2019/03/25 Add Improve BMP-HY, BMP-OP 2019/03/15 Selected Tri-Cnt, All-Edge-Cnt 2019/03/02 Triangle Counting Overview (GPU23-Synthetic) 2019/03/02 Triangle Counting Overview (GPU23) 2019/01/22 Experimental Overview (CPU/KNL)
add reordering
gro lack twitter/friendster
metis not correct yet (segment fault), cache not working yet (performance degradation)
2019/01/09-2019/01/14 Experimental Overview (CPU/KNL)
add emptyheaded, lemire, roaring, han's qfilter and bsr, bmiss/bmiss+sttni, tetzanks
not remove the construct time yet
2018/09/29 Effectiveness of Symmetric Assignment 2018/09/25 Bitmap 1D vs 2D 2018/09/23 CPU/KNL Scalability to number of Threads 2018/09/23 KNL Experiments 2018/09/20, 2018/09/23 CUDA Scalability results: data-md/lccpu12/09-23-lb-gpu-wildcard.md
2018/09/20 CUDA Load Balance Dynamic (Config: 8x) 2018/09/18 Performance Study On CPU and KNL 2018/09/17 CUDA Dedicated Performance Study (Hybrid Kernels / Hybird Computation per Thread / Bitmap) on Single-GPU 2018/09/13 CUDA With Co-Processing For Bin-Search on CPU All 2018/09/11 CUDA Multi-Pass To Process Snap-Friendster 2018/09/10 CUDA Multi-GPU 2018/09/08 CUDA Unified-Memory Bitmap (With 4x Index Range Size) 2018/09/08 CUDA Unified-Memory To Support Larger Dataset 2018/09/06 CUDA-Naive Implementation Time
exec
description
cuda-baseline-hybrid
one edge intersect(u,v) per cuda-thread
cuda-baseline
one edge intersect(u,v) per cuda-thread
cuda
global memory, each warp per edge intersect(u,v)
2018/09/03 Pre-Processing Edge-List to CSR Time To show this time is trivial compared to the intensive set-intersections for our application.
2018/09/01 Removing edge_src, Adopting BitVec
Attention: optimal workload in bitvec requires degree-descending traversal, with partial computation contraint u < v
BitVec Cost: num of threads * (number of vertices / 8) bytes
2018/08/28 Reordering with IPDPS rabbit_order, Not-Useful