Star-fst-l mailing list

[Tonko Ljubicic <tonko AT bnl.gov>]

Hello,

2 fresh items:

1) Updated documentation regarding pedestal calculation
and zero-suppression can be found here:

https://docs.google.com/document/d/1z9OgTVTCZ7JBdL5-49lBbnmfgTiH8m5lvJpd1I46-Rk/edit?usp=sharing

2) I optimized ZS a bit and I removed the CPU as the bottleneck.

The bottleneck now (as was before) is the raw data bandwidth over the
fiber links
which amounts to ~1450 Hz with 9 timebins.
So we can expect the max rate ~3.2 kHz with 4 timebins.

What I need from Gerard is the intrinsic frontend deadtime for the
9 and then for the 4 timebin case. I noted that 1000 RHIC
ticks is OK for 9 timebins and that a bit less is not but we
should have a precise number.