A subset of benchmarks (taken from the The Computer Language Benchmarks Game ) has been used to benchmark performance of Morfa programs. The benchmarking process followed the idea of The Computer Language Benchmarks Game and same caution at jumping to conclusions should be exercised with respect to the following results.
Judging based on these particular results, it seems that for the majority benchmarks Morfa delivers performance comparable with Mono C# and sometimes C/Java level in terms of CPU usage.
In terms of memory usage, Morfa performs very well except for the gc-intensive
Code from morfa-examples repository was used, see
Therein you can check the versions of codes from The Computer Language Benchmarks Game which were used to compare.
nbodyshould be further improved after an issue with
forloops is solved. It is estimated, that having done this the performance of these two benchmarks should level with C.
binarytreesis weak due to heavy use of GC, a known weakness to be improved in the future.
Tests were done using:
bencherscript of The Computer Language Benchmarks Game to perform measurements as well as
N(task size) and general approach taken from there