[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Is Churchill revolutionary?

>>>>> "Gert" == Gert Wollny <gw.fossdev@gmail.com> writes:

    Gert> On Wed, 2015-02-11 at 17:31 +0000, Jorge Sebastião Soares
    Gert> wrote:
    >> Hey,
    >> Maybe this is one more of those articles that have been published
    >> lately with the intent to verify the attention reviewers and
    >> journals pay to submissions:
    Gert> I don't think so, the paper appears in a BMC journal, and so
    Gert> far all review I got from other journals from this publisher
    Gert> were quite thorough. Besides, they really applied for a patent
    Gert> for the method:

    Gert> http://www.google.com/patents/US20130311106

    Gert> I skimmed the paper, and what they claim is that they found a
    Gert> parallelization strategy that gives close to optimal scaling
    Gert> which really seems to make a difference when using *lots* of
    Gert> processors. It's not my field of work, so I can't comment on
    Gert> the details.

Since I'm not a Bioinformtician, I ask myself: Are there tools in DebMed
(or otherwise free) to achieve the same results, even if currently not as
performing? If so, I'd be interested to look at them and see how they can
be optimized to achieve similar performance. If their tasks are
data-parallel a lot of the performance gain might come from a great HW setup.
What would be needed is some kind of benchmark everyone in the field is
familiar with.



http://www.q-leap.com / http://qlustar.com
          --- HPC / Storage / Cloud Linux Cluster OS ---

Reply to: