A Locality-Aware Energy-Efficient Accelerator for Graph Mining Applications

Yao, Pengcheng; Zheng, Long; Zeng, Zhen; Huang, Yu; Gui, Chuangyi; Jin, Hai; Xue, Jingling

doi:10.1109/micro50266.2020.00077

Cited by 35 publications

(27 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Its performance is limited due to the lack of symmetry breaking support and blindly processing graph data as a database table. GRAMER [73] is based on a much slower pattern-oblivious algorithm with expensive isomorphic check. Its execution time after speedup is even longer than directly executing pattern enumeration on commodity machines.…”

Section: Existing Architectures On Gpmmentioning

confidence: 99%

“…we implemented the cmap and simulated their access patterns. For GRAMER [73], we implemented its specialized memory hierarchy and simulated the access patterns. For TrieJax, Flexminer, and GRAMER, we all assume full overlapping of any non-dependent data access.…”

Section: Evaluation 61 Simulator and Configurationmentioning

confidence: 99%

“…PJR cache will deallocate large entries that exceed 1KB, which corresponds to only 256 vertices. However, for GPM applications, vertices with high degrees are more likely to be accessed [73], which usually cannot be placed in PJR. Hence, the PJR cache fail to cache the most frequently accessed data.…”

Section: Overall Performance Of Gpmmentioning

confidence: 99%

“…This difficulty gives rise to the recent interests of designing specific GPM accelerators. TrieJax [27] and GRAMER [73] are two early GPM accelerators. Their performance is limited since they did not accelerate the state-of-the-art algorithms.…”

mentioning

confidence: 99%

See 3 more Smart Citations

SparseCore: stream ISA and processor specialization for sparse computation

Rao

Chen

Yik

et al. 2022

Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems

View full text Add to dashboard Cite

Computation on sparse data is becoming increasingly important for many applications. Recent sparse computation accelerators are designed for specific algorithm/application, making them inflexible with software optimizations. This paper proposes SparseCore, the first general-purpose processor extension for sparse computation that can flexibly accelerate complex code patterns and fast-evolving algorithms. We extend the instruction set architecture (ISA) to make stream or sparse vector first-class citizens, and develop efficient architectural components to support the stream ISA. The novel ISA extension intrinsically operates on streams, realizing both efficient data movement and computation. The simulation results show that SparseCore achieves significant speedups for sparse tensor computation and graph pattern computation. CCS CONCEPTS• Computer systems organization → Architectures.

show abstract

Section: Existing Architectures On Gpmmentioning

confidence: 99%

Section: Evaluation 61 Simulator and Configurationmentioning

confidence: 99%

Section: Overall Performance Of Gpmmentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

SparseCore: stream ISA and processor specialization for sparse computation

Rao

Chen

Yik

et al. 2022

Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems

View full text Add to dashboard Cite

show abstract

“…Machine Eciency Analysis We show example analysis of CPU utilization, using the PAPI interface provided in GMS, see Figure 6b. The plots illustrate the attening of speedups with the increasing #threads, accompanied by the steady growth of stalled CPU cycles (both total counts and ratios), illustrating that maximal clique listing is memory bound [36,50,64,129,130]. Memory Consumption Analysis We illustrate example memory consumption results in Figure 6c; we compare the size of three GMS set-centric graph representations, showing both peak memory usage when constructing a representation (bars) and sizes of ready representations (all in GB).…”

Section: Additional Analysesmentioning

confidence: 99%

GraphMineSuite

et al. 2021

View full text Add to dashboard Cite

We propose GraphMineSuite (GMS): the first benchmarking suite for graph mining that facilitates evaluating and constructing high-performance graph mining algorithms. First, GMS comes with a benchmark specification based on extensive literature review, prescribing representative problems, algorithms, and datasets. Second, GMS offers a carefully designed software platform for seamless testing of different fine-grained elements of graph mining algorithms, such as graph representations or algorithm subroutines. The platform includes parallel implementations of more than 40 considered baselines, and it facilitates developing complex and fast mining algorithms. High modularity is possible by harnessing set algebra operations such as set intersection and difference, which enables breaking complex graph mining algorithms into simple building blocks that can be separately experimented with. GMS is supported with a broad concurrency analysis for portability in performance insights, and a novel performance metric to assess the throughput of graph mining algorithms, enabling more insightful evaluation. As use cases, we harness GMS to rapidly redesign and accelerate state-of-the-art baselines of core graph mining problems: degeneracy reordering (by >2X), maximal clique listing (by >9×), k -clique listing (by up to 1.1×), and subgraph isomorphism (by 2.5×), also obtaining better theoretical performance bounds.

show abstract