HSF Weekly Meeting #104, August 17, 2017
This meeting was virtual, with contributed notes on CWP status and other matters.
News, general matters
CWP
-
Simulation
- The 9 pages long Simulation Executive Summary is done and available
- The 61 pages long Detector Simulation White Paper only needs addressing a few comments. These remaining comments are assigned to people who are slowly responding or are on vacation. We have sent them email reminders. (They are either the originators of the text or the ones with the expertise.)
- We want to bring to your attention the two main issues that still need to be resolved:
- Figure 9 in Sec. 8 of the Simulation White Paper has triggered a lot of questions. This plot is about the CPU needs for the 1st year of HL-LHC operations and includes collider data processing plus the whole simulation chain from generation through reconstruction. It states that the need in 2026 would be 60 times the resources utilized in 2016. The information came from Ian Bird who collected the numbers from the individual experiments (computing coordinators) and plans to update it in the fall for the LHCC. This plot is likely to appear in many documents so we need to decide where it goes and if it is included at all. Some experiments believe that the factor 60 is an overestimate so we may want to update it before including it in some introductory section. Opinions?
- Section 8.2 of the Simulation White Paper should include an apples-to-apples comparison between ATLAS and CMS computing performance numbers. The two experiments agreed on the samples, machine, operating system, compiler to use and the goal is to provide relative numbers (ratios) of time performance as different full simulation features or approximations are taken out with respect to default, as well as ratios of time performance for different HL-LHC detector configurations with respect to the Run 1 detector. No absolute numbers (in seconds) will be included. ATLAS has not come up with the numbers yet so we may add their numbers when they become available or decide to slightly rewrite this subsection to present it as a CMS example rather than a comparison between experiments. The decision to rewrite the subsection will be taken shortly, in the next few days.
-
MachineLearning
-
Software Development
-
Software Trigger and Event Reconstruction
-
Data Access, Organisation and Management
-
Data and Software Preservation
-
Data Analysis and Interpretation
-
Workflow and Resource Management
-
Visualization
-
Event/Data Processing Frameworks
- Liz and Benedikt met and they are happy with the state of the roadmap part of the document. Resolving comments is missing. Start from the roadmap and ensure that the points in the roadmap are well motivated in the rest of the document.
-
Careers, Staffing and Training
-
Computing Model
- 5-pages available for review. The main document is now frozen (no more comments or edits) and will be worked on over the next 2 weeks. Goal is a draft for review by the end of August.
-
Conditions Access
- Awaiting feedback from review.
-
Next Steps
Activity updates
AOB
- The proposed WLCG/HSF workshop week of 23-27 April 2018 has been occupied by the RRB so does not work. Alternative dates proposed are:
- w/c March 5th (seems not good for LHCb or ATLAS; slightly disfavoured by CMS)
- w/c March 12th (not good for ALICE)
- w/c March 26th (N.B. Friday 30 March is Good Friday; 4 day week)
- w/c April 9th (slightly disfavoured by CMS)
If you know of any further constraints or favour a particular week please say here or send an email to Graeme