elssky@home:~$

Kvcc Experiments On Ssv

1.直接计算SSV的时候neighbor交集如何处理?

image-20220217145645357

solution

image-20220225081614778

2.如何更准确地利用lemma11, 12计算SSV?

image-20220217145658098

3.怎么平衡计算SSV的花费与其带来的sweep的收益?

CA-AstroPh.txt

k=11(20% kmax)

image-20220225081653786

k=23(40% kmax),compute SSV every step

image-20220225100606184

compute SSV once

image-20220225101023937

VCCE:

image-20220225100758816

只计算flow<k的LOC-CUT部分

image-20220225101931632

只计算flow<k的LOC-CUT部分

image-20220225102253043

flow>=k,

2491

2297

Conclusion

SSV能sweep的LOC-CUT大部分是flow<k的情况,这部分耗时不多,所以效果不显著

New Data

CA-AstroPh.txt

image-20220302193632171

K=23

image-20220301223551137

K=34

image-20220301224006815

K=11

image-20220301230102079

image-20220304105700895

web-Stanford.txt

K=28(40%)

image-20220302184635041

image-20220302190112386

5/5毕设实验结果

CA-AstroPh

image-20220505150805252

Stanford

image-20220505151132497

image-20220505151647904

image-20220505153558777

image-20220505181418595

image-20220505183722563

CA-CondMat

image-20220505160052610

image-20220505165503162

Cit