thanks to all for your hints. Performance is crucial for my task and with R the 
biggest progress was achieved by vectorization of functions. The difference 
between sequential and parallel calculation was stunning. For a function with 
50 conditional assignments it took about 0.1s per call and was scaling 
linearly. Putting the input of 100 calls in vectors and than do all 
calculations with one call took 0.3s and expanding the length of the vectors to 
1000 took 0.4s. Maybe this is not the case with Nim.

Reply via email to