Decreasing running time when using multiple cores.
π‘ Possible solution:
Not sure, the example provided uses an block.splsda object because it is what I was interested in but looking at the code, it seems that the use of BiocParallel is not consistent across all the perf() variants.
The problem looks similar to what was previously reported about lack of parallelisation in the tune() function in #214
π Describe the bug:
perf() is not using multicore processing even with BPPARAM set correctly.
π reprex results from reproducible example including sessioninfo():
Results:
sessionInfo()
π€ Expected behavior:
Decreasing running time when using multiple cores.
π‘ Possible solution:
Not sure, the example provided uses an block.splsda object because it is what I was interested in but looking at the code, it seems that the use of BiocParallel is not consistent across all the perf() variants.
The problem looks similar to what was previously reported about lack of parallelisation in the tune() function in #214