Investigate discrepancy between non-BLAS and BLAS versions of `linfa-ica`

[YuhanLiin]

According to these results, all ICA benchmarks are noticibly faster with BLAS than without, though this is less severe at higher sample sizes. We should profile the non-BLAS benchmark runs (using something like flamegraph) and see if we can fix the discrepancy.

[oojo12]

Fast ICA.zip ICA flamegraphs. Later in the week I'll see if a text format is exportable.