We are currently testing RCG2.8 on the front end systems which send their DAQ data to the second 10GE port on h1dc0. The full list of computers is:

h1oaf0, h1lsc0, h1asc0, h1pemmx, h1susauxey, h1iscey, h1susey, h1seiey, h1susauxex, h1iscex, h1susex, h1seiex.

All models running on these FEs were recompiled against the latest RCG trunk (the branch2.8 is a few days old, so we elected to compile against the trunk for this test). The exceptions are the SUSTMSX and SUSTMSY whose models are being worked on and did not compile, these were left at their 2.7.2 version.

The LVEA test stand was also left at 2.7.2 even though they are part of the block of FEs writing to this h1dc0 port.

The DAQ was reconfigured to remove the EDCU ini files associated with the upgraded models.

The front ends were restarted, no problems were observed.

We tested various slow (16Hz) channels which are now being transmitted along with the fast data over the DAQ network and not being acquired by the EDCU via the FrontEnd network, no problems have been observed.

The sus team are damping ETMX with the new version of h1susetmx with no problems being observed.

The ODC team checked their slow channels with no problems being seen.

As a demonstration of the synchronisity of the new 16Hz slow channels, the attachment shows 1 second of full data for a DuoTone slow channel for 2.8 (left plot) and 2.7.2 (right plot), showing true aliasing of the 960Hz Sine Wave signal with the new system.

We monitored the network bandwidth used by the second 10GE port, as measured by the switch, increased by 10% (200Mbps to 220Mbps).

The frame size did not increase beyond the noise level.