n+1 completely overlap. Gives a small but consistent performance increase in
non-intervaled whole exome traversals (2.79min original, 2.69min revised).
Needs a more in depth analysis of optimal shard sizing to determine a true
optimum.
Also renamed a variable because Khalid disapproved of my naming choices.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5595 348d0f76-0448-11de-a6fe-93d51630548a
- create size() method that returns an approximation of the uncompressed size in bytes of BAM span.
I'll use this method as a protoshard weighting function until we determine how to normalize the
weights across the different data access mechanisms (reads, reference, RODs).
- Implementations of basic union/intersection/subtraction mechanisms for BAM spans; should be enough
to get an accurate weight for two proto-shards put together.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5541 348d0f76-0448-11de-a6fe-93d51630548a
optimizer was mutating the data stored in the indices. Protosharding still
disabled by default.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5334 348d0f76-0448-11de-a6fe-93d51630548a
are works in progress, and this checkin will provide a baseline against which to gauge
improvements to both projects.
Low-memory BAM protoshards (disabled by default):
- Currently passing ValidatingPileupIntegrationTest.
- Gets progressively slower throughout the traversal, but should run at least as fast as original implementation.
- Uses 10+ file handles per BAM, but should use 3.
BAM performance microbenchmark test system:
- Currently tests performance of BAM reading using SAM-JDK vs. GATK
- Tests do not hit all GATK performance hotspots.
- New tests that require input data in a slightly different form are hard to implement.
- Output of test results is not easily parseable (investigating Google Caliper for possible improvements).
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5317 348d0f76-0448-11de-a6fe-93d51630548a
functionality for BAM index visualization.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5222 348d0f76-0448-11de-a6fe-93d51630548a
svn's logs:
- Copied BAM indexing engine from Picard back into the GATK anticipating
shard merging algorithm. Tried to leave most of the building blocks in
Picard. If this turns into a logistical nightmare, I'll merge the building
blocks into the GATK as well.
- Reorganized the org.broadinstitute.sting.gatk.datasources package, giving
better separation of query and management functionality for reads, ref, rmd,
and samples.
- Merged Shard building blocks into org.broadinstitute.sting.gatk.datasources.
reads package, indicating it's current strong relationship with the reads,
rather than the general unifying element I wish this would be.
- Collapsed BAMFormatAwareShard into Shard.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5184 348d0f76-0448-11de-a6fe-93d51630548a
codebase to a dependency on a jar file compiled from my private picard branch.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3034 348d0f76-0448-11de-a6fe-93d51630548a
infrastructure can be torn down:
1) New sharding system emulates old MonolithicSharding mechanism.
2) Better awareness of differences between fasta and BAM files when creating
shards.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2948 348d0f76-0448-11de-a6fe-93d51630548a
GATK users from the readers, which they could abuse in ways that could destabilize the GATK.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2923 348d0f76-0448-11de-a6fe-93d51630548a
the mode of operation is currently queryOverlapping rather than queryContained.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2899 348d0f76-0448-11de-a6fe-93d51630548a
support batched intervals in a single shard, but intervals are not yet compressed into a single
shard.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2730 348d0f76-0448-11de-a6fe-93d51630548a
hopefully I can find a way to kill these off before the Picard patch is ready.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2550 348d0f76-0448-11de-a6fe-93d51630548a
chunk space directly, stream data in block space, converting to chunk space
on demand.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2484 348d0f76-0448-11de-a6fe-93d51630548a
This commit can iterate over the BAM file, pulling out information about the blocks in the file without actually loading
or decompressing the reads.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2434 348d0f76-0448-11de-a6fe-93d51630548a
hanna
depristo