machines with convergence
play

Machines with Convergence Set Enumeration Youwei Zhuo , Jinglei - PowerPoint PPT Presentation

Jun 04, 2023 •261 likes •1.1k views

CSE: Parallel Finite State Machines with Convergence Set Enumeration Youwei Zhuo , Jinglei Cheng, Qinyi Luo, Jidong Zhai, Yanzhi Wang, Zhongzhi Luan, Xuehai Qian University of Southern California Tsinghua University Syracuse University

  1. CSE: Parallel Finite State Machines with Convergence Set Enumeration Youwei Zhuo , Jinglei Cheng, Qinyi Luo, Jidong Zhai, Yanzhi Wang, Zhongzhi Luan, Xuehai Qian University of Southern California Tsinghua University Syracuse University Beihang University

  2. Finite State Machine (FSM) • Network Intrusion Detection • Huffman Decoding • XML Parsing ALCHEM 2 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  3. “Embarrassingly Sequential” • State • Start state • Accept state S0 S0 • State transition *x / / /x * * * S0 S1 S2 S3 x x / ALCHEM 3 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  4. “Embarrassingly Sequential” • State • Start state / • Accept state S0 S0 S1 S1 • State transition *x / / /x * * * S0 S1 S2 S3 x x / ALCHEM 4 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  5. “Embarrassingly Sequential” • State • Start state / * • Accept state S0 S0 S1 S2 S1 S2 • State transition *x / / /x * * * S0 S1 S2 S3 x x / ALCHEM 5 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  6. “Embarrassingly Sequential” • State • Start state / * x • Accept state S0 S0 S1 S2 S1 S2 S2 S2 • State transition *x / / /x * * * S0 S1 S2 S3 x x / ALCHEM 6 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  7. “Embarrassingly Sequential” • State • Start state / * x x * * / • Accept state S0 S0 S1 S1 S2 S2 S2 S2 S2 S2 S3 S3 S3 S3 S0 S0 • State transition *x / / /x * * * S0 S1 S2 S3 x x / ALCHEM 7 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  8. Challenges of Parallel FSM • Divide Input into segments • Process segments in parallel? / * x x * * / S0 S0 ALCHEM 8 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  9. Challenges of Parallel FSM • Divide Input into segments • Process segments in parallel? / * x x * * / S0 S0 Same as sequential! ALCHEM 9 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  10. Challenges of Parallel FSM • Divide Input into segments • Process segments in parallel? / * x x * * / S0 S0 S1 S1 S2 S2 S2 S2 Same as sequential! ALCHEM 10 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  11. Challenges of Parallel FSM • Divide Input into segments • Process segments in parallel? / * x x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? Same as sequential! Unknown Start State? ALCHEM 11 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  12. Challenges of Parallel FSM • Divide Input into segments • Process segments in parallel? / * x x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? Same as sequential! Unknown Start State? ALCHEM 12 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  13. Enumeration • Enumerate every state as start state / * x x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? “Data -parallel finite- state machines.”, ASPLOS’14 ALCHEM 13 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  14. Enumeration • Enumerate every state as start state / * x x * * / S0 S0 S1 S2 S1 S2 S2 S2 ? ? ? ? ? S0 S0 S1 S1 S2 S2 S3 S3 “Data -parallel finite- state machines.”, ASPLOS’14 ALCHEM 14 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  15. Enumeration • Enumerate every state as start state / * x x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? S0 S0 S0 S0 S0 S0 S0 S0 S1 S1 S1 S1 S2 S2 S3 S3 S3 S3 S0 S0 S2 S2 S2 S2 S3 S3 S3 S3 S0 S0 S3 S3 S2 S2 S3 S3 S3 S3 S0 S0 “Data -parallel finite- state machines.”, ASPLOS’14 ALCHEM 15 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  16. Enumeration • Enumerate every state as start state / * x x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? S0 S0 S0 S0 S0 S0 S0 S0 S1 S1 S1 S1 S2 S2 S3 S3 S3 S3 S0 S0 S2 S2 S2 S2 S3 S3 S3 S3 S0 S0 S3 S3 S2 S2 S3 S3 S3 S3 S0 S0 Enumeration works! “Data -parallel finite- state machines.”, ASPLOS’14 ALCHEM 16 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  17. Enumeration • Enumerate every state as start state / * x x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? Overhead? S0 S0 S0 S0 S0 S0 S0 S0 S1 S1 S1 S1 S2 S2 S3 S3 S3 S3 S0 S0 S2 S2 S2 S2 S3 S3 S3 S3 S0 S0 S3 S3 S2 S2 S3 S3 S3 S3 S0 S0 Enumeration works! “Data -parallel finite- state machines.”, ASPLOS’14 ALCHEM 17 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  18. Enumeration • Enumerate every state as start state • Overhead=#State=4 / * x x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? Overhead? S0 S0 S0 S0 S0 S0 S0 S0 S1 S1 S1 S1 S2 S2 S3 S3 S3 S3 S0 S0 S2 S2 S2 S2 S3 S3 S3 S3 S0 S0 S3 S3 S2 S2 S3 S3 S3 S3 S0 S0 Enumeration works! “Data -parallel finite- state machines.”, ASPLOS’14 ALCHEM 18 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  19. Convergence • States transition to the same state * x / x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? S0 S0 S1 S1 S2 S2 S3 S3 ALCHEM 19 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  20. Convergence • States transition to the same state * x / x * * / S0 S0 S1 S2 S1 S2 S2 S2 ? ? ? ? ? S0 S0 S0 S0 S1 S1 S2 S2 S2 S2 S2 S2 S3 S3 S2 S2 ALCHEM 20 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  21. Convergence • States transition to the same state * x / x * * / S0 S0 S1 S2 S1 S2 S2 S2 ? ? ? ? ? S0 S0 S0 S0 S1 S1 S2 S2 S2 S2 S2 S2 S3 S3 S2 S2 ALCHEM 21 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  22. Convergence • States transition to the same state * x / x * * / S0 S0 S1 S2 S1 S2 S2 S2 ? ? ? ? ? S0 S0 S0 S0 S0 S0 S0 S0 S1 S1 S1 S1 S2 S2 S3 S3 S3 S3 S0 S0 S2 S2 S2 S2 S3 S3 S2 S2 ALCHEM 22 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  23. Convergence • States transition to the same state • # ReachableState in the current segment is non-increasing * x / x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? S0 S0 S0 S0 S0 S0 S0 S0 S1 S1 S1 S1 S2 S2 S3 S3 S3 S3 S0 S0 S2 S2 S2 S2 S3 S3 S2 S2 ALCHEM 23 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  24. Convergence • States transition to the same state • # ReachableState in the current segment is non-increasing • Overhead=#ReachableState=4 → 2 * x / x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? S0 S0 S0 S0 S0 S0 S0 S0 S1 S1 S1 S1 S2 S2 S3 S3 S3 S3 S0 S0 S2 S2 S2 S2 S3 S3 S2 S2 ALCHEM 24 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  25. Convergence • States transition to the same state • # ReachableState in the current segment is non-increasing • Overhead=#ReachableState=4 → 2 * x / x * * / S0 S0 S1 S1 S2 S2 S2 S2 ? ? ? ? ? S0 S0 S0 S0 S0 S0 S0 S0 S1 S1 S1 S1 S2 S2 S3 S3 S3 S3 S0 S0 S2 S2 S2 S2 Convergence works! S3 S3 S2 S2 ALCHEM 25 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  26. Speculation • Speculate one state as start state / * x x * * / S0 S0 S1 S2 S1 S2 S2 S2 ? ? ? ? ? S0 S0 S1 S1 S2 S2 S3 S3 Challenging the "embarrassingly sequential": parallelizing finite state machine- based computations through principled speculation. ASPLOS’14 ALCHEM 26 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  27. Speculation • Speculate one state as start state / * x x * * / S0 S0 S1 S2 S1 S2 S2 S2 ? ? ? ? ? S0 S0 S1 S1 S2 S2 S2 S2 S3 S3 S3 S3 S0 S0 S3 S3 Challenging the "embarrassingly sequential": parallelizing finite state machine- based computations through principled speculation. ASPLOS’14 ALCHEM 27 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

  28. Speculation • Speculate one state as start state / * x x * * / S0 S0 S1 S2 S1 S2 S2 S2 ? ? ? ? ? S0 S0 S1 S1 S2 S2 S2 S2 S3 S3 S3 S3 S0 S0 S3 S3 Challenging the "embarrassingly sequential": parallelizing finite state machine- based computations through principled speculation. ASPLOS’14 ALCHEM 28 CSE: Parallel Finite State Machines with Convergence Set Enumeration alchem.usc.edu

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

Some Thoughts on MC Convergence first, would like to define what I mean two kinds of

Some Thoughts on MC Convergence first, would like to define what I mean two kinds of

Some Thoughts on MC Convergence first, would like to define what I mean two kinds of convergence - convergence = experiments all working towards using same MC generator (common basis for comparison) - convergence =

239 views • 11 slides
Virtual Machines Uses for Virtual Machines There are several uses for virtual machines:

Virtual Machines Uses for Virtual Machines There are several uses for virtual machines:

Virtual Machines Uses for Virtual Machines There are several uses for virtual machines: Virtual machine technology, often just called virtualization , makes one computer behave as several Running several operating systems simultaneously on

254 views • 10 slides
CS 557 BGP Convergence Improved BGP Convergence via Ghost Flushing Bremler-Barr, Afek, Schwarz,

CS 557 BGP Convergence Improved BGP Convergence via Ghost Flushing Bremler-Barr, Afek, Schwarz,

CS 557 BGP Convergence Improved BGP Convergence via Ghost Flushing Bremler-Barr, Afek, Schwarz, 2003 BGP-RCN: Improving Convergence Through Root Cause Notification Pei, Azuma, Massey, Zhang, 2005 Spring 2013 BGP Path Exploration dest. ( ) Z

541 views • 31 slides
II of large Number Lattin in probability almost convergence convergence sure - - "

II of large Number Lattin in probability almost convergence convergence sure - - "

Chapters : Proofs of the laws II of large Number Lattin in probability almost convergence convergence sure - - " weak " strong " " cow corn Two tests for stray convergence - O almost sure conveyance . ) - X

339 views • 11 slides
19 FEM Convergence Requirements IFEM Ch 19 Slide 1 Introduction to FEM Convergence

19 FEM Convergence Requirements IFEM Ch 19 Slide 1 Introduction to FEM Convergence

Introduction to FEM 19 FEM Convergence Requirements IFEM Ch 19 Slide 1 Introduction to FEM Convergence Requirements for Finite Element Discretization Convergence: discrete (FEM) solution approaches the analytical (math model) solution

479 views • 17 slides
NSF Convergence Accelerator 1 Convergence Accelerator WHY: Leverage the science across all fields

NSF Convergence Accelerator 1 Convergence Accelerator WHY: Leverage the science across all fields

NSF Convergence Accelerator 1 Convergence Accelerator WHY: Leverage the science across all fields of NSF research to produce outcomes in an accelerated timeframe, with streamlined operations allowing for nimbleness to support the most innovative

327 views • 19 slides
HPC-AI CONVERGENCE Workshop on the Convergence of ML & HPC Gunter Roeth gunterr@nvidia.com

HPC-AI CONVERGENCE Workshop on the Convergence of ML & HPC Gunter Roeth gunterr@nvidia.com

NVIDIA TO ACCELERATE THE HPC-AI CONVERGENCE Workshop on the Convergence of ML & HPC Gunter Roeth gunterr@nvidia.com March 2020 GRAND CHALLENGES REQUIRE MASSIVE COMPUTING REINVENTING THE LI-ION BATTERY UNDERSTANDING HIVS STRUCTURE

693 views • 44 slides
NS NSF Convergence Accelerator Chaitan Baru Senior Science Advisor, Convergence Accelerator

NS NSF Convergence Accelerator Chaitan Baru Senior Science Advisor, Convergence Accelerator

NS NSF Convergence Accelerator Chaitan Baru Senior Science Advisor, Convergence Accelerator Office Office of the Director, NSF (on assignment from SDSC, UC San Diego) 1 NS NSF Bi Big Ideas 2 Con Convergence Research The grand challenges

477 views • 13 slides
Fixedmobile Convergence: Structural convergence Impact of radio coordinaTon on design of

Fixedmobile Convergence: Structural convergence Impact of radio coordinaTon on design of

Fixedmobile Convergence: Structural convergence Impact of radio coordinaTon on design of access and aggregaTon network Dirk Breuer Deutsche Telekom Laboratories, Germany Tibor Cinkler Budapest University of Technology and Economics,

244 views • 9 slides
Fixedmobile Convergence: Structural convergence MoSvaSon and technological triggers Dirk

Fixedmobile Convergence: Structural convergence MoSvaSon and technological triggers Dirk

Fixedmobile Convergence: Structural convergence MoSvaSon and technological triggers Dirk Breuer Deutsche Telekom Laboratories, Germany Tibor Cinkler Budapest University of Technology and Economics, Hungary Stphane Gosselin Orange

105 views • 8 slides
Fixedmobile convergence FuncRonal convergence Converged subscriber and session management

Fixedmobile convergence FuncRonal convergence Converged subscriber and session management

Fixedmobile convergence FuncRonal convergence Converged subscriber and session management Dirk Breuer Deutsche Telekom Laboratories, Germany Tibor Cinkler Budapest University of Technology and Economics, Hungary Stphane Gosselin Orange

212 views • 9 slides
Fixedmobile Convergence: Structural convergence Technology dimensioning and cost modelling

Fixedmobile Convergence: Structural convergence Technology dimensioning and cost modelling

Fixedmobile Convergence: Structural convergence Technology dimensioning and cost modelling Dirk Breuer Deutsche Telekom Laboratories, Germany Tibor Cinkler Budapest University of Technology and Economics, Hungary Stphane Gosselin

395 views • 5 slides
Fixedmobile convergence FuncRonal convergence Advanced interface selecRon and route control

Fixedmobile convergence FuncRonal convergence Advanced interface selecRon and route control

Fixedmobile convergence FuncRonal convergence Advanced interface selecRon and route control Dirk Breuer Deutsche Telekom Laboratories, Germany Tibor Cinkler Budapest University of Technology and Economics, Hungary Stphane Gosselin

279 views • 3 slides
Fixedmobile convergence IntroducSon The COMBO perspecSve: focus on network convergence Dirk

Fixedmobile convergence IntroducSon The COMBO perspecSve: focus on network convergence Dirk

Fixedmobile convergence IntroducSon The COMBO perspecSve: focus on network convergence Dirk Breuer Deutsche Telekom Laboratories, Germany Tibor Cinkler Budapest University of Technology and Economics, Hungary Stphane Gosselin Orange

335 views • 10 slides
Fixedmobile Convergence: Structural convergence Backhaul Dirk Breuer Deutsche Telekom

Fixedmobile Convergence: Structural convergence Backhaul Dirk Breuer Deutsche Telekom

Fixedmobile Convergence: Structural convergence Backhaul Dirk Breuer Deutsche Telekom Laboratories, Germany Tibor Cinkler Budapest University of Technology and Economics, Hungary Stphane Gosselin Orange R&D, France Annie Gravey

84 views • 5 slides
Fixedmobile convergence FuncRonal convergence Advanced interface selecRon and route control

Fixedmobile convergence FuncRonal convergence Advanced interface selecRon and route control

Fixedmobile convergence FuncRonal convergence Advanced interface selecRon and route control Dirk Breuer Deutsche Telekom Laboratories, Germany Tibor Cinkler Budapest University of Technology and Economics, Hungary Stphane Gosselin

90 views • 4 slides
Web Information Retrieval Lecture 6 Vector Space Model Recap of the last lecture Parametric

Web Information Retrieval Lecture 6 Vector Space Model Recap of the last lecture Parametric

Web Information Retrieval Lecture 6 Vector Space Model Recap of the last lecture Parametric and field searches Zones in documents Scoring documents: zone weighting Index support for scoring tf idf and vector spaces This

857 views • 32 slides
Part 5: Scoring, Term Weighting and the Vector Space Model Francesco Ricci Most of these slides

Part 5: Scoring, Term Weighting and the Vector Space Model Francesco Ricci Most of these slides

Part 5: Scoring, Term Weighting and the Vector Space Model Francesco Ricci Most of these slides comes from the course: Information Retrieval and Web Search, Christopher Manning and Prabhakar Raghavan 1 Content p Ranked retrieval p

527 views • 48 slides
The Web document collection No design/co-ordination I Unstructured (text, html, ),

The Web document collection No design/co-ordination I Unstructured (text, html, ),

Web Data Management Advanced Topics in Database Management (INFSCI 2711) Textbooks: Database System Concepts - 2010 Introduction to Information Retrieval - 2008 Vladimir Zadorozhny, DINS, SCI, University of Pittsburgh The Web document

470 views • 14 slides
XACML-Grid and XACML-NRP Attributes and Policy Profiles and Policy Obligations Handling

XACML-Grid and XACML-NRP Attributes and Policy Profiles and Policy Obligations Handling

XACML-Grid and XACML-NRP Attributes and Policy Profiles and Policy Obligations Handling Overview by Yuri Demchenko On behalf of OSG/EGEE AuthZ Interoperability WG and Phosphorus project SNE Group, University of Amsterdam TF-EMC2 Meeting 3

633 views • 50 slides
BIOE 301 HIV/AIDS Vaccine Update HIV/AIDS Vaccine Update Health Care Reform Update

BIOE 301 HIV/AIDS Vaccine Update HIV/AIDS Vaccine Update Health Care Reform Update

Lecture Thirteen BIOE 301 HIV/AIDS Vaccine Update HIV/AIDS Vaccine Update Health Care Reform Update http://www.npr.org/templates/story/story. php?storyId= 113781751&ps= rs Review of Cancer Lectures What is the goal of cancer

317 views • 30 slides
The Value of Evidence-Based Medicine in Establishing Ob-Gyn Clinical Practice Guidelines l l d

The Value of Evidence-Based Medicine in Establishing Ob-Gyn Clinical Practice Guidelines l l d

The Value of Evidence-Based Medicine in Establishing Ob-Gyn Clinical Practice Guidelines l l d l Diane Horvath-Cosper, MD, MPH, FACOG American Congress of Obstetricians and Gynecologists What is evidence-based medicine? What is evidence

257 views • 8 slides
Health and Health Care for Asian Americans, Native Hawaiians and Other Pacific Islanders (NHOPIs)

Health and Health Care for Asian Americans, Native Hawaiians and Other Pacific Islanders (NHOPIs)

Health and Health Care for Asian Americans, Native Hawaiians and Other Pacific Islanders (NHOPIs) in the United States May 2017 Figure 1 Nearly 16.5 million Asians and NHOPIs individuals self- account for about 6% identify as Asian or of the

452 views • 9 slides
WHOse New in WHOse New in What are the trends and what can we do Thyroid Pathology Thyroid

WHOse New in WHOse New in What are the trends and what can we do Thyroid Pathology Thyroid

L.D.R. Thompson Overall Objectives What is the current management of papillary carcinoma? WHOse New in WHOse New in What are the trends and what can we do Thyroid Pathology Thyroid Pathology differently? Supporting data

434 views • 17 slides

More recommend