Closed seoanezonjic closed 6 years ago
The old version of Haploclique did parallelization by simply partitioning the reads in several datasets, and then running haploclique with GNU parallel on every dataset. This meant that edges between reads in different partitions were not considered for the first iteration of clique enumeration, leading to some loss of information. After rewriting the original bash script handling the iterations directly into the C++ code, the parallelization part of the script didn't make it into the new version.
The remark about losing information in the old version is wrong.
Too bad that parallelization didn't make it.
On Apr 3, 2017 11:57 PM, "Bernhard Lang" notifications@github.com wrote:
The old version of Haploclique did parallelization by simply partitioning the reads in several datasets, and then running haploclique with GNU parallel on every dataset. This meant that edges between reads in different partitions were not considered for the first iteration of clique enumeration, leading to some loss of information. After rewriting the original bash script handling the iterations directly into the C++ code, the parallelization part of the script didn't make it into the new version.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/cbg-ethz/haploclique/issues/40#issuecomment-291286338, or mute the thread https://github.com/notifications/unsubscribe-auth/ABXexpfbDKL5WuWPbu6JhGeg9qE4PYMwks5rsWtmgaJpZM4Mw1vY .
I'm using the last release of hapliclique and I had seen that use pthreads. How can I set more than one thread to do the analysis? Thanks Pedro