[d2n-analysis-talk] Output variables for farm replay

[Diana Parno]

Hello all,

I've implemented the BB.tr.n>0 cut that we discussed in yesterday's  
meeting, and it seems to work well. As expected, we get a reduction by  
a factor of 4.5 in the amount of disk space occupied by the replayed  
ROOT files for a sample run. Combined with the newly freed space on  
our work disk, I think it's worthwhile to try again at replaying a  
large portion of the 4-pass dataset. (The whole thing may have to wait  
for the new 5TB disk space, but we'll see.) Tonight I'm doing some  
test runs, and if they go well, I'll start more serious replays over  
the weekend.

I've attached a copy of the replay_farm_BB.odef file that I'm  
currently planning to use. I've pared it down a bit from the regular  
version, in hopes of saving some more disk space. I've left the old  
version commented out rather than deleted, so that you can see at a  
glance what I've removed. I'd be much obliged if you could take a  
quick look and see whether there's something I've left out that really  
ought to be left in. (Or, perhaps, if I've left something in that we  
really don't need.)

A quick summary of the changes:

- I've removed the bbite scaler tree in favor of the evbbite scaler  
tree, which has the same data but is synchronous with the regular  
datastream. bbite seemed redundant under these circumstances.

- I've added a number of epics variables, including beam energy and  
all the target variables that we were monitoring via logged  
stripcharts during the experiment.

- I've removed the BB golden track, which never worked properly;  
PriKineBB, which relied on the golden track; and the rastered beam  
package, since BB optics are not sensitive to raster information.

- I've removed several pieces of the DBB decoder data, e.g. Cerenkov  
ADCs.

- I've removed some of the BB.cer data, especially the TDC info from  
that class, since we're getting the multihit TDC data from the DBB  
data. Most of that multihit data can probably then be thrown out  
later, since it's tuned to provide us with one useful hit per event,  
but I'm not sure whether BB.cer.t draws from the right part of the  
multihit TDC buffer.

I'd appreciate any suggestions. I realize that we'll almost certainly  
go through multiple rounds of replay in the end, but it would be nice  
to start out with a good approximation of the data we'll need. Thanks!

Best,
Diana
-------------- next part --------------
A non-text attachment was scrubbed...
Name: replay_farm_BB.odef
Type: application/octet-stream
Size: 3638 bytes
Desc: not available
Url : https://mailman.jlab.org/pipermail/d2n-analysis-talk/attachments/20101203/d8fb6511/attachment.obj