The RNAFold package

[Tags: gpl, library, program]

RNAfold v2 using the ADPfusion library. The RNAfold algorithm is used to determine how fast we can be compared to a highly optimized C program.

If possible, build using the GHC llvm backend, and GHC-7.2.2. GHC-7.4.x produces very bad code on my system, please benchmark using 7.2.2.

NOTE I'd like to rename this package to RNAfold, like the C implementation. Do not install globally, especially if you normally use RNAfold from the ViennaRNA package, for obvious reasons.

NOTE I am reluctant to call this v2 for now.

[Skip to ReadMe]


Change logNone available
DependenciesADPfusion (==, base (==4.*), BiobaseVienna (==, BiobaseXNA (==, mtl (>=2), primitive (==0.4.*), PrimitiveArray (==, strict (>=0.3.2), vector (==0.9.*) [details]
CopyrightChristian Hoener zu Siederdissen, 2010-2012
AuthorChristian Hoener zu Siederdissen (Haskell), Ivo L. Hofacker et al (ViennaRNA), 2010-2012
Home page
Source repositoryhead: git clone git://
UploadedSun May 20 16:49:47 UTC 2012 by ChristianHoener
Downloads1483 total (31 in last 30 days)
0 []
StatusDocs uploaded by user
Build status unknown [no reports yet]




llvmbuild using llvm backendEnabledAutomatic

Use -f <flag> to enable a flag, or -f -<flag> to disable that flag. More info


Maintainers' corner

For package maintainers and hackage trustees

Readme for RNAFold-

ViennaRNA RNAfold v2, MFE variant using the ADPfusion library


This algorithm is the second, and much larger, test case for ADPfusion. We implement "RNAfold v2" in the MFE variant using "-d2" dangles. Both a library version and an executable are created. The "RNAFold" binary expects single sequences, one per line. Backtracking tracks all co-optimal structures.


A simple "cabal update && cabal-dev install RNAFold" should be enough.

Runtime notes

Using Haskell and ADPfusion, we come to within x3-x4 for this package. Between the initial test case / submission (in I have traded in some performance improvements for much better readability in BioInf.RNAfold.Energy. The C version of RNAfold employs some other methods to improve performance. Consider:

base -~+ inner-1 +~- base base -~+ inner-2 +~- base

where it is advantageous to calculate the outer basepair only once, not twice as we are doing. It is probably better to try to improve the handling of fusioned code and/or final assembler generation than finding calculations common to different parts of CFG's.