Fig. 1
From: VariantSpark: population scale clustering of genotype information
Comparison of method and genome-wide scaling experiment. Left Runtime for clustering variants from chromosome 22 is given in seconds with 32 GB of memory on 8 threads (except for the pre-processing in R and Python where this was not supported). Right Scaling from 20 % to 100 % of variants in the genome with maximal number of executors and lowest possible memory assignment