Hi Chris,

I was swamped with other urgent issues to solve. I remember that megatest
setup was in my todo list for quite a long time. I will setup as soon as I
finished some jobs.

Right now, there is no megatest for BlastDB reading/writing performance
checking. I need to add a simple time check when I open BlastDB. If it takes
more than a few seconds, there are problems.

I posted a bug report in Google Code. It has nothing to do with
building/indexing BlastDB. Just opening step.

Thanks,
Namshin Kim


On Wed, Oct 15, 2008 at 12:42 AM, Christopher Lee <[EMAIL PROTECTED]>wrote:

>
> Namshin,
> shouldn't our megatests have caught any such gross performance
> decrease?!?  I haven't seen a single megatest report in months, let
> alone any indication of a performance problem.  As I have said before,
> I am extremely concerned that we do not seem to be regularly running
> the extensive performance tests we have written.  Not running the test
> suite nightly is like having no test suite at all.
>
> Regarding this performance problem, please submit a bug report that we
> can reproduce, to the issue tracker, not just an email saying "BlastDB
> has gotten slow".  I've posted several times that people should file
> reproducible bug reports, which means
> - include an exact series of steps to reproduce the bug.  For example,
> based on your email I have no idea if the problem occurs when opening
> a FASTA sequence file for the first time (which requires building a
> shelve index file), or when opening an existing shelve index file...
> - include any data files needed to reproduce the bug
> - include precise build information (e.g. commit ID) for the version
> that has the performance problem.  Ideally please indicate the most
> recent version that DID NOT have the performance problem.
> - include detailed platform information (OS version info, memory size,
> etc.)
>
> This should be trivial to fix, if you can give us a reproducible.  As
> far as I know, nothing has changed in our seqdb indexing code, so this
> highlights how fragile and unreliable shelve / bsddb's performance
> is...  We do not do anything to load the index into memory.
>
> -- Chris
>
> >
>

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"pygr-dev" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/pygr-dev?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to