For general information see: https://uimon.cern.ch/twiki/bin/view/Atlas/AtlasTriggerDAQ Steering configuration (high) ------------------------------ (we have discussed this to some extent already) Join and eventually take over the work on trigger configuration in Athena by Johannes Haller - i.e. adapting the current TrigConfig (menus, sequences, steps and chains for the HLT steering) and TrigT1Config (for LVL1 simulation) to work from the new config database, to support the new requirements that come from the document we recently reviewed, and to provide tools to visualize and edit the configuration. Asses and develop a GUI panel for configuration if needed. Evaluate the different needs (physics analysis, trigger online configuration, etc...) and get a coherent system that can be used at all these levels. Monitoring of the steering (medium) ---------------------------------- Introduce monitoring into the steering in the form of scalars for the signatures at each step. Interface the information with the IS server or OH server (in case histograms are needed) and to the gatherer. Develop a system that allows to see the scalers per subfarm, node, or all together. Monitoring panels for HLT running (medium) -------------------------------------------- We will need a panel that shows these scalers in a graphical way Maybe also we need some interactivity on this panel (when some number becomes "bad" click to see the history, for example). The work will, at the end have to be connected to the configuration (HLT), where they also might like to be involved. It is something where experience from a running experiment can be profited. It is a project in itself and a product that can be seen as "sole" contribution from the group. A lot of the development/ideas can be done without the integration with the complicated tdaq ATLAS software (all the programming of the panel, etc...). Only need to interact with IS/confDB, at the end. So, can be done even from a remote site. Region Selector (low) --------------------- There are some small projects that need volunteers here, e.g. - modify to correctly handle time-dependent data (beginRun incident) - implement readout mapping and dynamic maps for muon detectors - write/retrieve maps from detector description database if available rather than computing them - RegionSelector for LVL1 calo trigger towers - requested for jet trigger studies, working with Kyle Cranmer (BNL). Trigger "persistency" (urgent, maybe unrealistic for people starting up) -------------------------- It would be great to have someone with more time take on a leading role in this. At the moment all the work is done (with the greatest respect to those of us involved) by people with only a few hours spare a week so we have lots of ideas but little implementation. These talks cover the area quite well: http://agenda.cern.ch/askArchive.php?base=agenda&categ=a051610&id=a051610s3t3/transparencies http://agenda.cern.ch/askArchive.php?base=agenda&categ=a055881&id=a055881s1t11/moreinfo And in summary, the aim is to provide the means to: - write trigger data to byte stream for seeding EF from L2 - write trigger data to byte stream from L2 and EF for debugging - write trigger data to ESD, AOD and TAG database It would involve working with Gianluca Comune on the steering aspects and PESA groups like e/gamma and tau/jet/etmiss for the analysis use cases, and the offline "Physics Analysis Tools" group for general offline persistency issues. It needs quite a lot of investment in time up front to get into Athena. Producing examples and documentation is part of the job. Extension of Test Suite for nightly tests of the HLT (high) ---------------------------------------------------------------- John Baines provides a lot of algorithm integration tests done in an offline environment. We need to perform tests for these same algorithms in the online environment, i.e. using AthenaMT or pt_test tools from the HLT release. These tests would be run as part of the HLT nightly build integration tests and could also be run in RTT (offline testing framework). Also, we would like to have small partitions running for these tests (in the nightlies with targetted Tdaq+off+HLT releases). For this last step, a tool that uses python to start partitions already exists -----> In essence is to make and extend the integration work running continuously in the nightlies -----> It is a job that can be done remotely in consultation with Jiri Masik -----> It will change when we use the RTT test framework and maybe when we use the tag collector -----> Gives a good overview of all the software packages and integration Stability tests (low) ----------------- Set up a system that can be left running for hours and hours. Maybe computing facilities at SLAC could be available for this? Once the system is running, try to "badly play with it". Make something crash, see what happens if we apply back-pressure, etc... Understand the behavior of the system to these incidents In this respect also find ways to monitor incidents (for example, realize when a L2PU crashes). It can lead to development of new tools/monitoring in a lot of areas Code Profiling (medium) ------------------ Devise and implement methods to profile the code. How to evaluate the timing spent in loading libraries and also in every algorithm in an online way. Interface results with the monitoring system. EF Infrastructure (medium or high depending on manpower in this area) ---------------------- Work in the area of interfacing the DataFlow with Athena (the so-called pt application). A lot of things need to be done here, like for example, adapting it to the state transitions, etc... Also providing state transition exerciser for use in the offline environment in which algorithms are developed.