[CHiPPS] Questions on Blis
Alessandro Tiberi
alessandro.tiberi at gmail.com
Thu Mar 12 12:12:30 EDT 2009
Yan Xu ha scritto:
> As to gap, you are right. The best bound seems do not move. The parallel log was changed to style of commercial solvers last year, I did not test much. I will take a look.
>
Great, thanks!
> For a child node, Blis only store the difference with its parent. I haven't thought about zipping and don't know how to do this yet. Any suggestion?
>
Well, my very preliminary idea is just to apply a simple serialization
mechanism, so the steps would be:
1) Represent a node description as a memory stream
2) fed this stream to a zipper
3) The Node Desc. wrapper would hence contain the result of 2 plus the
ability to uncompress it back to Node Des.
But as I said it is very preliminary, and I still do not have a global
picture...
best,
Alessandro
More information about the CHiPPS
mailing list