Leyuan Wang
Leyuan Wang
Hi @weihanmines, the formats of the graphs are the same as those when doing query without labels. And labels are sent in as an NDArray. To pass in labels in...
@wzbxpy it looks from the output that the reference is wrong instead of the Gunrock results . I’ll double check it. Also the tc in Gunrock is based on this...
@wzbxpy we haven't published the code for it yet. We are working on cleaning the code.
For the reference code bug, I confirmed that it's not working any more with any dataset. But before this commit (b4717f22106bccdeab1df88af583a90ff6f57ec7), the reference code was tested and worked properly. @neoblizz...
@wzbxpy yes, we are working on getting the code to Gunrock now. I’ll ping you once it gets merged. For now if you want to compare with the performance, please...
@ChuckHastings Thanks for your detailed exploration on the bug, I'll investigate into it and make changes accordingly.
@archenroot I use simple K/V pair to store the edge/vertex info
Hi @xianglizhi123 , you can lower the mem_limit here: https://github.com/gunrock/gunrock/blob/dev/gunrock/app/sm/sm_problem.cuh#L188 to reduce the pre-allocated memory size if you get run out of memory error. Pls lemme know if that helps....
@xianglizhi123 The memory limit is to ensure that the intermediate results generated in this app needs memory space less than the maximum GPU global memory. Could you try to print...
@AKKamath Thanks for detecting the data race. I should have assigned the atomic results to a different array.