Out-of of-GPU-Memory ry Graph Processing Amir Hossein Nodehi - - PowerPoint PPT Presentation

Su Subway: : Min inimizing Data Transfer during Out-of

• f-GPU-Memory

ry Graph Processing

Amir Hossein Nodehi Sabet, Zhijia Zhao, Rajiv Gupta Computer Science and Engineering UC Riverside

1

Background and Motivation

• GPUs enable massive parallelism for graph processing
• CuSha [1]
• Gunrock [2]
• Tigr [3]
• …
• Graphs can be large and tend to grow over time
• Web graphs
• Social networks
• But GPU memory is limited!!
• Out-of-GPU-Memory Graph Processing

[1] Khorasani, Farzad, et al. "CuSha: vertex-centric graph processing on GPUs.” HPDC’14 [2] Wang, Yangzihao, et al. "Gunrock: A high-performance graph processing library on the GPU.” PPoPP’16 [3] Nodehi Sabet, Amir Hossein, Junqiao Qiu, and Zhijia Zhao. "Tigr: Transforming irregular graphs for gpu-friendly graph processing.” ASPLOS’18 2

Partition-based Graph Processing

Computation Transferring

Main Memory GPU Memory

3

A Key Observation

Ratio of active vertices (edges) is often low in most iterations

Algo. friendster Uk-2007 SSSP 9.1% 5.1% BFS 4.1% 0.6% CC 9.8% 3.2%

Average Ratio of Active Edges across Iterations

4

Only Load Active Edges to GPU?

Main Memory GPU Memory

5

Too expensive to generate ?!

Efficient Subgraph Generation

Subway:

• a concise subgraph representation, called SubCSR
• a highly parallel algorithm for subgraph generation
• an efficient GPU-accelerated implementation

6

SubCSR Generation Cost

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 FS UK FS UK FS UK SSSP BFS CC Relative Cost PT (Transfer) Subway-sync (SubCSR + Transfer)

3% 17%

Costs: Partitioning-based vs. Subway (subgraph generation)

7

Takeaway

Too expensive to dynamically generate subgraphs! Improve performance up to 28X !

Subway

8

Thank you

Amir Nodehi: anode001@ucr.edu or on Slack The source code (to be posted soon): https://github.com/AutomataLab/Subway

9