depristo d8fbda17ab O(N^2) bug found and removed -- very subtle and hard to find. ArrayLists underlying read backed pileups were being initialized with size() from the entire pileup up all samples, not the sample-specific sizes. So in 1000 samples at 4x, we were creating 1000 x 4000 element array lists, instead of 1000 x 4x element array lists. This fix results in a 2-3x speedup for 900 sample calling, and moves UG.map() back into the main CPU cost of UG with many samples. ... 900 samples in a single BAM: Release: 64.29 With sample-specific size: 24s - 35s git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5519 348d0f76-0448-11de-a6fe-93d51630548a		2011-03-26 12:38:19 +00:00
..
config	Provide a default logger, some config settings, and some doc updates.	2009-04-29 02:06:05 +00:00
src	O(N^2) bug found and removed -- very subtle and hard to find. ArrayLists underlying read backed pileups were being initialized with size() from the entire pileup up all samples, not the sample-specific sizes. So in 1000 samples at 4x, we were creating 1000 x 4000 element array lists, instead of 1000 x 4x element array lists. This fix results in a 2-3x speedup for 900 sample calling, and moves UG.map() back into the main CPU cost of UG with many samples.	2011-03-26 12:38:19 +00:00
test	Now uses the NO_HEADER option to avoid breaking MD5s due to changes in GATKArgumentCollection	2011-03-26 12:00:37 +00:00