thesis about concurrency frameworks

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

thesis about concurrency frameworks

thomas.kuehner
Hello gpars support,

I currently work on my bachelor thesis " Comparison of concurrency  
frameworks for the JVM" at the university of ulm in Germany . The  
comparison will regard qualitative as well as quantitative properties  
and the main goal is to find out which frameworks are more appropriate  
in different situations so that I'm able to write a kind of guideline  
for concurrency on the JVM. To figure out that I will specifiy some  
generic concurrency problems and evaluate how easy it is to implement  
them with the frameworks and additionally make some performance tests  
on the implementations. So far I apologize for the long introduction  
and now my actual questions.

1. Can you  tell me situations in which gpars faciliates programming  
concurrency very much or provides great performance but also  
situations were it's not that supportive. Are there papers available  
about such studies of gpars or even comparisons between gpars and  
other frameworks for the JVM.

2.If I send you my the specification of the problems and my actual  
implementations with gpars, is it possible to give me feedback where  
is space for improvement or advice to other approaches. I guess that's  
necessary to ensure that my implementations are good and don't lead to  
distorted results.  That would be great. Thank you very.

greetings
Thomas Kühner


---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email


Reply | Threaded
Open this post in threaded view
|

Re: thesis about concurrency frameworks

Vaclav
Administrator
Hi Thomas,

I'm happy you included GPars in your list of technologies to evaluate.

To your questions:

1. GPars is a multi-paradigm framework, which offers several different abstractions (actors, agents, dataflow, CSP, parallel collections, ans a few more) and let's you pick the one that works best for the problem at hand or combine several of these abstractions, where it makes sense. I tried to summarize my thoughts on how to apply GPars to different types of concurrency problems in my recent talk "GPars - a how-to guide", delivered at GR8Conf EU 2013. You my find the slide deck at http://www.slideshare.net/VaclavPech/g-pars-howto and I hope the recording should be made available soon, as well (perhaps at http://gr8conf.eu/Presentations/GPars-how-to-guide).
You may also get some ideas from the user guide (http://gpars.org/1.0.0/guide/index.html) or from the other presentations listed at http://gpars.codehaus.org/Presentations

2. Feel free to send me your implementations for review, perhaps through this mailing list so that others can chime it, too.

If there are any other questions I can help you with, please let me know.

Best regards,

Vaclav




On Sat, Jun 29, 2013 at 4:31 PM, <[hidden email]> wrote:
Hello gpars support,

I currently work on my bachelor thesis " Comparison of concurrency frameworks for the JVM" at the university of ulm in Germany . The comparison will regard qualitative as well as quantitative properties and the main goal is to find out which frameworks are more appropriate in different situations so that I'm able to write a kind of guideline for concurrency on the JVM. To figure out that I will specifiy some generic concurrency problems and evaluate how easy it is to implement them with the frameworks and additionally make some performance tests on the implementations. So far I apologize for the long introduction and now my actual questions.

1. Can you  tell me situations in which gpars faciliates programming concurrency very much or provides great performance but also situations were it's not that supportive. Are there papers available about such studies of gpars or even comparisons between gpars and other frameworks for the JVM.

2.If I send you my the specification of the problems and my actual implementations with gpars, is it possible to give me feedback where is space for improvement or advice to other approaches. I guess that's necessary to ensure that my implementations are good and don't lead to distorted results.  That would be great. Thank you very.

greetings
Thomas Kühner


---------------------------------------------------------------------
To unsubscribe from this list, please visit:

   http://xircles.codehaus.org/manage_email





--
E-mail: [hidden email]
Blog: http://www.jroller.com/vaclav
Linkedin page: http://www.linkedin.com/in/vaclavpech