- Esc-team - mailman3.cels.anl.gov

Important ESC News
by Pete Beckman 01 Jul '11

01 Jul '11

Friends, Well, it seems the ESC review did not go well and the plan will not be funded. Rick got a call yesterday evening from Dan Hitchcock, who explained the reviews were not positive enough to move forward. A new joint ASCR/NNSA call for proposals would probably need to be issued to address exascale software. I'd like to thank everyone for their fantastic help over the last year building such strong collaborations between all the partners. I think we can be very proud of the ESC team and the plan we put together. While this is a setback for creating a new R&D community to focus on HPC open source software, I'm glad we proposed something daring, bold, and exciting. We pushed the limits --- proposing a new partnership between ASCR and NNSA, an improved model for identifying and attacking hard research problems and delivering them quickly on cutting-edge hardware, and a new plan for working with vendors to co-design architectures. Never before had such a clear and cohesive community plan for improving HPC software been built. However disappointing, I feel we did the right thing and the current plan is a tremendous achievement. Science is best when the outcome is unknown, the path perilous, the work gooey, and the team amiable and goodnatured. Thanks again for everyone's amazing and excellent work. Together we built an A+ team and audacious plan. It may be several weeks before we get official news and substantive feedback from the program office. However, as you hear informal news from reviewers, please share. -Pete

1 0

ESC Review this week
by Pete Beckman 27 Jun '11

27 Jun '11

Hello, As you know, this week the program offices (ASCR with a bit of help from NNSA) will review the ESC plan. Our understanding is that they are doing a panel, and we won't be providing presentations or participating -- we will simply be waiting anxiously. We offered several times to provide additional material, do a PPT overview, but Bill told us the material we already provided would be enough for a review. Please be attentive to email this week. If we get any questions from Bill or the review team, we may have some quick turn-around homework. Hopefully at SciDAC we will start to hear some quiet whispers about how the review panel went. -Pete

1 0

Improved instructions for additional materials
by Pete Beckman 01 Jun '11

01 Jun '11

Ok, after some discussions here on how best to do this, this is what we will do. We have 2 deadlines... Deadline late today / tomorrow early morning: 0) Jason is putting together the per-person budget. Folks at each site have already been working on this (it was not discussed on this email list yet). We got a 1 day extension, and can hand it in tomorrow, but all the numbers need to go to Jason today. So if you are one of those folks working to get Jason budget numbers and people rates, please finalize things quickly. Most sites are done, we are still waiting for numbers from UHouston, ORNL, some SNL numbers and some UIUC numbers, and LBNL. All other sites have completed this part of our exercise. If you don't have numbers for individual people, and are stuck there, please just give us an average rate, and we will note that and pass it forward, rather than leaving holes in the spreadsheet. Deadline Friday evening: 1) *EVERYONE* in the list below should pass us a 2 page CV (people who had 1/2 page bios in the original proposal). We will check it into the repo and build a single PDF to give Lucy and Bill. 2) Anyone who is listed as getting funds in the budget exercise that Jason has been working on (see above) should give us a C&P. I know, I had hoped to just have "PIs" do it, but I don't think that will work for the program office, so we will have to do the full exercise. If you have folks who are 100% on a project or NNSA funded, that makes it very easy. For example, the folks listed as QA folks already existing at labs probably are 100% funded on ASC or LCF projects, etc. So just note that to Jason with something like "Bob Smith: 100% ASC system software" or something like that. -Pete > al_geist > allen_malony > andy_white > barbara_chapman > barney_maccabe > barry_smith > bronis_de_supinski > dan_quinlan > david_skinner > franck_cappello > galen_shipman > gary_grider > george_bosilca > hank_childs > jack_dongarra > james_ahrens > jason_budd > jeff_hammond > jeffrey_nichols > jeffrey_vetter > jim_demmel > john_mellor-crummey > john_morrison > kamil_iskra > kathy_yelick > lee_ward > marc_snir > mike_heroux > nathan_tallent > pat_mccormick > paul_hargrove > paul_messina > pavan_balaji > pete_beckman > rajeev_thakur > ray_bair > rick_stevens > rob_falgout > rob_latham > rob_ross > ron_brightwell > rusty_lusk > sameer_shende > shane_canon > sudip_dosanjh > sunita_chandrasekaran > teresa_quinn > terry_moore > vinod_tipparaju > william_gropp > william_kramer > yili_zheng

1 1

Re: [Esc-team] Additional materials needed for ESC plan
by Pete Beckman 01 Jun '11

01 Jun '11

That is the traditional way, but given the short notice, I think the easier thing is to NOT prepare a special C&P with the current project listed. What do others think? -Pete On Jun 1, 2011, at 1:08 PM, John Mellor-Crummey wrote: > Pete, > > Do you want the support proposed under the ESC plan to be listed on the current and pending documents? > -- > John Mellor-Crummey Professor > Dept of Computer Science Rice University > email: johnmc(a)rice.edu phone: 713-348-5179 > > On Jun 1, 2011, at 11:25 AM, Pete Beckman wrote: > >> Some clarifications... >> >> The cc: list is: Jennifer Salazar <salazar(a)mcs.anl.gov>, Cheryl Drugan <cdrugan(a)mcs.anl.gov>, Gail Pieper <pieper(a)mcs.anl.gov>, Laura Wolf <wolf(a)mcs.anl.gov>, Jason R. Budd <jbudd(a)anl.gov>, Pete Beckman <beckman(a)mcs.anl.gov> >> >> Please send C&P and Bios to that group. >> >> Also, folks have asked who do we need to include for the C&P? I think it is reasonable to only include C&Ps for the 2 or 3 "PIs or Co-PIs" from your institution. We have avoided the "PI" label in the plan, since we have clearly named technical contacts, managers, and roles and responsibilities, and the teams are large. However, I think to respond to Lucy's request, we need to identify 2 or 3 people from each institution, rather than EVERY person who we included a short-bio for. >> >> Rob or I will send out a shorter list in a few moments for folks we really need the C&P for. >> >> -Pete >> >> _______________________________________________ >> Esc-team mailing list >> Esc-team(a)lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/esc-team >> >

1 0

Re: [Esc-team] Additional materials needed for ESC plan
by Pete Beckman 01 Jun '11

01 Jun '11

Some clarifications... The cc: list is: Jennifer Salazar <salazar(a)mcs.anl.gov>, Cheryl Drugan <cdrugan(a)mcs.anl.gov>, Gail Pieper <pieper(a)mcs.anl.gov>, Laura Wolf <wolf(a)mcs.anl.gov>, Jason R. Budd <jbudd(a)anl.gov>, Pete Beckman <beckman(a)mcs.anl.gov> Please send C&P and Bios to that group. Also, folks have asked who do we need to include for the C&P? I think it is reasonable to only include C&Ps for the 2 or 3 "PIs or Co-PIs" from your institution. We have avoided the "PI" label in the plan, since we have clearly named technical contacts, managers, and roles and responsibilities, and the teams are large. However, I think to respond to Lucy's request, we need to identify 2 or 3 people from each institution, rather than EVERY person who we included a short-bio for. Rob or I will send out a shorter list in a few moments for folks we really need the C&P for. -Pete

2 1

Additional materials needed for ESC plan
by Pete Beckman 01 Jun '11

01 Jun '11

Hello, This morning I had a phone call with Bill Harrod and Lucy to discuss their upcoming reading review of the ESC plan. During the call, Lucy asked that we put together full 2 page bios for everyone as well as C&Ps. So... yes, a last minute request.... We need to organize 2 page bios for everyone and Current-and-Pending for PIs/Co-PIs. In other news, as some of you know, we are finalizing some of the additional budget materials they requested. There are still 2 labs an a university that have not submitted the information to Jason so he can do the spreadsheet roll-up. Please get him your information today. For the 2 page bios and C&P, please send them to the people on the cc: list, and we will organize them into the repo. Lucy gave us until Friday evening to collect the 2 page bios and C&P, and over the weekend we are to PDF them together and send them to DC. -Pete

1 0

next steps
by Pete Beckman 14 Apr '11

14 Apr '11

Ok, Things are looking quite good indeed for our planning documents. The plan looks nice and it is very comprehensive. Thanks in advance to all the folks who read, corrected, rewrote, and then started all over again. At this point we have all the material together and are just updating figures, graphs, and correcting mistakes. Later today at ANL we will read through the entire plan, end to end, all 192 pages of it, and correct things as we go. We hope to finish up by the evening. The plan is due at high noon, eastern time, on Friday. So that means we will send mail a few minutes before 11:00 central on Friday, and then retire to food and drink. I encourage everyone to follow suit in their timezone. What's next? Well, since the DOE now has a budget, they are quickly deciding how to fund co-design centers, and other things. You may have heard they expect May 23 through 27 to be the site visit times for the co-design centers. I don't know yet what they plan for us, but after Bill Harrod looks through the plan, there will probably be some phone conversations. We hope to have a review at the end of May or beginning of June, so we can start by July 1st. At this point, if you are so inclined, download the PDFs and read them, sending mail to this entire list if you find bugs. On Friday, I'll send out confirmation that the plan was submitted, and then wait for instructions from Bill Harrod. -Pete

1 0

Getting Close
by Pete Beckman 13 Apr '11

13 Apr '11

The materials are getting very close to being ready. Thanks to everyone for their help. Hard to believe our first meeting was Feb 22, 2010 at O'hare. We have come a long way. I've checked in the draft PDF versions into ./esc/review/TotalPackageOutput. The sections still undergoing work.... Narrative: * Intro being tweaked * Last two sections being worked Everything else ready to be proofed. QA: Terri is doing some editing, should be ready for proof reading later today. SoftwareStack: Pretty much done, just double checking things. Ready for proof reading GapAnalysis: Pretty much done, going to double check Kamil's figures, should be ready for proof reading later today The Project management PDF that the PM holds is being edited by Ray and Jason, and will be PDFed shortly and added into the plan. If you want to see what folks have been working on, look through the ToDo.txt list (and previous versions) to see all the goodies. We are up to SVN revision 1010. We even had time to add an easter egg bonus... -Pete

1 0

Budget Pages
by Pete Beckman 12 Apr '11

12 Apr '11

I've been asked by several folks if we need to prepare budget pages. I'm going with no. I think the plan will be reviewed, and if they can spend money in 2011, they will update the current FWPs we have, etc. I don't think we should go to all the trouble to have budget pages created when DOE has to hack on budgets now anyway. The good news, is that the word we got from our DC guy is that in the latest budget deal, only about 2% was shaved off the 2010 budget for Office of Science. Of course, that is the 2010 level, not the increased 2011 numbers. Furthermore, the 2% could get unequally distributed across the Office of Science departments, of which ASCR is only one. However, compared to other areas of the budget, we *seem* (key word) to have done pretty well, since the White House really believed in innovation, research, and energy. -Pete

2 2

Standard Team Names
by Pete Beckman 12 Apr '11

12 Apr '11

We are trying to standardize some things in the documents. The team names for the technical work (in this order) are: Programming Models (PM) Application Programming Tools (TLS) Operating and Run-time Systems (ORS) Linear Algebra and Math Software (LAMS) Data Management and Analysis (DMA) I'll try and correct a bunch tonight, but if you see things in the morning that are not right, please fix them. -Pete

6 9