It would be interesting to see benchmarks comparin...
# mathematicsc
corneil
10/17/2024, 1:09 PMIt would be interesting to see benchmarks comparing to other languages / tools that are being used for solving same problems.
a
altavir
10/17/2024, 3:38 PMIt is not possible to directly compare complex algorithms in different languages because all implementations are different. For simple operations we have a baseline: KMath "naive implementation" slower than BLAS-based Multik, but is faster than Tensorflow (which is also native).
c
corneil
10/21/2024, 11:20 AMIf you want people to consider using Kotlin for ML or data science then a piece of work analysing a set of data using numpy or some other library and kotlin library should be comparable.
a
altavir
10/22/2024, 6:11 AMAs soon as you start using complex computations it is not the case. For example, Numpy/Multik gives you an edge when working with large number arrays in memory, but it fails on function processing and streaming data analysis. When we talk about complex matrix operations like decomposition, it strongly depends not on the language, but on implementations.
7 Views