Thanks, David! Much appreciated!
~brian
On Sun, Apr 27, 2014 at 8:21 PM, David Eccles (gringer) <
bioinformat...@gringene.org> wrote:
> On 26.04.2014 05:50, Brian Haas wrote:
> > David - you know quite a bit about Ray. Do you understand how Ray is
> leveraging MPI?
>
> Short answer, no.
>
> I have used it many times, and managed to hack it a bit. I wouldn't say my
> knowledge about Ray is extensive, more like my lack of knowledge
> is extensive. Your best option for this would be to contact Sébastien
> Boisvert directly (CCd), or via the denovoassembler-users mailing
> list, because I only have a very vague idea about how the parallel stuff
> works in Ray. I just know it works really well, mostly from other
> people saying how wonderful Ray is in their blogs.
>
> I recall Sébastien writing a couple of blog-like posts about message
> passing and virtual messages (which I can no longer find), but that was
> a long time ago and before the RayPlatform interface was developed.
> There's a bit of documentation in the RayPlatform Documentation directory:
>
> https://github.com/sebhtml/RayPlatform/tree/master/Documentation
>
> But it's had most of the assembler-specific things removed from it. I've
> have quite a lot of trouble finding anything that summarises what's
> going on behind the scenes. I have found a header file that has message
> names in it (which was informative for me previously when hacking in
> a couple of things for testing):
>
>
> https://github.com/sebhtml/ray/blob/master/code/MessageProcessor/MessageProcessor.h
>
> And it looks like Sébastien gave a presentation of this topic:
>
>
> http://boisvert.info/dropbox/SciNet/SciNet-presentation-for-computer-scientists.pdf
>
> [message stuff starts at pg. 32, but again there's not much specifically
> about assembly]
>
> It's possible you could get your best idea about what's going on by
> running Ray on a toy dataset and just seeing the message output. Ray is
> very verbose, which is a very nice thing to have when an assembly can take
> many, many hours to finish.
>
> Hope this helps,
>
> - David
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos. Get
> unparalleled scalability from the best Selenium testing platform available.
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Trinityrnaseq-developers mailing list
> trinityrnaseq-develop...@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/trinityrnaseq-developers
>
--
--
Brian J. Haas
The Broad Institute
http://broad.mit.edu/~bhaas
------------------------------------------------------------------------------
"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
Instantly run your Selenium tests across 300+ browser/OS combos. Get
unparalleled scalability from the best Selenium testing platform available.
Simple to use. Nothing to install. Get started now for free."
http://p.sf.net/sfu/SauceLabs
_______________________________________________
Denovoassembler-users mailing list
Denovoassembler-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/denovoassembler-users
