graphp
play

GraphP phP : Reducing Communication for PIM-based Graph Processing - PowerPoint PPT Presentation

Nov 16, 2022 •411 likes •1.12k views

GraphP phP : Reducing Communication for PIM-based Graph Processing with Efficient Data Partition Mingxing Zhang, Youwei Zhuo (equal contribution ), Chao Wang, Mingyu Gao, Yongwei Wu, Kang Chen, Christos Kozyrakis, Xuehai Qian Tsinghua

  1. GraphP phP : Reducing Communication for PIM-based Graph Processing with Efficient Data Partition Mingxing Zhang, Youwei Zhuo (equal contribution ), Chao Wang, Mingyu Gao, Yongwei Wu, Kang Chen, Christos Kozyrakis, Xuehai Qian Tsinghua University University of Southern California Stanford University

  2. Outline • Motivation • Graph applications • Processing-In-Memory • The drawbacks of the current solution • GraphP • Evaluation ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  3. Graph Applications • Social network analytics • Recommendation system • Bioinformatics • … ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  4. Challenges • High bandwidth requirement • Small amount of computation per vertex • Data movement overhead ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  5. Challenges • High bandwidth requirement • Small amount of computation per vertex • Data movement overhead comp L1 L2 L3 mem ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  6. PIM: Processing-In-Memory ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  7. PIM: Processing-In-Memory • Idea : Computation logic inside memory • Advantage : High memory bandwidth ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  8. PIM: Processing-In-Memory • Idea : Computation logic inside memory • Advantage : High memory bandwidth • Example : Hybrid Memory Cubes (HMC) ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  9. PIM: Processing-In-Memory • Idea : Computation logic inside memory • Advantage : High memory bandwidth • Example : Hybrid Memory Cubes (HMC) mem … .. mem mem 320GB/s intra-cube mem 4x120GB/s comp inter-cube ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  10. HMC: Hybrid Memory Cubes 320 Intra-cube bandwidth (GB/s) ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  11. HMC: Hybrid Memory Cubes 120 320 Intra-cube Inter-cube bandwidth (GB/s) ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  12. HMC: Hybrid Memory Cubes 120 320 Intra-cube Inter-cube bandwidth (GB/s) ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  13. HMC: Hybrid Memory Cubes 120 320 Intra-cube Inter-cube bandwidth (GB/s) ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  14. HMC: Hybrid Memory Cubes 120 120 320 Intra-cube Inter-cube Inter-group bandwidth (GB/s) ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  15. HMC: Hybrid Memory Cubes 120 120 320 Intra-cube Bottleneck: Inter-cube communication Inter-cube Inter-group bandwidth (GB/s) ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  16. Outline • Motivation • Graph applications • Processing-In-Memory • The drawbacks of the current solution • GraphP • Evaluation ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  17. Current Solution: Tesseract • First PIM-based graph processing architecture Ahn, J., Hong, S., Yoo, S., Mutlu, O., & Choi, K. A scalable processing-in- memory accelerator for parallel graph processing. In 2015 ACM/IEEE 42nd Annual International Symposium on Computer Architecture (ISCA) . ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  18. Current Solution: Tesseract • First PIM-based graph processing architecture • Programming model • Vertex program Ahn, J., Hong, S., Yoo, S., Mutlu, O., & Choi, K. A scalable processing-in- memory accelerator for parallel graph processing. In 2015 ACM/IEEE 42nd Annual International Symposium on Computer Architecture (ISCA) . ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  19. Current Solution: Tesseract • First PIM-based graph processing architecture • Programming model • Vertex program • Partition • Based on vertex program Ahn, J., Hong, S., Yoo, S., Mutlu, O., & Choi, K. A scalable processing-in- memory accelerator for parallel graph processing. In 2015 ACM/IEEE 42nd Annual International Symposium on Computer Architecture (ISCA) . ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  20. PageRank in Vertex Program for (v: vertices ) { update = 0.85 * v.rank / v.out_degree; } ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  21. PageRank in Vertex Program for (v: vertices ) { update = 0.85 * v.rank / v.out_degree; for (w: edges.destination ) { put(w.id, function{ w.next_rank += update; }); } } ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  22. PageRank in Vertex Program for (v: vertices ) { update = 0.85 * v.rank / v.out_degree; for (w: edges.destination ) { put(w.id, function{ w.next_rank += update; }); } } barrier(); ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  23. Graph Partition hmc0 0 1 2 0 1 2 5 3 4 vertex 1 3 4 5 hmc1 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  24. Graph Partition hmc0 0 1 2 0 1 2 5 3 4 intra vertex 1 3 4 5 edge inter hmc1 edge ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  25. Graph Partition hmc0 0 1 2 0 1 2 5 3 4 intra vertex 1 3 4 5 edge inter hmc1 comm edge put(w.id, function{ w.next_rank += update; }); ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  26. Graph Partition hmc0 0 1 2 0 1 2 5 3 4 intra vertex 1 3 4 5 edge inter hmc1 comm edge communication = # of cross-cube edges ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  27. Drawback of Tesseract • Excessive data communication • Why? Data Programming Graph Communication Model Partition Tesseract ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  28. Drawback of Tesseract • Excessive data communication • Why? Data Programming Graph Communication Model Partition Tesseract ? ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  29. Drawback of Tesseract • Excessive data communication • Why? Data Programming Graph Communication Model Partition Tesseract ? ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  30. Drawback of Tesseract • Excessive data communication • Why? Data Programming Graph Communication Model Partition Tesseract ? ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  31. Outline • Motivation • GraphP • Evaluation ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  32. GraphP • Consider graph partition first. • Graph Partition • Source-Cut • Programming model • Two-phase vertex program • Reduces inter-cube communication ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  33. Source-Cut Partition 0 1 2 0 1 2 hmc0 5 3 4 vertex 1 3 4 5 hmc1 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  34. Source-Cut Partition 0 1 2 0 1 2 hmc0 5 3 4 intra vertex 1 edge inter edge 3 4 5 hmc1 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  35. Source-Cut Partition 0 1 2 0 1 2 hmc0 5 3 4 intra vertex 1 edge 2 inter replica 2 edge 3 4 5 hmc1 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  36. Source-Cut Partition 0 1 2 0 1 2 hmc0 5 3 4 intra vertex 1 edge 2 inter replica 2 edge 3 4 5 hmc1 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  37. Source-Cut Partition 0 1 2 0 1 2 hmc0 5 3 4 intra vertex 1 edge 2 inter replica 2 edge 3 4 5 hmc1 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  38. Two-Phase Vertex Program for (r: replicas) { r.next_rank = 0.85 * r.next_rank / r.out_degree; } //apply updates from previous iterations 2 5 3 4 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  39. Two-Phase Vertex Program for (r: replicas) { r.next_rank = 0.85 * r.next_rank / r.out_degree; } //apply updates from previous iterations 2 5 3 4 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  40. Two-Phase Vertex Program 2 3 4 5 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  41. Two-Phase Vertex Program for (v: vertices ) { for (u: edges.sources ) { } 2 3 4 5 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  42. Two-Phase Vertex Program for (v: vertices ) { for (u: edges.sources ) { update += u.rank; } 2 3 4 5 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  43. Two-Phase Vertex Program for (v: vertices ) { for (u: edges.sources ) { update += u.rank; } 2 3 4 5 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  44. Two-Phase Vertex Program for (r: replicas ) { put(r.id, function { r.next_rank = update}); } 0 2 } barrier(); 5 3 4 3 4 ALCHEM GraphP: A PIM-based Graph Processing Framework alchem.usc.edu

  45. Benefits • Strictly less data communication • Enables architecture optimizations ALCHEM GraphP: A PIM-based Graph Processing Framework 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

PUBLIC KEY INFRASTRUCTURE Nina Bindel Cryptography for the IoT+Cloud Udyani Herath Bochum,

PUBLIC KEY INFRASTRUCTURE Nina Bindel Cryptography for the IoT+Cloud Udyani Herath Bochum,

TRANSITIONING TO A QUANTUM-RESISTENT PUBLIC KEY INFRASTRUCTURE Nina Bindel Cryptography for the IoT+Cloud Udyani Herath Bochum, Germany Matthew McKague 11/06/2017 Douglas Stebila 1 7 chance of breaking RSA-2048 (Michele Mosca Nov 2015)

614 views • 17 slides
Using Processor Partitioning to Using Processor Partitioning to Evaluate the Performance of MPI,

Using Processor Partitioning to Using Processor Partitioning to Evaluate the Performance of MPI,

http://prophesy.cs.tamu.edu Using Processor Partitioning to Using Processor Partitioning to Evaluate the Performance of MPI, Evaluate the Performance of MPI, OpenMP and Hybrid Parallel OpenMP and Hybrid Parallel Applications on Dual- and

483 views • 29 slides
Solution Concepts www.unisi.it and W ell-posedness of Hybrid Systems Maurice Heemels Embedded

Solution Concepts www.unisi.it and W ell-posedness of Hybrid Systems Maurice Heemels Embedded

st 1 HYCON PhD School on Hybrid Systems www.ist-hycon.org Solution Concepts www.unisi.it and W ell-posedness of Hybrid Systems Maurice Heemels Embedded Systems Institute (NL) maurice.heemels@embeddedsystems.nl scimanyd suounitnoc enibmoc

907 views • 63 slides
CHAPTER 5: REACTIVE AND HYBRID ARCHITECTURES An Introduction to Multiagent Systems

CHAPTER 5: REACTIVE AND HYBRID ARCHITECTURES An Introduction to Multiagent Systems

CHAPTER 5: REACTIVE AND HYBRID ARCHITECTURES An Introduction to Multiagent Systems http://www.csc.liv.ac.uk/mjw/pubs/imas/ Chapter 5 An Introduction to Multiagent Systems 0.1 Reactive Architectures There are many

421 views • 19 slides
Towards Hybrid Isolation for Shared Multicore Systems 23rd Workshop on Job Scheduling Strategies

Towards Hybrid Isolation for Shared Multicore Systems 23rd Workshop on Job Scheduling Strategies

1 Towards Hybrid Isolation for Shared Multicore Systems 23rd Workshop on Job Scheduling Strategies for Parallel Processing, (JSSPP), 2020 Yoonsung Nam , Byeonghun Yoo , Yongjun Choi , Yongseok Son , and Hyeonsang Eom Seoul

635 views • 19 slides
formulation and single resolution experiments with real data for NCEP GFS Ting Lei, Xuguang Wang

formulation and single resolution experiments with real data for NCEP GFS Ting Lei, Xuguang Wang

GSI-based four dimensional ensemble-variational (4DEnsVar) data assimilation: formulation and single resolution experiments with real data for NCEP GFS Ting Lei, Xuguang Wang University of Oklahoma, Norman, OK, USA Wang and Lei, MWR, 2013

575 views • 21 slides
Hybrid scheme Kerberos Protocol Public-key: nice solution for key distribution, but

Hybrid scheme Kerberos Protocol Public-key: nice solution for key distribution, but

Hybrid scheme Kerberos Protocol Public-key: nice solution for key distribution, but Motivation : In a multi-national company, its computational expensive distributed information services are usually maintained by various business units

461 views • 7 slides
Distributed hybrid control synthesis for multi-agent systems from high level specifcations

Distributed hybrid control synthesis for multi-agent systems from high level specifcations

Distributed hybrid control synthesis for multi-agent systems from high level specifcations OptHySYS Workshop, Trento, January 2017 Dimos V. Dimarogonas (joint work with Jana Tumova, Dimitris Boskos and Meng Guo) Automatic Control, KTH Royal

782 views • 62 slides
Hybrid Co-scheduling Optimizations for Concurrent Applications in Virtualized Environments

Hybrid Co-scheduling Optimizations for Concurrent Applications in Virtualized Environments

Hybrid Co-scheduling Optimizations for Concurrent Applications in Virtualized Environments Yulong Yu School of Software Dalian University of Technology The 6 th International Conference on Networking, Architecture, and Storage (NAS), July

455 views • 22 slides
Hybrid CLS-Opamp/ZCBC Pipelined ADC Using a 300mV Output Swing Opamp Benjamin Hershberg Skyler

Hybrid CLS-Opamp/ZCBC Pipelined ADC Using a 300mV Output Swing Opamp Benjamin Hershberg Skyler

A 1.4V Signal Swing Hybrid CLS-Opamp/ZCBC Pipelined ADC Using a 300mV Output Swing Opamp Benjamin Hershberg Skyler Weaver Un-Ku Moon Oregon State University, Corvallis, OR Preview 1. Background Correlated Level Shifting

712 views • 54 slides
Strip Module Design Short Strip Modules for Barrel region feature: Silicon sensor plate

Strip Module Design Short Strip Modules for Barrel region feature: Silicon sensor plate

LBL Weekly Meeting (May 23rd 2017) 1 Strip Module Design Short Strip Modules for Barrel region feature: Silicon sensor plate Two hybrids 10 Atlas Binary Chip (ABC 130) with 256 readout channels. 1 Hybrid Control Chip (HCC)

526 views • 16 slides
A/acker'Knowledge ' Frdric'Besson,'Nataliia'Bielova ,' Thomas'Jensen' INRIA' '

A/acker'Knowledge ' Frdric'Besson,'Nataliia'Bielova ,' Thomas'Jensen' INRIA' '

Hybrid'Monitoring'of'' A/acker'Knowledge ' Frdric'Besson,'Nataliia'Bielova ,' Thomas'Jensen' INRIA' ' IEEE'Computer'Security'Founda@ons'2016' June'29,'2016' ' Informa@on'flow'control'' Quan=ta=ve' Noninterference' A/ackers'knowledge'

677 views • 38 slides
Hybrid SMR Drives Fenggang Wu , Bingzhe Li, Zhichao Cao, Baoquan Zhang Ming-Hong Yang, Hao Wen,

Hybrid SMR Drives Fenggang Wu , Bingzhe Li, Zhichao Cao, Baoquan Zhang Ming-Hong Yang, Hao Wen,

ZoneAlloy : Elastic ZoneAlloy Data and Space Management for Hybrid SMR Drives Fenggang Wu , Bingzhe Li, Zhichao Cao, Baoquan Zhang Ming-Hong Yang, Hao Wen, David H.C. Du University of Minnesota, Twin Cities Jul. 8, 2019. HotStorage19 C

624 views • 33 slides
P h y s i c s o f b i o l o g i c a l s y s t e ms P H 5 4 9 L

P h y s i c s o f b i o l o g i c a l s y s t e ms P H 5 4 9 L

P h y s i c s o f b i o l o g i c a l s y s t e ms P H 5 4 9 L E C T U R E 1 2 : c h r o ma t i n p a c k i n g mithun@phy.iitb.ac.in Office: Physics 303 D N A t o c h r o mo s

567 views • 29 slides
RESEARCH & METHODS RNA-RNA interaction prediction Jerome Waldispuhl School of Computer

RESEARCH & METHODS RNA-RNA interaction prediction Jerome Waldispuhl School of Computer

COMP598: ADVANCED COMPUTATIONAL BIOLOGY RESEARCH & METHODS RNA-RNA interaction prediction Jerome Waldispuhl School of Computer Science, McGill From slides from Ivo Hofacker (University of Vienna) Motivation Experimental and

481 views • 36 slides
PTT 207 Biomolecular and Genetic Engineering Semester 1 2012/2013 BY: PUAN NURUL AIN HARMIZA

PTT 207 Biomolecular and Genetic Engineering Semester 1 2012/2013 BY: PUAN NURUL AIN HARMIZA

PTT 207 Biomolecular and Genetic Engineering Semester 1 2012/2013 BY: PUAN NURUL AIN HARMIZA ABDULLAH Chapter 2 The Structure of DNA Imagine yourself inside one of those cooling DNA solutions, observing the rebirth of beautifully

506 views • 46 slides
1 Molecular characters Nucleotide sequences structural genes (protein, RNA, regulatory)

1 Molecular characters Nucleotide sequences structural genes (protein, RNA, regulatory)

Phylogenetics 3: Methods to reconstruct phylogenies A generalized protocol for molecular phylogenetics, and the associated concerns with each step Concerns: Collect homolgous sequences gene tree-species tree / paralogyorthology / trees

527 views • 26 slides
Synthesis of N -acetyl and N -formyl pyrazoline derivatives from vanillin and their antigenotoxic

Synthesis of N -acetyl and N -formyl pyrazoline derivatives from vanillin and their antigenotoxic

Synthesis of N -acetyl and N -formyl pyrazoline derivatives from vanillin and their antigenotoxic activity Jovana Mukinja 1,2, *, Sanja Mati 1,3 , Sneana Stani 3 and Zoran Ratkovi 2 1 Department of Science, Institute for Information

203 views • 16 slides
Computational Study on the Structure of N- (2-Amino-benzoyl)-N-phenyl hydrazine Ibrahim SEN,

Computational Study on the Structure of N- (2-Amino-benzoyl)-N-phenyl hydrazine Ibrahim SEN,

Computational Study on the Structure of N- (2-Amino-benzoyl)-N-phenyl hydrazine Ibrahim SEN, Akn AZIZOGLU Laboratory of Computational Chemistry, Department of Chemistry , University of Balikesir, 10145-Balikesir, Turkey

220 views • 5 slides
1 Miniature Mode Estimation Example ACS Subsystem GHe Pressure Transducer (S) Model

1 Miniature Mode Estimation Example ACS Subsystem GHe Pressure Transducer (S) Model

Outline Motivation Previous Mode Estimation Approaches Example ACS System Improving Model-based Mode Estimation Miniature Mode Estimation System Through Offline Compilation Rule-based system comparison Conclusions

495 views • 4 slides
Overview of Controlled re-entry activities Tiago Soares 26/10/2017 ESA UNCLASSIFIED - For

Overview of Controlled re-entry activities Tiago Soares 26/10/2017 ESA UNCLASSIFIED - For

Overview of Controlled re-entry activities Tiago Soares 26/10/2017 ESA UNCLASSIFIED - For Official Use Outline - Why controlled re-entry? - What does it entail? Technical challenges - Current technology solution - Possible solutions and

236 views • 9 slides
Goals Stress the global importance of iden:fying laboratory

Goals Stress the global importance of iden:fying laboratory

3/11/16 Common Laboratory Hazards Rich Wittman, MD, MPH Clinical Assistant Professor rwi$man@stanford.edu Medical Director, Stanford and SLAC Occupational Health Centers Disclosures I have nothing to disclose 1 3/11/16

528 views • 16 slides
PREPARATION, PROPERTIES PREPARATION OF METALLIC PARTICLES AND APPLICATIONS Phase break

PREPARATION, PROPERTIES PREPARATION OF METALLIC PARTICLES AND APPLICATIONS Phase break

PREPARATION, PROPERTIES PREPARATION OF METALLIC PARTICLES AND APPLICATIONS Phase break down - Milling/grinding - Atomization OF Phase transformation - Thermolysis/Pyrolysis HIGHLY DISPERSED METALLIC - Reduction

145 views • 12 slides
Research & Development Frank Doyle Director, Research & Development 1 Oct 6, 2009

Research & Development Frank Doyle Director, Research & Development 1 Oct 6, 2009

CANDU Owners Group Inc. Strength Through Cooperation Research & Development Frank Doyle Director, Research & Development 1 Oct 6, 2009 CNS Ottawa Branch Meeting CANDU Owners Group Inc. Program Areas Strength Through

1.11k views • 51 slides

More recommend