Jon,
What events are you using to time an image. Upon your recommendation (or
maybe it was Robyn's), I'm using
stageleader.main.MPI_Barrier1.{start,end}. Is this what you are using?
It sounds like our numbers differ by 5-10 s per image.
My results are:
Time per image (s)
stage 0-4 stage 0-5
Consumerfull 82.8 +/- 1.9 138.5 +/- 3.1
Consumerless 12.6 +/- 1.9 15.7 +/- 2.3
cheers,
Ray
_______________________________________________
LSST-data mailing list
[email protected]
http://www.lsstmail.org/mailman/listinfo/lsst-data
