April 2015 - Acme-se - mailman3.cels.anl.gov

namelist diffs on master
by Foucar, James G 01 May '15

01 May '15

Hi all, I'm seeing some namelist diffs on master that I don't understand. Will anyone who has merged a PR to master in the last couple days look here: http://my.cdash.org/viewTest.php?onlynotrun&buildid=760165 ... and tell me if these are OK to bless? Thanks, -Jim

3 5

Re: [ACME-SE] [acme-team-leaders] ACME Executive Committee Decision on CIME
by Robert Jacob 20 Apr '15

20 Apr '15

(Fixing address of acme-se. In replies to past messages, please remove "-bounces" from the acme-se email On 4/20/15 11:45 AM, Bader, David C. wrote: > (Second response, with grammar corrected and several important additions) > > Bill, et al, > > This decision was not based on the value of CIME. We applaud NCAR and > CESM for recognizing this need after the Modeling Summit and encourage > CESM to continue its development, within the resources it has available. > > Our decision about joint participation in further development is based > on the likelihood that the Summit and Aurora architectures and software > stacks will make it difficult, if not impossible, to stay compatible > with CIME, and that software engineering resources will need to be > dedicated to those foreseeable needs ( with unforeseen specifics). In > our phone conversation, Mariana stated that most of the requirements for > compatibility between CESM and ACME in the near-to-intermediate term > could be accomplished by a one-time CIME adoption, with the possibility > of additions as they are released to the public CIME repository. > > Furthermore, we view CIME as a CSEG/CESM component, subject to the CESM > governance structure, and as such, not compatible with ACME governance. > > In our view, the decision does not have a negative impact on the > collaboration agreements we have established with NCAR as an > institutional partner in the ACME team. We anticipate continued close > communication and cooperation between CSEG and ACME SE's, which can take > place in the absence of agreement to collaborate on CIME. > > Please feel free to give me a call if you want to discuss this further. > > Dave > > David C. Bader (bader2(a)llnl.gov) > Lawrence Livermore National Laboratory > +1-925-422-4843(w) > +1-925-960-3981(c) > > > From: Bill Large <wily(a)ucar.edu <mailto:[email protected]>> > Date: Monday, April 20, 2015 8:37 AM > To: Dave Bader <bader2(a)llnl.gov <mailto:[email protected]>> > Cc: ACME-Leaders <acme-team-leaders(a)lists.llnl.gov > <mailto:[email protected]>>, > "acme-se-bounces(a)lists.mcs.anl.gov > <mailto:[email protected]>" > <acme-se-bounces(a)lists.mcs.anl.gov > <mailto:[email protected]>>, Jim Hurrell > <jhurrell(a)ucar.edu <mailto:[email protected]>>, Mariana Vertenstein > <mvertens(a)ucar.edu <mailto:[email protected]>>, Jean-Francois Lamarque > <lamar(a)ucar.edu <mailto:[email protected]>> > Subject: Re: [acme-team-leaders] ACME Executive Committee Decision on CIME > > ACME > > We are very disappointed with the ACME decision of at most a one-time > adoption of CIME as a framework to develop common climate modeling > infrastructure. We believe CIME is a very positive step toward sharing > infrastructure code and developments and to reduce duplication and we > very much appreciate the contributions of DOE/ACME SEs since inception. > > As you know, our intention is to put CIME forward to USGCRP/IGIM as a > positive outcome from the February Summit, and we had hoped ACME would > be on board. Nevertheless, we will move forward in engaging other > groups with CIME as a way to accelerate progress in the US climate > modeling community. > > Clearly ACMEï¿½s decision on CIME has ramifications for NCAR and CESM, but > we will need some time to assess the situation and how we can best move > forward. > > > Jim Hurrell, NCAR Director > > W.G. Large, CGD Laboratory Director > > J.-F. Lamarque, CESM Chief Scientist > > Mariana Vertenstein, CSEG Head > > > On Fri, Apr 17, 2015 at 9:53 AM, Bader, David C. <bader2(a)llnl.gov > <mailto:[email protected]>> wrote: > > TO: ACME Group Leaders > > Below is a copy of a post on the ACME-CIME discussion page: > > https://acme-climate.atlassian.net/wiki/pages/viewpage.action?pageId=201533… > > After a week of discussions concerning the status of various pieces > of the ACME modeling system, and the general uncertainties and > ambiguities associated with maintaining compatibility with CESM, the > Executive Committee, with concurrence from the Program Manager has > determined that the SE/CPL GLs should continue the evaluation of the > current public CIME and make a recommendation as to whether a > one-time adoption of CIME is in the best interests of ACME. > > Moving forward, ACME cannot make a commitment to stay compatible > with CIME changes, nor commit to joint CIME development with > NCAR/CSEG. Flexibility in our software to support machines such as > Aurora and Summit is required. Further, a principal of the project > has been no critical development paths outside of direct ACME control. > > > Dave > > David C. Bader (bader2(a)llnl.gov <mailto:[email protected]>) > Lawrence Livermore National Laboratory > +1-925-422-4843 <tel:%2B1-925-422-4843>(w) > +1-925-960-3981 <tel:%2B1-925-960-3981>(c) > > > > > -- > Dr. William G. Large > Associate Director, National Center for Atmospheric Research > Director, Climate and Global Dynamics Laboratory > phone (303) 497-1364; cell (720) 402-8993; fax (303) 497-1314

1 0

ACME Land Model Development unification
by Bader, David C. 17 Apr '15

17 Apr '15

TO: ACME Council, Land Group and SE/CPL Group The executive committee and Peter Thornton spoke today on a conference call regarding the multiple independent branches for land model development. To eliminate the ambiguity resulting from our prior decision to allow the development of independent branches, we decided to accelerate the time table to achieve a single merged model for development. Please see the discussion related to the Land Group on https://acme-climate.atlassian.net/wiki/display/CNCL/2015-04-13+Week+Execut… Dave David C. Bader (bader2(a)llnl.gov) Lawrence Livermore National Laboratory +1-925-422-4843(w) +1-925-960-3981(c)

1 0

another problem in our repo
by Foucar, James G 16 Apr '15

16 Apr '15

All, Looks like our repo is in a bad state again: % git merge jgfouca/scripts/fix_for_jenkins_script_on_blues Auto-merging models/atm/cam/bld/namelist_files/namelist_defaults_cam.xml CONFLICT (content): Merge conflict in models/atm/cam/bld/namelist_files/namelist_defaults_cam.xml Automatic merge failed; fix conflicts and then commit the result. My branch did not modify namelist_defaults_cam.xml Is there a commit on master that never made it to next? -Jim

5 14

Re: [ACME-SE] non-BFB commit on master
by Burrows, Susannah 10 Apr '15

10 Apr '15

Hi James, Can you clarify what your acme_bisect script is telling us? The commit you identified (adeddb4) was merged to master in adeddb4. Are you saying that the non-BFB differences in master were introduced by that merge commit? Thanks, Susannah From: <Singh>, Balwinder Singh <Balwinder.Singh(a)pnnl.gov<mailto:[email protected]>> Date: Monday, April 6, 2015 at 12:35 PM To: "Foucar, James G" <jgfouca(a)sandia.gov<mailto:[email protected]>>, "acme-se(a)lists.mcs.anl.gov<mailto:[email protected]>" <acme-se(a)lists.mcs.anl.gov<mailto:[email protected]>> Subject: Re: [ACME-SE] non-BFB commit on master This commit should be BFB when MAM4 is not turned on (it is turned off by default). It was BFB when I tested it in my branch. I don’t know what went wrong during the merging process. I will take a look and report back. Thanks, Balwinder From: Foucar, James G [mailto:[email protected]] Sent: Monday, April 06, 2015 12:28 PM To: acme-se(a)lists.mcs.anl.gov<mailto:[email protected]>; Singh, Balwinder Subject: non-BFB commit on master Hi all, I used my new acme_bisect script to find the commit that introduced the BFB change on master. It reported this commit: commit 1386774cb4035aa977deb740cf79cb044f772fa0 Author: Balwinder Singh <balwindersingh(a)gmail.com<mailto:[email protected]>> Date: Thu Aug 28 15:46:24 2014 -0700 Add MAM4 (4 aerosol modes: aitken, accu., coarse, primary carbon) List of changes: 1. MAM4 package (use -chem trop_mam4 while configuring the model) 2. Added a flag .fix_g1_err_ndrop. which when turned on (.true.) fixes a bug introduced by a repeated g1 equation in ndrop.F90 3. Added bug fix from NCAR for .create_production_test. script. 4. Added compiler debug flags for intel compiler and added .CCSM_BASELINE. in config_machine.xml file for PNNL-Cascade machine Model testing: 5. Model passes the exact restart test with MAM4 turned on. 6. Model passes b4b test when run in the default configuration. 7. Model ran fine for couple months with NAG compiler with full debugging in MAM4 configuration. And it caused this change: Comparing hist file with baseline hist file, /home/jgfouca/acme/scratch/ERS.f45_g37.B1850C5.melvin_gnu.C.1386774/run/ERS.f45_g37.B1850C5.melvin_gnu.C.1386774.cpl.hi.0001-01-12-00000.nc /home/jgfouca/acme/baselines/master/ERS.f45_g37.B1850C5.melvin_gnu/cpl.hi.nc ncdump1 done ncdump2 done comparing split files x[a-z][a-z] xaa 5132,5138c5132,5138 < 0, 0, 0, 0, 0, 0, 0, 0, 0.00479342661287574, 0.0402775963645312, < 0.017706122474612, 0.0264561009472818, 0.0658332477827422, < 0.0918150389464547, 0.149059009179461, 0.267001236695145, < 0.303681815485963, 0.268437849398892, 0.0483376970978501, 0, 0, \ 0, 0, 0, FAIL hist file comparison is FAIL Please advise. -Jim

4 22

Re: [ACME-SE] POC for missing input dataset on NERSC
by Taylor, Mark A 10 Apr '15

10 Apr '15

I see no reason for us to switch to an ACME specific inputdata directory, and I vote we continue to use /project/projectdirs/ccsm1/inputdata on NERSC. We've already made a similar decision for Titan and Mira. In general, the centers would prefer we not duplicate this ~1TB data set by having a copy just for ACME. That link below discusses the policy - any responsible developer should have write access to this directory. (The Titan admins disagree, as described in that confluence page). I don't know what the NERSC policy is, but Gautam should try and get write access, and update that confluence page with information on NERSC's policy. Mark -- Mark Taylor, mataylo(a)sandia.gov, +1-505-284-1874 1446 Multiphysics Applications Sandia National Labs, Albuquerque NM 87185-1320 From: jnjohnson <jnjohnson(a)lbl.gov<mailto:[email protected]>> Date: Wednesday, April 8, 2015 at 11:05 AM To: Gautam Bisht <gbisht(a)lbl.gov<mailto:[email protected]>> Cc: Benjamin Mayer <mayerbw(a)ornl.gov<mailto:[email protected]>>, "acme-se(a)lists.mcs.anl.gov<mailto:[email protected]>" <acme-se(a)lists.mcs.anl.gov<mailto:[email protected]>> Subject: [EXTERNAL] Re: [ACME-SE] POC for missing input dataset on NERSC Guys, It is true that we are using /project/projectdirs/ccsm1/inputdata on NERSC for ACME for the time being. You might be able to get the dataset(s) you’re missing from the ACME input data repository at Oak RIdge. Details about this repo and getting an account are at: https://acme-climate.atlassian.net/wiki/display/WORKFLOW/ACME+Input+Data+Re… Is that not presented as an option when you run the build script and it attempts to download them? Jeff On Apr 8, 2015, at 9:57 AM, Gautam Bisht <gbisht(a)lbl.gov<mailto:[email protected]>> wrote: Hmm... it seems this V5 of Qian forcing dataset isn't available on NCAR's svn input server. Since it is only need for one test, I'm not going to worry about it. -Gautam. On Wed, Apr 8, 2015 at 9:48 AM, Patrick Worley <worleyph(a)ornl.gov<mailto:[email protected]>> wrote: Almost all error messages, of the form: svn: E170000: URL 'https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata/atm/datm7/atm_forci…' doesn't exist On 4/8/15 12:42 PM, Gautam Bisht wrote: I want these datasets to go into "/project/projectdirs/ccsm1/inputdata" because I'm trying to create baselines for 57 tests (aux_clm45) using NCAR's clm4_5_r097 and then plan to compare those baselines with the ACME branch that brings in clm4_5_r097. -Gautam. On Wed, Apr 8, 2015 at 9:38 AM, Patrick Worley <worleyph(a)ornl.gov><mailto:[email protected]> wrote: I'd like a confirmation from "those who know" (whoever you are) that Gautam's datasets should go into /project/projectdirs/ccsm1/inputdata and not into /project/projectdirs/acme/data/inputdata Thanks. Pat On 4/8/15 12:11 PM, Patrick Worley wrote: I might. Point me at the script and I'll give it a try. Pat On 4/8/15 12:09 PM, Gautam Bisht wrote: Hi, On NERSC, there are few missing datasets in /project/projectdirs/ccsm1/inputdata. I'm wondering if there is someone on the ACME SE team has privileges to pull in the missing datasets. I can sent a shell script for pulling the missing datasets. Thanks, -Gautam. _______________________________________________ ACME-SE mailing list ACME-SE(a)lists.mcs.anl.gov<mailto:[email protected]>https://lists.mcs.anl.gov/mailman/listinfo/acme-se _______________________________________________ ACME-SE mailing list ACME-SE(a)lists.mcs.anl.gov<mailto:[email protected]>https://lists.mcs.anl.gov/mailman/listinfo/acme-se _______________________________________________ ACME-SE mailing list ACME-SE(a)lists.mcs.anl.gov<mailto:[email protected]> https://lists.mcs.anl.gov/mailman/listinfo/acme-se

3 6

dashboard change
by Foucar, James G 09 Apr '15

09 Apr '15

Hi all, Just a heads up: tests that only have a namelist diff will now show up as yellow, not red, on our dashboard. The column of these yellow tests is labelled "Not Run" which is a misnomer for us but there's no way I can see to rename the column. -Jim

1 0

ACLF CORAL machine announcement today
by Robert Jacob 09 Apr '15

09 Apr '15

Watch announcement of ALCF CORAL machine today. 11:15am CT. -------- Forwarded Message -------- Subject: Labcast: Major announcement live online Date: Thu, 9 Apr 2015 15:30:53 +0000 From: Argonne Newsroom <ArgonneNewsRoom(a)anl.gov> To: labcast(a)anl.gov <labcast(a)anl.gov> *Labcast: Major announcement live online* U.S. Department of Energy Under Secretary for Science and Energy Lynn Orr will make an important announcement involving Argonne today at 11:15 a.m. The announcement will be broadcast online at http://livestream.com/argonnelive. Argonne Director Peter Littlewood will join Orr for the announcement, along with 11^th District U.S. Rep. Bill Foster, 3^rd District U.S. Rep. Dan Lipinkski and14^th District U.S. Rep. Randy Hultgren. ----------------------------------------------------------------------------------------------------------------------------- Please share this message with co-workers who do not have email. Please do not reply to this message.

1 0

Re: [ACME-SE] namelist change on next
by Taylor, Mark A 09 Apr '15

09 Apr '15

Hi Jim, This is from Erika's pull request - not sure if she or the integrator are on the 'acme-se' email list, so you may not get a response. But in this case, I happen to know that Eirka's work added the ability to set 'zmconv_tau' in the name list, so it should change the name list. Mark -- Mark Taylor, mataylo(a)sandia.gov, +1-505-284-1874 1446 Multiphysics Applications Sandia National Labs, Albuquerque NM 87185-1320 From: <Foucar>, James G <jgfouca(a)sandia.gov<mailto:[email protected]>> Date: Tuesday, April 7, 2015 at 1:08 PM To: "acme-se(a)lists.mcs.anl.gov<mailto:[email protected]>" <acme-se(a)lists.mcs.anl.gov<mailto:[email protected]>> Subject: [EXTERNAL] [ACME-SE] namelist change on next Hi all, On next, we have ERS.f45_g37.B1850C5 failing with a namelist problem: In namelist 'zmconv_nl', found extra variable: 'zmconv_tau' FAIL: Namelist diff between files /home/jgfouca/acme/baselines/next/ERS.f45_g37.B1850C5.melvin_gnu/CaseDocs/atm_in and /home/jenkins/acme/scratch/jenkins/ERS.f45_g37.B1850C5.melvin_gnu.C.150407-013208/CaseDocs/atm_in Please advise, -Jim

1 0

Q3 reporting pages ready
by Robert Jacob 08 Apr '15

08 Apr '15

The draft pages for our Q3 reports are ready. https://acme-climate.atlassian.net/wiki/pages/viewpage.action?pageId=189409… There is a page for each component in our SE/CPL project including one for Support. We need words from Jim, Jeff, Doug, Jayesh, Steve and Jed by Thursday evening in the "Summary of Progress" section for component pages with your name on them. It should be short, no more then a paragraph. For the "Support" page, a couple of sentences on some support activity is fine. Andy and I will take care of word-smithing and the summary report. Thanks, Rob

1 0