On Fri, 25 Jul 2014 18:23:14 +0200 berenger.morel@neutralite.org wrote: > _ a lot of them are biased (at least in programming world): Not too much for FS (but it exist), you also must know how to read results. And you must read _some_ papers to make an average (not very different from programming: I've discovered that ~70% of time is dedicated to read docs). Some tests are also 'biased' but you're warned (eg: I'm testing to find the best FS for handling billions of small pictures). > I can remember having tried to read lot of bench about> > filesystems... I did not learned a lot, except that I was not > skilled enough to understand the raw data and often the conclusion > seemed biased to me, but this impression was maybe due to lacks of > skills, too. Most of benchmarks are made with bonnie++ that have a good doc. -- <You> Hi <Stranger> Hi <You> MOM WHAT THE FUCK ARE YOU DOING ON CHATROULETTE!??
Attachment:
signature.asc
Description: PGP signature