NPPS Leadership Team

[Torre Wenaus <wenaus AT gmail.com>]

Comments much appreciated! This is a new version, following the wise advice to drop the travel part, and adding a paragraph about how additionally the validation suite can be used. With a few tweaks to perhaps make it clearer what I'm proposing.

Hi Jamie,
I have a proposal to make an addition to the EIC software effort (over and above Kolja, including his promotion), while at the same time activating the PHENIX DAP support.

I haven't forgotten that you have allocated 1/2 FTE for PHENIX data & analysis preservation :-) I haven't found a way to use it on an existing person. I have been looking for an opportunity to add a complementary 1/2 FTE to turn it into a full person we can hire, and I think there's a good case now for the other 1/2 FTE to be EIC software support. This new 1/2 FTE would be complemented by another 1/2 FTE of new EIC software support being requested to JLab by Markus (which would be applied to JLab based effort).

Here's why we need an additional FTE of EIC software support.

In EIC software, the YR developments and the most recent workshop last week have clarified our needs. The suite of simu tools we support in the software working group are accepted and in use across the YR working groups. It is clear that for the indefinite future we will have two full simu + reco frameworks, fun4all and escalate (g4e + jana2). Actually we have three because EicRoot is still getting a lot of use. On the fast simu side we support one, eic-smear, though some in the community are using delphes. What is clear now, and what we're getting requests for, is we need cross-validation between these tools. Between full and full, fast and full, and fast and fast if delphes catches on. Kolja is ready to help here, in the context of eic-smear; he can help both with the general infrastructure and applying it to eic-smear. But he can't do it all, and the software working group conveners are asking for help in the form of an FTE of new postdoc-level effort. This effort would be used to help develop the infrastructure (common format for kinematic outputs from simu and reco, standard set of validation ntuples and plots, validation workflows, etc.) and to perform the validation of our supported software packages and cross-validation between them.

This validation suite and infrastructure will have important value beyond validating the software itself. By enabling the comparison of full and fast simu based on physics performance metrics, it will provide quantitative means of assessing the relative utility and roles of fast and full simu, enabling us to optimize their use with their very different processung costs and latencies. It will also provide a means of objectively comparing detector models and monitoring the changes in their performance as models evolve.

Hence I am asking you for 1/2 FTE postdoc level support, which I could combine with the DAP support to hire someone. DAP and the validation work fit very well; validation is part of DAP. And Markus is asking JLab management for 1/2 FTE support.  The two labs collaborating to support an FTE worth of effort (across two people) on a task that is manifestly important and generic -- and which is all the more important because we have not converged on one framework, though the work would be needed even if we had -- would be a very good message of support for the immediate software needs of EIC.

Interested to hear what you think. This is a bit convoluted, I'm happy to clarify as needed!

  Torre

On Wed, May 27, 2020 at 12:33 PM Laycock, Paul <laycock AT bnl.gov> wrote:

Hi Torre,

I would completely drop the travel support request, it's not the right time and risks weakening your argument by mixing a high priority request with something not currently relevant.  When travel restarts we will be scrutinised more than before and there will be an assumption that we will travel less than before corona.  That may even require some prioritisation exercise, asking for new travel budgets in this context is bad timing.

For the software position, and thinking about the bottom line (you're discussing money) as described there is no guiding strategy.  It should be clear that by investing effort now in validation of these tools, we can more accurately assess future simulation needs.  By distinguishing between the use of full and fast sim based on physics performance needs, we will be able to optimise their use and the associated costs.

Paul

On 27 May 2020, at 11:59, Torre Wenaus via Phys-npps-mgmt-l <Phys-npps-mgmt-l AT lists.bnl.gov> wrote:

Hi all,

I plan to make the following proposal to Jamie following discussion among the EIC SW WG conveners. If you think it is misguided, or you have any comments at all, please let me know!

  Torre

Hi Jamie,
A couple questions about EIC software support. The deadline for new EIC detector R&D proposals is coming next month. Up to now we've relied on a small allocation to software consortium R&D to support travel. That account (which I manage now) is down to $7k or so. How can we replenish our travel funds? A new software R&D proposal is probably not the sensible way to do it. Can we get $20k or so (when we start to travel again) from your detector budget?

The other question is about effort. I haven't forgotten that you have allocated 1/2 FTE for PHENIX data & analysis preservation :-) I haven't found a way to use it on an existing person. I have been looking for an opportunity to add a complementary 1/2 FTE to turn it into a full person we can hire, and I think there's a good case now for the other 1/2 FTE to be EIC software support, with another 1/2 FTE of new EIC software support being requested to JLab by Markus (which would presumably be applied to JLab based effort). Here's why we need an additional FTE of EIC software support.

In EIC software, the YR developments and the most recent workshop last week have clarified our needs. The suite of simu tools we support in the software working group are accepted and in use across the YR working groups. It is clear that for the indefinite future we will have two full simu + reco frameworks, fun4all and escalate (g4e + jana2). Actually we have three because EicRoot is still getting a lot of use. On the fast simu side we support one, eic-smear, though some in the community are using delphes. What is clear now, and what we're getting requests for, is we need cross-validation between these tools. Between full and full, fast and full, and fast and fast if delphes catches on. We are going to ask Kolja to help here, in the context of eic-smear; he can help both with the general infrastructure and applying it to eic-smear. But he can't do it all, and the software working group conveners are asking for help in the form of an FTE of new postdoc-level effort. This effort would be used to help develop the infrastructure (common format for kinematic outputs from simu and reco, standard set of validation ntuples and plots, validation workflows, etc.) and to perform the validation of our supported software packages and cross-validation between them.

I am asking you for 1/2 FTE postdoc level support, which I could combine with the DAP support to hire someone. DAP and the validation work fit very well; validation is part of DAP. And Markus is asking JLab management for 1/2 FTE support.  The two labs collaborating to support an FTE worth of effort (across two people) on a task that is manifestly important and generic -- and which is all the more important because we have not converged on one framework, though the work would be needed even if we had -- would be a very good message of support for the immediate software needs of EIC.

Interested to hear what you think on these points.

  Torre

--

-- Torre Wenaus, BNL NPPS Group, ATLAS Experiment

-- BNL 510A 1-222 | 631-681-7892 |  wenaus AT gmail.com | npps.bnl.gov | wenaus.org

-- NPPS Mattermost room: https://chat.sdcc.bnl.gov/npps/channels/town-square

_______________________________________________
Phys-npps-mgmt-l mailing list
Phys-npps-mgmt-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/phys-npps-mgmt-l

--

-- Torre Wenaus, BNL NPPS Group, ATLAS Experiment

-- BNL 510A 1-222 | 631-681-7892 |  wenaus AT gmail.com | npps.bnl.gov | wenaus.org

-- NPPS Mattermost room: https://chat.sdcc.bnl.gov/npps/channels/town-square