The system is structured as follows (coordinators are given in brakets):
DAQ
|
(John Zweizig & Dominik Ressing) |
SLT | (Mike Medinnis ) |
TLT,4LT ``Farm'' | (Andreas Gellrich ) |
SCS | (Ulrich Uwer ) |
These categories have many items in common and we hence view the subsystems
rather as being a common system and have common meetings and coordination.
An overview over the active projects in the common field gives the DAQ
Projects page. If you search the meaning of an abbreviation you can
find it in the
An overview and explanation of the System is in the now slightly outdated Architecture note. The up to date Architecture scheme is availlable as click-able inline image PS and EPS. There is also a recent paper rom DAQ96 and a talk from Beauty 97. The Data path software architecture is availlable as click-able inline image, as PS and EPS. The amount of data going into a specific trigger level, the reduction at every level and the detector components used are summarized in this table.
The main components are
The data path software is implemented and (partially) documented in the software list. Its backbone is a UDP based message passing system: RPM and a extension to the SHARC world: RPS in the KRNL library.
The effort here is to use as view as possible bus systems and sensor types. The main interface to the general system (from the individual systems) is the VME bus. Thus the main work is located in the subsystems. A partial list of used items is on the projects page.
The system software is mostely implemented and tested. An overview is given in the system breakdown.
The SLT will be build as a farm of Pentium Processors running Linux. The nodes are requesting data by way of a switch build out of SHARC boards from the Second Level Buffers. Since the nodes are of-the-shelf items the only homemade unit is a SHARC Link to PCI Interface card (apart from the SHARC boards which are already under heavy use in the DAQ system as SLBs). The whole system will be controlled by a further SHARC board. The system is described in th enow slightly outdated proposal.
The system oftware running on the nodes will be besides Linux a driver for the Interface card. This card interfaces the SHARC link protocoll to the PCI bus of th elinux PCs. The framework for the Algorithms running on the SLT is identical to the Simulation frame L2SIMU.
There is a recent overview talk, paper and an abstract submitet to RT97. Further information and pointers to the recent documents is availlable by way of the SLT Project page.
The TLT farm will be made out of the same hardware then the SLT farm. The system software will be identical. It will be interfaced to the 4th level farm by way of fast ethernet. Recent talks are enlisted on the farm homepage. For example form Beauty 97.
editors: D.Ressing Last modified: Fri Jul 3 01:26:11 MET DST 1998