Implementing Push-Pull Efficiently in GraphBLAS · HackerPapers discussions

Implementing Push-Pull Efficiently in GraphBLAS — HN submission

Wed, 24 Oct 2018 18:56:13 GMT

Implementing Push-Pull Efficiently in GraphBLAS

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Implementing Push-Pull Efficiently in GraphBLAS — mentioned by ctchocula

Tue, 16 Oct 2018 03:54:07 GMT

Hi Todd! Our recent work implements a subset of GraphBLAS operations for GPU and compares them to Gunrock in breadth-first-search [1]. Our implementation of a subset of GraphBLAS is comparable to Gunrock performance for power law graphs, but are worse for mesh graphs. Gunrock uses a different load-balancer in Advance for those graphs and the load-balancer we use in the analogous operation (matrix-vector multiplication) isn't as optimized for mesh graphs. We definitely want to collect data in more applications than just BFS, so we're working on that now. The code is open-source, so feel free…

Paper and discussion history. A mention is not an endorsement.

Implementing Push-Pull Efficiently in GraphBLAS — mentioned by ctchocula

Mon, 01 Oct 2018 05:44:14 GMT

GraphBLAS is an open standard, and I believe the implementation in the link provided targets single-threaded CPU, which rules it out for comparison with nvgraph, custinger and gunrock. My recent work actually implements some GraphBLAS operations for the GPU and compares them to Gunrock in breadth-first-search [1]. Our findings are our implementation of a subset of GraphBLAS is comparable to Gunrock in performance for power law graphs, but not for mesh graphs. Gunrock uses a different load-balancer in Advance for those graphs and the load-balancer we use in the analogous operation…

Paper and discussion history. A mention is not an endorsement.