Sebastian Hanß

Just a quick update: This turns out to be quite a rabbit hole ;-) I went a fair bit deeper and it seems that our patch finding algorithm is quite...

I'd say that's the plan. I'll hack some C++ code and you go for quality control?

just a few intermediate ideas and thoughts: - the new rcpp-code is half finished, I might come to push it on the weekend - R seems to be inherently not...

I kinda lost track, too ;-) I'll look into this next week.

> I would also like to add `focal` and `focalCpp` to the wishlist since focal operations can be computationally expensive (and I use them a lot sweat_smile). This would be...

``` r # remotes::install_github("r-spatialecology/landscapemetrics", ref = "mw_classes") library(landscapemetrics) library(raster) #> Loading required package: sp landscape

Sure, good idea!

@mhesselbarth This would produce much redundant code that may diverge over time and make it complicated to maintain the code. I see the benefit of having fewer function arguments for...

We could introduce a wrapper function that calls the RCpp functions with the argument like so:\ ```R # calculate all metrics on patch level (single threaded) calculate_lsm(landscape, level = "patch")...

I toyed around a bit on my laptop (i7 dual core with hyperthreading = 4 threads):  ncores = 0 is the original function, without any OMP; ncores 1-4 are...