U.S. ATLAS HL-LHC Upgrade LAr Level 2 and Level 3 Managers Mailing List

[John Parsons <parsons AT nevis.columbia.edu>]

Hi all,

	As you can see from the thread below, the PO has given pre-approval for 
2 years of the proposed NYU effort on 6.4.3.  While that is all they 
feel they can commit to at this time, other aspects of our proposal will 
be revisited next year, once there is (hopefully) more clarity about the 
impacts of COVID, the project is further advanced, ...

	Andy, the next step is to get your NYU task list for 2 years put into 
P6.  You should revisit and update what you prepared earlier, shift for 
a slightly later start date for the funding (eg. Dec. 1).  Let us know 
when you have everything ready, and then we will talk with Penka to 
arrange a 1-on-1 meeting with Jenn or some other way to get things into 
P6.

        Regards,
                John

-------- Forwarded Message --------
Subject: Re: [Usatlas-hllhc-management-l] discussion of LAr priorities
Date: Mon, 12 Oct 2020 16:38:16 -0400
From: Gustaaf Brooijmans <gusbroo AT nevis.columbia.edu>
To: John Parsons <parsons AT nevis.columbia.edu>, 
usatlas-hllhc-management-l AT lists.bnl.gov 
<usatlas-hllhc-management-l AT lists.bnl.gov>
CC: Hong Ma <hma AT bnl.gov>


Hi John,


> OK, thanks.  I understand well that there exist significant 
> uncertainties now, and that it is very difficult for the PO to make 
> additional commitments.  So I appreciate the support for the NYU effort, 
> which we feel is really very strongly needed now; I will contact Andy to 
> let him know.  How do we proceed now with that?  Prepare a BCP?  Or work 
> with Penka/Jenn to get things into P6?  I assume both, but in what order?

The next step is P6.  The process is roughly 1) discuss with PO and get 
pre-approval (just done), 2) implement in P6, 3) check that P6 roughly 
matches anticipated impact, 4) CCB, 5) final approval.

> One question about the LSBs: do you have any comment about whether it 
> might be possible to incorporate this at some point in DOE scope, or is 
> it clear any future chance to include it must fall under NSF?  (or is it 
> just too early to know?)

It's too early.  It pretty much depends on what the agencies do about 
COVID.  DOE is *very* low now, contingency-wise, whereas NSF is low-ish.

Best,

Gustaaf


>      Regards,
>          John
>
> On 10/12/20 4:17 PM, Gustaaf Brooijmans wrote:
> > Hi John,
> >
> > Yes; we came to a decision in our management meeting earlier today.  
> > At this time, we're ready to proceed with NYU for a two year period.  
> > Given our contingency level, we're not in a position to commit to more 
> > than that at this time.
> >
> > Going forward we assume we'll need a decision on the LSBs sometime 
> > early next year, and about a year from now we can reassess the 
> > firmware situation to see if we can commit to more then.
> >
> > So this is the green light to implement two years of work at NYU in P6.
> >
> > Best,
> >
> > Gustaaf
> >
> > On 10/12/20 4:05 PM, John Parsons wrote:
> > > Hi all,
> > >
> > >      Just checking again that you all received this.  Let Hong and me 
> > > know how you would like to proceed.
> > >      Regards,
> > >          John
> > >
> > > On 10/1/20 4:31 PM, John Parsons wrote:
> > > > Hi all,
> > > >
> > > >      As requested, we have discussed within LAr the priorities for 
> > > > the various new efforts proposed.  Those efforts include the following:
> > > >
> > > > 1. NYU participation in BE firmware
> > > >      - cost ~$55k/year, beginning in FY21 = ~$250k over MREFC period
> > > >
> > > > 2. SMU participation in BE firmware
> > > >      - cost ~$90k/year for 0.5 FTE, beginning in FY22, increasing to 
> > > > $180k for 1 FTE over time = ~$315k - $540k over MREFC period 
> > > > (depending on speed of rampup from 0.5 to 1 FTE)
> > > >
> > > > 3. U Pittsburgh participation in new LSBs
> > > >      - cost ~$250k (based roughly on Phase I; Jim Mueller is working 
> > > > on a more definite estimate), assuming US provides 50% of LSBs 
> > > > (Saclay is expected to commit to providing other 50%)
> > > >
> > > >      As explained in the document previously submitted to you, the 
> > > > BE firmware increases are based on the Phase 1 experience, where the 
> > > > firmware team needed to be modified to be both deeper (ie. more 
> > > > experts) and broader (ie. more institutions).  We will not repeat 
> > > > those arguments here.  One update is that, since our last 
> > > > communication with you about the firmware, we have examined 
> > > > potential impacts on the BE entries in the RR. If the NYU effort and 
> > > > at least the fixed 0.5 FTE/yr at SMU are funded, we expect to reduce 
> > > > the cost impacts of f/w risks by a combined $250k.
> > > >      For the LSBs, this request comes as it was recognized that 
> > > > producing new LSBs, needed for the HL-LHC Level 0 trigger, was 
> > > > overlooked when defining fully the needed LAr upgrade.  There is now 
> > > > a proposal to add this to CORE, assuming the resources can be 
> > > > identified. Since U Pittsburgh provided all the LSBs for both the 
> > > > original construction and Phase 1, they are uniquely positioned to 
> > > > contribute also to the HL-LHC LSBs.
> > > >
> > > >      We have considered various possible approaches to reducing the 
> > > > overall cost impact on the MREFC.  We present below a summary of 
> > > > possible options, ranked from highest to lowest desirability:
> > > >
> > > > Option A (MREFC cost ~ $1040k)
> > > >      - fund all 3 efforts described above from NSF scope
> > > >
> > > > Option B (MREFC cost ~ $790k)
> > > >      - fund BE firmware efforts from MREFC, and fund LSB effort as 
> > > > part of DOE scope.  This makes logical sense, given the LSBs are 
> > > > responsible for handling the trigger sums from the PA/S chips, which 
> > > > are already part of DOE scope.  In addition, U Pittsburgh is a 
> > > > DOE-funded group. While there are DOE-funded groups who receive 
> > > > MREFC funding for NSF scope, recent communications from DOE make 
> > > > clear that, in case of resource limits in future, they would need to 
> > > > prioritize base funding for groups working on DOE scope as opposed 
> > > > to groups working on NSF MREFC scope; therefore it would be 
> > > > advantageous to fund the Pittsburgh effort from DOE scope.
> > > >      Note that this DOE increase would be in addition to the LAr 
> > > > electronics database work which would be done at BNL (and which 
> > > > therefore must be on the DOE side of the project)
> > > >
> > > > Option A2 (MREFC cost ~ $815k)
> > > >      - a variation of option A above, where the LSBs are funded from 
> > > > MREFC scope, would include also funding the NYU effort and funding 
> > > > the SMU effort at a fixed level of 0.5 FTE EE.
> > > >
> > > > Option C  (MREFC cost ~ $790k)
> > > >      - another alternative if the LSBs cannot be funded as part of 
> > > > DOE scope, is that then we would have to drop it, and instead 
> > > > prioritize the success of existing scope (ie. BE firmware) over 
> > > > picking up new scope. Note that not funding the LSBs would likely 
> > > > receive considerable pushback from ATLAS/LAr.
> > > >
> > > > Option B2  (MREFC cost ~ $565k)
> > > >      - a variant of option B above (ie. LSBs funded from DOE scope), 
> > > > if even further MREFC reductions were necessary, would be to fund 
> > > > the NYU effort and to fund the SMU effort at a fixed level of 0.5 
> > > > FTE EE.
> > > >
> > > >
> > > >      Some features to note:
> > > >
> > > > - the NYU effort is highest priority, for a number of reasons.  
> > > > Given that the BE firmware was well behind schedule, even during the 
> > > > preMREFC period before COVID, it is clear we need additional effort 
> > > > NOW.  Only NYU can provide that, and in a very cost-effective 
> > > > manner. In addition, to provide effective oversight and a better 
> > > > integrated effort across institutions, we need Andy to be more 
> > > > involved in the day-to-day BE work; as has been pointed out in past 
> > > > reviews, that is an unreasonable expectation if he has no local 
> > > > effort at NYU.  Also, adding (modest) funding to NYU would provide 
> > > > an additional NSF-funded institution on the MREFC.
> > > >
> > > > - the LSB production at Univ Pittsburgh is very well motivated, and 
> > > > we would very much like to include this effort.  However, adding 
> > > > this scope must take lower priority than ensuring the success of our 
> > > > current scope, including the BE firmware
> > > >
> > > > - while we would greatly prefer to fund the full SMU proposal, as 
> > > > you see we do consider options with SMU funded instead at a fixed 
> > > > level of 0.5 FTE EE.  This eventuality would give SMU at least 2 
> > > > years to negotiate the other 0.5 FTE from ATLAS Ops for Phase 1 M&O, 
> > > > or to find another funding source (eg. another project) for the 
> > > > other 0.5 FTE. It would also allow time for the upgrade to advance 
> > > > and evaluate later whether the funding level could be increased, 
> > > > depending on available contingency at that future time.
> > > >
> > > >      Let us know any questions/comments.
> > > >      Regards,
> > > >                 John and Hong