1. TDAQ Upgrade Software Plans John Baines, Tomasz Bold Contents: Future Framework Exploitation of future Technologies Work for Phase-II IDR

2. Future Framework ATLAS Future Framework project: Requirements capture completed: https://cds.cern.ch/record/1974156https://cds.cern.ch/record/1974156 Recommendations: Framework should support Trigger and offline evolution of athena (rather than replacement) based on Gaudi Minimise changes needed to algorithmic code i.e. maintain interfaces as far as possible Formal approval of project in progress Initial effort estimate: 110 staff-months effort to deliver core functionality by end 2016 Design/prototyping continuing in parallel: Workshop planed for week of 6 th July TDAQ plans: TDAQ effort focussed on Views demonstrator View provide RoI context for HLT algorithms Aims: athena scheduler + Views provide functionality of current Steering & Navigation Transparent to algorithms  offline algorithm can run unchanged in HLT Status: Currently working on stand-along demonstrator to test various use-cases Plan to have functional athena prototype by end of year

3. Timescales: driven by Run-3 readiness 2014 Q3Q4 LS 1 New Framework Delivered and production-ready Basic tests running in initial framework prototype Run 3 Framework Requirements Capture Complete 3 Decision on core technological solution Design, Prototyping, Initial Implement n Continued Development Refinement & further development as needed Complete migration Validate & Commission Migration of a limited set of algos. Migration of extended set of algos. Bug fixes, further refinement as needed HLT software Commissioning Complete Migration of Algorithms and Tools to new framework complete Framework Algs & Tools Core Framework Functionality delivered

4. Trigger GPU Demonstrator Aims: Implement ID tracking, Topoclustering & Muon tracking on GPU Integrate with athena (using APE) Assess potential performance gain - compare throughput of : i) standard TDAQ box (2U with 8 CPU) ii) GPU server (2U with 2 CPU 4 GPU) Investigate resource sharing schemes and optimum balance of GPU/CPU Inform: Estimate of effort to port & maintain GPU algorithms Decisions on future HLT farm hardware (for Run 3) Estimates of resources needed for Phase II with/wo accelerators Assess feasibility of GPU implementation of 100 kHz ID tracking Timescale: May 2015 : Export of data from athena to GPU modules - Integrate ID, Calo, Muon July 2015: Initial implementation of algorithmic code in GPU modules September 2015: Prototype completed ready for performance measurements December 2015: Initial evaluation complete & documented

5. Fix PC architecture Timescales: Framework, Steering & New Technologies 2014 Q3Q4 LS 1 Design & Prototype Implement core functionality Extend to full functionality CommissioningRun EvaluateImplement Infrastructure Exploit New. Tech. in Algorithms Speed up code, thread-safety, investigate possibilities for internal parallelisation Implement Algorithms in new framework. Narrow h/w choices e.g. Use or not GPU Run 3 Full menu complete Simple menu Framework New Tech. 5 Prototype with 1 or 2 chains New Framework Delivered and production- ready Basic tests running in initial framework prototype Framework Requirements Capture Complete Decision on core technological solution HLT software Commissioning Complete Migration of Algorithms and Tools to new framework complete incl. accelerator modules Core Framework Functionality delivered incl. accelerator support

6. Work for Phase-II IDR Event Filter software section of IDR will cover: Precision tracking in the Event Filter Options for 100(50)kHz full scan tracking at Phase-II Upgrades to selection software Upgrading the Event Selection software for new farm hardware Preparatory work: Updated estimate of Tracking times at Phase-II luminosity with and without FTK++ seeding Possible input from GPU demonstrator (tight timescale)

7. Effort