Skip to content

Commit 5f7fe28

Browse files
graeme-a-stewartgraeme-a-stewart
authored
Steering group minutes 2024-10-14 (#1605)
* Steering group minutes 2024-10-14 * Spelling corrections Also for a GSOC blog
1 parent df402a7 commit 5f7fe28

File tree

2 files changed

+120
-1
lines changed

2 files changed

+120
-1
lines changed

‎_gsocblogs/2024/blog_SOFIE_VedantMehra.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -46,7 +46,7 @@ intro: |
4646
The Toolkit for Multivariate Analysis (TMVA) is a part of the ROOT Project, providing a machine learning framework used primarily in High-Energy Physics for tasks like multivariate classification and regression. Given the massive data output (often reaching petabytes daily) from research activities, and the need for fast, real-time predictions, a highly efficient inference engine is necessary for using trained machine learning models regularly. With this objective SOFIE (System for Optimized Fast Inference code Emit) was developed.
4747

4848
### SOFIE (System for Optimized Fast Inference code Emit)
49-
ROOT/TMVA SOFIE generates C++ functions easily invokable for the fast inference of trained neural network models. It takes ONNX model files as inputs and produces C++ header files that can be included and utilized in a “plug-and-go” style.
49+
ROOT/TMVA SOFIE generates C++ functions easily invocable for the fast inference of trained neural network models. It takes ONNX model files as inputs and produces C++ header files that can be included and utilized in a “plug-and-go” style.
5050

5151
```c
5252
// This generates a .hxx file containing the C++ code for model inference.

‎organization/_posts/2024/2024-10-14-steering.md‎

Lines changed: 119 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,119 @@
1+
---
2+
title: "HSF Steering Group Meeting, 14 October 2024"
3+
layout: plain_toc
4+
---
5+
6+
## Attending
7+
8+
Present/Contributing: Graeme Stewart, Stefan Roiser, Torre Wenaus, Eduardo Rodrigues, Mark Neubauer
9+
10+
Apologies: Andrea Valassi, Paul Laycock
11+
12+
## Discussions with Activity Areas
13+
14+
Any discussions to report on?
15+
16+
| Activity | SG Responsible | Met? |
17+
| ------------------------------------ | --------------- | ------------ |
18+
| Data Analysis | Eduardo + Mark | Not yet |
19+
| Detector Simulation | Torre + Graeme | This week |
20+
| Physics Generators | Liz + Stefan | Y |
21+
| JuliaHEP | Pere | Y (implicit) |
22+
| PyHEP | Eduardo | Y (implicit) |
23+
| Reconstruction and Software Triggers | Michel + Paul | |
24+
| Tools and Packaging | Liz + 1? | Y |
25+
| Training | Graeme + Nicole | Y |
26+
27+
Liz met with the Tools and Packaging conveners (as reported in last week's Coodination Meeting).
28+
29+
We should discuss regular slots that could be used for activity area meetings - topic for next coordination meeting.
30+
31+
Generators - Sapta had an idea for a follow-up to last year's workshop, looking at event generation on GPUs.
32+
33+
## European Strategy Update 2025
34+
35+
We devoted last week's coordination meeting to this topic (see [slides](https://indico.cern.ch/event/1355758/contributions/5708314/attachments/2945044/5175250/HSF%20Inputs%20from%20European%20Strategy%20Update.pdf)).
36+
37+
Intention is to cover as a core: Event generators, simulation, reconstruction, analysis, plus training and careers.
38+
39+
Those working groups that were there responded positively.
40+
41+
Our proposed timeline is:
42+
43+
- Now -> beginning of January: activity meetings to get the community feedback on the topic you mentioned
44+
- A one afternoon mini-workshop looks like it would be a good model
45+
- We can certainly allow people to submit short white papers as well
46+
- January : writing of chapter drafts
47+
- Feb 1-15: assembling of chapters, first paper draft
48+
- Feb 15 - March 1 : draft circulation and finalization
49+
- March: community endorsement (something like what we did for the CWP, asking for people to sign it)
50+
51+
Length is not set in stone, but we think about 3-5 pages per activity will be about right. 20-25 pages in total.
52+
53+
Michel and Graeme have followed up with the key activity areas to help generate momentum - there is not much time.
54+
55+
And additional help would be very welcome (Andrea is happy to contribute to the generator chapter).
56+
57+
## Seminar Series
58+
59+
[Indico](https://indico.cern.ch/category/18810/); Coordinators: Benedikt, Michel, ~~Andrea~~
60+
61+
We need an additional helper here!
62+
63+
Got an initial boost from JuliaHEP:
64+
65+
- 1 October, Julia in high-energy physics: a paradigm shift or just another tool?
66+
- 3 October, Julia as a Statically-Compiled Language
67+
68+
Future dates:
69+
70+
- 30 October (probably this one slips..)
71+
- Summaries/highlights of community meetings (JuliaHEP, Rucio, CHEP)?
72+
- Torre will suggest this to the seminar organisers
73+
- 27 November, Generators?
74+
- 29 January
75+
- 26 February
76+
77+
Nothing *confirmed* yet?
78+
79+
Note the [4D Tracking for Future Colliders Workshop](https://indico.slac.stanford.edu/event/8992) on November 7, 2024. Could provide some useful input on topics and/or speakers for a future HSF Seminar.
80+
81+
## HSF Affiliated Projects
82+
83+
The description of [HSF Affiliated Projects and Software](https://hepsoftwarefoundation.org/projects/affiliated.html) is now on the website. It motivates the activity and sets out some of the parameters for things like reviews, defining requirements and providing reference implementations.
84+
85+
The actual [guidelines for projects](https://hepsoftwarefoundation.org/projects/guidelines.html) are on a separate page.
86+
87+
- Projects (theirs or others) who would like to trial the process of becoming an affiliate?
88+
- Graeme: [prmon](https://github.com/HSF/prmon); [JetReconstruction.jl](https://github.com/JuliaHEP/JetReconstruction.jl)
89+
90+
Need a volunteer from the SG to check the criteria and report to the SG for approval.
91+
92+
- Eduardo will check prmon.
93+
- Ask Pere about JetReconstruction.jl.
94+
95+
## Advisory Group
96+
97+
The [Advisory Group mandate](https://hepsoftwarefoundation.org/organization/advisory-group.html) is published. Invitations have been sent to the large HEP experiments (ALICE, ATLAS, Belle II, DUNE, CMS, LHCb), WLCG and the theory community (MCnet).
98+
99+
So far positive responses from MCnet, LHCb and ALICE.
100+
101+
Intention would be to have a first AG meeting early next year.
102+
103+
## AOB
104+
105+
### Coordination Meetings
106+
107+
As ever... [chairing sign up document](https://docs.google.com/spreadsheets/d/1Z1Z4payCpieOLiVFcC6y9j-KCj71u6xX232LHUgIHfI/edit). We need one more volunteer for 2024, for 7 November.
108+
109+
And we have to start planning for next year!
110+
111+
### Domain
112+
113+
Torre made enquiries with BNL who seem to be willing to take over ownership of the domain. Any news?
114+
115+
- Procedure is established, but this will take a bit of time (there is a prior knowledge on this from <https://www.usatlas.org>).
116+
117+
### Next Meeting
118+
119+
Next meeting in 4 weeks time? 11 November. (US, French public holiday, so not ideal - we may poll.)

0 commit comments

Comments
 (0)