The SCROLL solution consists of 4 variants with different performance properties. In increasing order of performance, they are:
RSyncBDT, and RSyncBDT-unordered, roughly at the same level
RSyncBDD, and RSyncBDD-unordered, roughly at the same level
All solutions are faster than the provided ATL implementation.
However only small instances are transformed within the 10 min timeout.
On top of time and memory (which reports the number additional objects needed to perform the transformation), this approach also reports:
the number of leaves,
the number of subtrees,
the minimum, maximum and average depth of the BDT/BDD.
Currently the approach lacks a DSL, all transformations are written in Scala and thus are not easy to understand.
Adding examples of the different portions of code needed in the transformation would help understanding the code of the transformation.
The solution uses the provided benchmark structure.
The SCROLL solution consists of 4 variants with different performance properties. In increasing order of performance, they are:
All solutions are faster than the provided ATL implementation. However only small instances are transformed within the 10 min timeout.
On top of time and memory (which reports the number additional objects needed to perform the transformation), this approach also reports:
Currently the approach lacks a DSL, all transformations are written in Scala and thus are not easy to understand. Adding examples of the different portions of code needed in the transformation would help understanding the code of the transformation.
The solution uses the provided benchmark structure.