Executors.newCachedThreadPool().getMaximumPoolSize() return Integer.MAX_VALUE.
This caused some problems (number overflow) trying to parallelize Gauss convolution.
This PR makes some changes that improve the code for parallelization when the number
of threads is very high.
Fix https://github.com/imglib/imglib2-algorithm/issues/77
Executors.newCachedThreadPool().getMaximumPoolSize() return Integer.MAX_VALUE. This caused some problems (number overflow) trying to parallelize Gauss convolution. This PR makes some changes that improve the code for parallelization when the number of threads is very high.