Hi Jon,

This is great - just what we need.  I think we can finish up the report
with these numbers, and don't need a further dbingest run.   If there's an
argument to the contrary from anyone, let's hear it soon...

Tim

> Hi all,
>
> I did a short run with 8 nodes and 48 CCDs (8 nodes * 2 CPUs/node, so 3
> CCDs / CPU).  However, it was still missing dbIngest since I have no
> idea how to get that component working (of course, it was not a
> restrictively slow step in our experiments anyway).  Here are some
> numbers:
>
> Average Latency:  76.3180193468 sec
> Minimum Latency:  74.1756629944 sec
> Maximum Latency:  78.4021351337 sec
>
> -----------------------------------------------
> TOTALS:
>         Total Ops:              7.25660430726e+12
>         Total Runtime:          10629.051662 sec
>         Total GFlops/sec:       0.635827129392
> -----------------------------------------------
> Minimum GFlops:         0.05518
> Maximum GFlops:        1.0631
>
> Total time spent on I/O:  666704.838166 milliseconds
>
>
> And here are the numbers for the corresponding do-nothing run:
>
> Average Run Time:  9.65193891525
> Minimum Run Time:  7.21981501579
> Maximum Run Time:  14.6085150242
>
> Shall I get some screenshots as well, or would we be better waiting for
> one with dbIngest (note: I may be gone by that time).
>
> -Jon
>
>
> _______________________________________________
> LSST-data mailing list
> [email protected]
> http://www.lsstmail.org/mailman/listinfo/lsst-data
>

_______________________________________________
LSST-data mailing list
[email protected]
http://www.lsstmail.org/mailman/listinfo/lsst-data

Reply via email to