Difference between revisions of "Ghent Functional Programming Group/BelHac/Projects"
(Project page instantiation.) |
(Added HaBench as a project.) |
||
| Line 12: | Line 12: | ||
Hackers: Hacker 1, Hacker 2 |
Hackers: Hacker 1, Hacker 2 |
||
--> |
--> |
||
| + | |||
| + | |||
| + | === HaBench === |
||
| + | |||
| + | Two years ago, boegel ([http://www.elis.ugent.be/~kehoste Kenneth Hoste]) and Itkovian ([http://www.itkovian.net Andy Georges]) tried to fire up an effort to construct a new, representative and neat benchmark suite for Haskell, under the name [[HaBench|HaBench]]. Since then, not much happened, except people kept contributing to and making (small) changes in the nofib suite. Unfortunately, nofib mainly consists of short running applications. Given the fact that Haskell programs come with a built-in runtime management system, chances are your measurements on this suite might be perturbed by the runtime more than you'd like. |
||
| + | |||
| + | The plan is to develop a framework for the new suite, modelled on nofib, but using best practices, such as criterion. The proposal on the main [http://github.com/itkovian/HaBench github repo] reflects what I think should be considered in order to obtain a decent framework. The initial benchmarks that will take place in the framework will most likely be those from nofib, the language shootout, etc. We shall see where we can take this beast, but I'd love it if we could have some finalised version (0.9 :-) by the deadline for the next Haskell Symposium, so we can submit something there. |
||
Revision as of 13:55, 29 July 2010
Projects
If you have a project that you want to work on at the Hackathon, please describe it here.
Since Hackathons are great for teamwork, consider joining one of the projects mentioned below. If you're interested in one of these projects, add your name to the list of hackers under that project.
HaBench
Two years ago, boegel (Kenneth Hoste) and Itkovian (Andy Georges) tried to fire up an effort to construct a new, representative and neat benchmark suite for Haskell, under the name HaBench. Since then, not much happened, except people kept contributing to and making (small) changes in the nofib suite. Unfortunately, nofib mainly consists of short running applications. Given the fact that Haskell programs come with a built-in runtime management system, chances are your measurements on this suite might be perturbed by the runtime more than you'd like.
The plan is to develop a framework for the new suite, modelled on nofib, but using best practices, such as criterion. The proposal on the main github repo reflects what I think should be considered in order to obtain a decent framework. The initial benchmarks that will take place in the framework will most likely be those from nofib, the language shootout, etc. We shall see where we can take this beast, but I'd love it if we could have some finalised version (0.9 :-) by the deadline for the next Haskell Symposium, so we can submit something there.