Re: [Bigdata-developers] [bigdata®] #740: performance impact of NSPIN

[Jeremy J C. <jj...@sy...>]

On Oct 4, 2013, at 3:17 PM, Bryan Thompson <br...@sy...> wrote:

> Let me suggest that we create a system propery for nspin that will let you override it and gain experience with that without pushing a change to everyone.



I am very happy with this as a next step - I may split the variable into two: one for the read end (the iterator) and one for the write end. If I do that I would only expose the former as a system property.

From my point of view, it would get Syapse off a forked version of the code. I see the fork as a necessary evil, use of which should be minimized.

Jeremy J Carroll
Principal Architect
Syapse, Inc.