[ome-devel] Feedback: write performance

[Melissa Linkert]

Hi Rubén,

> With bioformats 4.3 now 768 images convert in 6 minutes which means more than 2 images per second are read, processed and written.  This depends on conversion used (microscope format) but here the performance increase was 100%. 
> 
> The chosen solution was definitively the correct one, the larger buffers seemed to accelerate all the conversion process. 
> 
> Once again sincere thanks for your prompt reply and follow up of the incidence. 

Thanks for the feedback!  I am glad to hear that the new builds are now
suitably performant.

If any further problems arise, please do not hesitate to let us know.

Regards,
-Melissa

On Tue, Aug 30, 2011 at 09:57:11AM +0200, Rubén Muñoz wrote:
> Dear Melissa, Curtis & Josh, 
> 
> Following the resolution of the ticket #5965, I have been testing the new 4.3 to convert a medium size dataset (384 planes and 2 timepoints) to a network share.
> 
> The command generates multiple OME.TIFs with compression:
> 
> bfconvert -compression LZW source/experiment_descriptor.xml target/%n--T%t--Z%z--C%c.ome.tif
> 
> The connection to the server is not limiting because copy is faster but previous bioformats versions gave roughly one 1.5MB file (slice) per second converted.
> 
> With bioformats 4.3 now 768 images convert in 6 minutes which means more than 2 images per second are read, processed and written.  This depends on conversion used (microscope format) but here the performance increase was 100%. 
> 
> The chosen solution was definitively the correct one, the larger buffers seemed to accelerate all the conversion process. 
> 
> Once again sincere thanks for your prompt reply and follow up of the incidence. 
> 
> Great job!
> 
> Rubén
> _______________________________________________
> ome-devel mailing list
> ome-devel at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel