experimental Algorithms (SEA 2024)

Abstract

In the field of quantum computing, simulating quantum systems on classical computers is crucial. Tensor networks are fundamental in simulating quantum systems. A tensor network is a collection of tensors, that need to be contracted into a result tensor. Tensor contraction is a generalization of matrix multiplication to higher order tensors. The contractions can be performed in different orders, and the order has a significant impact on the number of floating point operations (flops) needed to get the result tensor. It is known that finding an optimal contraction order is NP-hard. The current state-of-the-art approach for finding efficient contraction orders is to combinine graph partitioning with a greedy strategy. Although heavily used in practice, the current approach ignores so-called free indices, chooses node weights without regarding previous computations, and requires numerous hyperparameters that need to be tuned at runtime. In this paper, we address these shortcomings by developing a novel graph cut strategy. The proposed modifications yield contraction orders that significantly reduce the number of flops in the tensor contractions compared to the current state of the art. Moreover, by removing the need for hyperparameter tuning at runtime, our approach converges to an efficient solution faster, which reduces the required optimization time by at least an order of magnitude.

Cite as

Christoph Staudt, Mark Blacher, Julien Klaus, Farin Lippmann, and Joachim Giesen. Improved Cut Strategy for Tensor Network Contraction Orders. In 22nd International Symposium on Experimental Algorithms (SEA 2024). Leibniz International Proceedings in Informatics (LIPIcs), Volume 301, pp. 27:1-27:19, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2024)

Copy BibTex To Clipboard

@InProceedings{staudt_et_al:LIPIcs.SEA.2024.27,
  author =	{Staudt, Christoph and Blacher, Mark and Klaus, Julien and Lippmann, Farin and Giesen, Joachim},
  title =	{{Improved Cut Strategy for Tensor Network Contraction Orders}},
  booktitle =	{22nd International Symposium on Experimental Algorithms (SEA 2024)},
  pages =	{27:1--27:19},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-325-6},
  ISSN =	{1868-8969},
  year =	{2024},
  volume =	{301},
  editor =	{Liberti, Leo},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.SEA.2024.27},
  URN =		{urn:nbn:de:0030-drops-203924},
  doi =		{10.4230/LIPIcs.SEA.2024.27},
  annote =	{Keywords: tensor network, contraction order, graph partitioniong, quantum simulation}
}

Document

DOI: 10.4230/LIPIcs.ECRTS.2024.1

JuMP2start: Time-Aware Stop-Start Technology for a Software-Defined Vehicle System

Authors: Anam Farrukh and Richard West

Published in: LIPIcs, Volume 298, 36th Euromicro Conference on Real-Time Systems (ECRTS 2024)

Abstract

Software-defined vehicle (SDV) systems replace traditional ECU architectures with software tasks running on centralized multicore processors in automotive-grade PCs. However, PC boot delays to cold-start an integrated vehicle management system (VMS) are problematic for time-critical functions, which must process sensor and actuator data within specific time bounds. To tackle this challenge, we present JuMP2start: a time-aware multicore stop-start approach for SDVs. JuMP2start leverages PC-class suspend-to-RAM techniques to capture a system snapshot when the vehicle is stopped. Upon restart, critical services are resumed-from-RAM within order of milliseconds compared to normal cold-start times. This work showcases how JuMP2start manages global suspension and resumption mechanisms for a state-of-the-art dual-domain vehicle management system comprising real-time OS (RTOS) and Linux SMP guests. JuMP2start models automotive tasks as continuable or restartable to ensure timing- and safety-critical function pipelines are reactively resumed with low latency, while discarding stale task state. Experiments with the VMS show that critical CAN traffic processing resumes within 500 milliseconds of waking the RTOS guest, and reaches steady-state throughput in under 7ms.

Cite as

Anam Farrukh and Richard West. JuMP2start: Time-Aware Stop-Start Technology for a Software-Defined Vehicle System. In 36th Euromicro Conference on Real-Time Systems (ECRTS 2024). Leibniz International Proceedings in Informatics (LIPIcs), Volume 298, pp. 1:1-1:27, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2024)

Copy BibTex To Clipboard

@InProceedings{farrukh_et_al:LIPIcs.ECRTS.2024.1,
  author =	{Farrukh, Anam and West, Richard},
  title =	{{JuMP2start: Time-Aware Stop-Start Technology for a Software-Defined Vehicle System}},
  booktitle =	{36th Euromicro Conference on Real-Time Systems (ECRTS 2024)},
  pages =	{1:1--1:27},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-324-9},
  ISSN =	{1868-8969},
  year =	{2024},
  volume =	{298},
  editor =	{Pellizzoni, Rodolfo},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.ECRTS.2024.1},
  URN =		{urn:nbn:de:0030-drops-203046},
  doi =		{10.4230/LIPIcs.ECRTS.2024.1},
  annote =	{Keywords: Time-aware stop-start, Real-time power management, Suspend-to-RAM, Partitioning hypervisor, Vehicle management system, Vehicle-OS, Software-defined vehicles (SDV)}
}

Document

DOI: 10.4230/DARTS.3.2.7

Parallelizing Julia with a Non-Invasive DSL (Artifact)

Authors: Todd A. Anderson, Hai Liu, Lindsey Kuper, Ehsan Totoni, Jan Vitek, and Tatiana Shpeisman

Published in: DARTS, Volume 3, Issue 2, Special Issue of the 31st European Conference on Object-Oriented Programming (ECOOP 2017)

Abstract

This artifact is based on ParallelAccelerator, an embedded domain-specific language (DSL) and compiler for speeding up compute-intensive Julia programs. In particular, Julia code that makes heavy use of aggregate array operations is a good candidate for speeding up with ParallelAccelerator. ParallelAccelerator is a non-invasive DSL that makes as few changes to the host programming model as possible.

Cite as

Todd A. Anderson, Hai Liu, Lindsey Kuper, Ehsan Totoni, Jan Vitek, and Tatiana Shpeisman. Parallelizing Julia with a Non-Invasive DSL (Artifact). In Special Issue of the 31st European Conference on Object-Oriented Programming (ECOOP 2017). Dagstuhl Artifacts Series (DARTS), Volume 3, Issue 2, pp. 7:1-7:2, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2017)

Copy BibTex To Clipboard

@Article{anderson_et_al:DARTS.3.2.7,
  author =	{Anderson, Todd A. and Liu, Hai and Kuper, Lindsey and Totoni, Ehsan and Vitek, Jan and Shpeisman, Tatiana},
  title =	{{Parallelizing Julia with a Non-Invasive DSL (Artifact)}},
  pages =	{7:1--7:2},
  journal =	{Dagstuhl Artifacts Series},
  ISSN =	{2509-8195},
  year =	{2017},
  volume =	{3},
  number =	{2},
  editor =	{Anderson, Todd A. and Liu, Hai and Kuper, Lindsey and Totoni, Ehsan and Vitek, Jan and Shpeisman, Tatiana},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/DARTS.3.2.7},
  URN =		{urn:nbn:de:0030-drops-72888},
  doi =		{10.4230/DARTS.3.2.7},
  annote =	{Keywords: parallelism, scientific computing, domain-specific languages, Julia}
}

Document

DOI: 10.4230/LIPIcs.ECOOP.2017.4

Parallelizing Julia with a Non-Invasive DSL

Authors: Todd A. Anderson, Hai Liu, Lindsey Kuper, Ehsan Totoni, Jan Vitek, and Tatiana Shpeisman

Published in: LIPIcs, Volume 74, 31st European Conference on Object-Oriented Programming (ECOOP 2017)

Abstract

Computational scientists often prototype software using productivity languages that offer high-level programming abstractions. When higher performance is needed, they are obliged to rewrite their code in a lower-level efficiency language. Different solutions have been proposed to address this trade-off between productivity and efficiency. One promising approach is to create embedded domain-specific languages that sacrifice generality for productivity and performance, but practical experience with DSLs points to some road blocks preventing widespread adoption. This paper proposes a non-invasive domain-specific language that makes as few visible changes to the host programming model as possible. We present ParallelAccelerator, a library and compiler for high-level, high-performance scientific computing in Julia. ParallelAccelerator's programming model is aligned with existing Julia programming idioms. Our compiler exposes the implicit parallelism in high-level array-style programs and compiles them to fast, parallel native code. Programs can also run in "library-only" mode, letting users benefit from the full Julia environment and libraries. Our results show encouraging performance improvements with very few changes to source code required. In particular, few to no additional type annotations are necessary.

Cite as

Todd A. Anderson, Hai Liu, Lindsey Kuper, Ehsan Totoni, Jan Vitek, and Tatiana Shpeisman. Parallelizing Julia with a Non-Invasive DSL. In 31st European Conference on Object-Oriented Programming (ECOOP 2017). Leibniz International Proceedings in Informatics (LIPIcs), Volume 74, pp. 4:1-4:29, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2017)

Copy BibTex To Clipboard

@InProceedings{anderson_et_al:LIPIcs.ECOOP.2017.4,
  author =	{Anderson, Todd A. and Liu, Hai and Kuper, Lindsey and Totoni, Ehsan and Vitek, Jan and Shpeisman, Tatiana},
  title =	{{Parallelizing Julia with a Non-Invasive DSL}},
  booktitle =	{31st European Conference on Object-Oriented Programming (ECOOP 2017)},
  pages =	{4:1--4:29},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-035-4},
  ISSN =	{1868-8969},
  year =	{2017},
  volume =	{74},
  editor =	{M\"{u}ller, Peter},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.ECOOP.2017.4},
  URN =		{urn:nbn:de:0030-drops-72693},
  doi =		{10.4230/LIPIcs.ECOOP.2017.4},
  annote =	{Keywords: parallelism, scientific computing, domain-specific languages, Julia}
}

Refine by Author
2 Anderson, Todd A.
2 Kuper, Lindsey
2 Liu, Hai
2 Shpeisman, Tatiana
2 Totoni, Ehsan
Show More...

Refine by Classification
1 Applied computing → Physics
1 Computer systems organization → Embedded systems
1 Computer systems organization → Real-time system architecture
1 Mathematics of computing → Solvers
1 Theory of computation → Algorithm design techniques

Refine by Keyword
2 Julia
2 domain-specific languages
2 parallelism
2 scientific computing
1 Partitioning hypervisor
Show More...

Refine by Type
4 document

Refine by Publication Year
2 2017
2 2024