June 2015 - Allocations-admins - mailman3.cels.anl.gov

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by John Blaas 23 Jun '15

23 Jun '15

Granted 80k core hours as requested. - JB On Mon, Jun 22, 2015 at 11:17 AM, accounts(a)lcrc.anl.gov <accounts(a)lcrc.anl.gov> wrote: > Hello, > > A change in allocation has been requested: > > Requester: sinclair (Donald Sinclair) > Project: Lattice-QCD > Title: Lattice simulations of Conformal and Walking > Technicolor. > Description: We perform simulations to evaluate the functional integrals of QCD-like theories formulated on a discrete space-time lattice, to enable determination of the non-perturbative aspects these theories. These include the properties > of these theories at non-zero temperature, including the scales of confinement and chiral symmetry breaking, and such zero temperature properties as spectra (including the Higgs mass), decay constants and the running of the gauge > coupling constant. We are particularly interested in those theories where the coupling constant evolves very slowly, since these are candidate 'Walking Technicolor' theories. Related to these are theories with an infrared fixed point (conformal field theories). Our first goal is to differentiate between these two different types of behaviour for candidate theories. > > We are performing simulations of QCD-like theories which are models for Walking or Conformal Technicolor. We have been studying theories which are essentially QCD but with colour-sextet rather than colour-triplet quarks. We hope to measure the running of the QCD coupling constant. For 2 or 3 flavours, > 2-loop perturbation theory predicts an infrared fixed point. > > For 2 flavours, it is possible that a chiral condensate forms before this fixed point is reached. If so, the fixed point is avoided, the theory is confining, and chiral symmetry breaks spontaneously. However, there is a region where the coupling constant evolves very slowly. These are the properties required for a walking technicolor theory. Simulations we have performed so far at finite temperature suggest that this theory might walk. So far the results are inconclusive. > > For 3 flavours we know that the theory should be conformal. Simulations at N_t=6 (12^3 X 6 lattice) and N_t=8 (16^3 X 8 lattice), some of which used Fusion, did not yet show evidence of conformality. We are thus moving to N_t=12 (24^3 X 12 lattice). We started these simulations on Blues, and intend to continue them in FY2015 on Blues, while performing lower mass simulations on Edison at NERSC. > > Our sextet quark codes are based on our earlier triplet quark codes and use the RHMC simulation method. The Rational Hybrid Monte Carlo (RHMC) is a stochastic molecular dynamics algorithm. The functional integral of QCD is written as a partition function of a classical field theory evolving in a fictitious time. The determinant of the Dirac operator raised to a fractional power is calculated by introducing bosonic fields (pseudofermions), and sandwiching this Dirac operator raised to minus said fractional power between them. This fractional power of the Dirac operator is approximated to machine accuracy by a rational approximation. After defining this this theory on a discrete space-time lattice, the inversions required by the partial-fraction > expansion of the rational approximation are performed using Krylov space methods, in particular a multi-shift extension of the conjugate gradient algorithm. A global Metropolis Monte-Carlo accept/reject step applied at the > end of each trajectory removes discretization errors introduced by the numerical integration of these stochastic equations of motion. We parallelize the code by assigning a fixed number of adjacent lattice sites to each MPI > task. Network bandwidth ultimately limits how small a chunk of the lattice can be assigned to each task. > > Our 24^3 X 12 runs will be performed on 288 cores (18 nodes) of Blues. > > A short benchmark run of the 24^3 X 12 code yielded 667 Gflops or 2.3 Gflops/core. Since we have seen performances of > 3 Gflops/core on machines with similar processors, we suspect that at least 3 Gflops/core should be achievable on this machine. Since we have exhausted our allocation on Blues, we cannot perform more extensive tests. Scaling tests on Fusion and on Edison at NERSC for numbers of cores from 24 up to 288 cores indicate that the per core performance increases up to 144 or 288 cores, which we believe to be related to cache usage. > > Our new project, simulating QCD at finite mu using Complex Langevin with Gauge Cooling, will require some code development. Our old codes do not as yet have gauge cooling and our parallel code uses SHMEM rather than MPI for communication. Since the real Langevin is the limit of our Hybrid Molecular Dynamics code (the forerunner of the RHMC method) in which there is a single update per trajectory, most of what we have discussed for the RHMC applies, and will not be repeated here. We plan to use Blues and Carver at NERSC for the early, small-lattice simulations, 8^3 X 4 on 32 cores and 8^4 on 64 cores. Because of the smaller memory requirements, allowing more efficient cache usage, we expect to get a minimum performance of 3-4 Gflops/core for this code. > > Our requested allocation is based on running 1 288 core job half of the time at times when the NERSC machines are experiencing slowdowns for amall jobs. > > > Current: undetermined amount > Justification: We have not yet performed a detailed scaling analysis on Blues, however, we have such analyses performed on Fusion as well as Edison at NERSC. > For Fusion, for our QCD with sextet quarks code on a 24^3 X 12 lattice running on Fusion we observed the following performances: > 24 cores = 45 Gflops = 1.9 Gflops/core > 48 cores = 95 Gflops = 2.0 Gflops/core > 72 cores = 148 Gflops = 2.1 Gflops/core > 96 cores = 220 Gflops = 2.3 Gflops/core > 144 cores = 369 Gflops = 2.6 Gflops/core > 288 cores = 784 Gflops = 2.7 Gflops/core > For the same code and lattice size running on Edison at NERSC we observed the following performances. > 24 cores = 53 Gflops = 2.2 Gflops/core > 48 cores = 115 Gflops = 2.4 Gflops/core > 72 cores = 191 Gflops = 2.7 Gflops/core > 96 cores = 245 Gflops = 2.6 Gflops/core > 144 cores = 510 Gflops = 3.5 Gflops/core > 288 cores = 959 Gflops = 3.3 Gflops/core > We use a custom assignment of tasks to nodes in order to minimize communications. Earlier recoding reduced the number of global reductions in the routines, which use most of the CPU time, by a factor of 2. The node codes are transparently vectorizable, to allow use of SSE instructions. > > > Requested: 80000 > > A specific reason has been given: > Needed to complete the runs at the current set of parameters. This set of runs is 90% completed > > This needs to be approved and the final allocation amount decided upon. > > Thank You, > The LCRC Accounts System > _______________________________________________ > allocations-admins mailing list > allocations-admins(a)lcrc.anl.gov > https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins

1 0

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by Bair, Raymond A. 22 Jun '15

22 Jun '15

Since 80K is all that is needed to finish the work, then lets provide it. Ray ----------------------------------------- Ray Bair Computing, Environment, and Life Sciences Argonne National Laboratory and the University of Chicago TCS Building 240, Room 4122 9700 South Cass Avenue Argonne, IL 60439 email: rbair(at)anl.gov Phone: (630)252-5751 On 6/22/15, 11:17 AM, "accounts(a)lcrc.anl.gov" <accounts(a)lcrc.anl.gov> wrote: >Hello, > >A change in allocation has been requested: > > Requester: sinclair (Donald Sinclair) > Project: Lattice-QCD > Title: Lattice simulations of Conformal and Walking >Technicolor. > Description: We perform simulations to evaluate the functional >integrals of QCD-like theories formulated on a discrete space-time >lattice, to enable determination of the non-perturbative aspects these >theories. These include the properties >of these theories at non-zero temperature, including the scales of >confinement and chiral symmetry breaking, and such zero temperature >properties as spectra (including the Higgs mass), decay constants and the >running of the gauge >coupling constant. We are particularly interested in those theories where >the coupling constant evolves very slowly, since these are candidate >'Walking Technicolor' theories. Related to these are theories with an >infrared fixed point (conformal field theories). Our first goal is to >differentiate between these two different types of behaviour for >candidate theories. > >We are performing simulations of QCD-like theories which are models for >Walking or Conformal Technicolor. We have been studying theories which >are essentially QCD but with colour-sextet rather than colour-triplet >quarks. We hope to measure the running of the QCD coupling constant. For >2 or 3 flavours, >2-loop perturbation theory predicts an infrared fixed point. > >For 2 flavours, it is possible that a chiral condensate forms before this >fixed point is reached. If so, the fixed point is avoided, the theory is >confining, and chiral symmetry breaks spontaneously. However, there is a >region where the coupling constant evolves very slowly. These are the >properties required for a walking technicolor theory. Simulations we have >performed so far at finite temperature suggest that this theory might >walk. So far the results are inconclusive. > >For 3 flavours we know that the theory should be conformal. Simulations >at N_t=6 (12^3 X 6 lattice) and N_t=8 (16^3 X 8 lattice), some of which >used Fusion, did not yet show evidence of conformality. We are thus >moving to N_t=12 (24^3 X 12 lattice). We started these simulations on >Blues, and intend to continue them in FY2015 on Blues, while performing >lower mass simulations on Edison at NERSC. > >Our sextet quark codes are based on our earlier triplet quark codes and >use the RHMC simulation method. The Rational Hybrid Monte Carlo (RHMC) is >a stochastic molecular dynamics algorithm. The functional integral of QCD >is written as a partition function of a classical field theory evolving >in a fictitious time. The determinant of the Dirac operator raised to a >fractional power is calculated by introducing bosonic fields >(pseudofermions), and sandwiching this Dirac operator raised to minus >said fractional power between them. This fractional power of the Dirac >operator is approximated to machine accuracy by a rational approximation. >After defining this this theory on a discrete space-time lattice, the >inversions required by the partial-fraction >expansion of the rational approximation are performed using Krylov space >methods, in particular a multi-shift extension of the conjugate gradient >algorithm. A global Metropolis Monte-Carlo accept/reject step applied at >the >end of each trajectory removes discretization errors introduced by the >numerical integration of these stochastic equations of motion. We >parallelize the code by assigning a fixed number of adjacent lattice >sites to each MPI >task. Network bandwidth ultimately limits how small a chunk of the >lattice can be assigned to each task. > >Our 24^3 X 12 runs will be performed on 288 cores (18 nodes) of Blues. > >A short benchmark run of the 24^3 X 12 code yielded 667 Gflops or 2.3 >Gflops/core. Since we have seen performances of > 3 Gflops/core on >machines with similar processors, we suspect that at least 3 Gflops/core >should be achievable on this machine. Since we have exhausted our >allocation on Blues, we cannot perform more extensive tests. Scaling >tests on Fusion and on Edison at NERSC for numbers of cores from 24 up to >288 cores indicate that the per core performance increases up to 144 or >288 cores, which we believe to be related to cache usage. > >Our new project, simulating QCD at finite mu using Complex Langevin with >Gauge Cooling, will require some code development. Our old codes do not >as yet have gauge cooling and our parallel code uses SHMEM rather than >MPI for communication. Since the real Langevin is the limit of our Hybrid >Molecular Dynamics code (the forerunner of the RHMC method) in which >there is a single update per trajectory, most of what we have discussed >for the RHMC applies, and will not be repeated here. We plan to use Blues >and Carver at NERSC for the early, small-lattice simulations, 8^3 X 4 on >32 cores and 8^4 on 64 cores. Because of the smaller memory requirements, >allowing more efficient cache usage, we expect to get a minimum >performance of 3-4 Gflops/core for this code. > >Our requested allocation is based on running 1 288 core job half of the >time at times when the NERSC machines are experiencing slowdowns for >amall jobs. > > > Current: undetermined amount >Justification: We have not yet performed a detailed scaling analysis on >Blues, however, we have such analyses performed on Fusion as well as >Edison at NERSC. >For Fusion, for our QCD with sextet quarks code on a 24^3 X 12 lattice >running on Fusion we observed the following performances: > 24 cores = 45 Gflops = 1.9 Gflops/core > 48 cores = 95 Gflops = 2.0 Gflops/core > 72 cores = 148 Gflops = 2.1 Gflops/core > 96 cores = 220 Gflops = 2.3 Gflops/core >144 cores = 369 Gflops = 2.6 Gflops/core >288 cores = 784 Gflops = 2.7 Gflops/core >For the same code and lattice size running on Edison at NERSC we observed >the following performances. > 24 cores = 53 Gflops = 2.2 Gflops/core > 48 cores = 115 Gflops = 2.4 Gflops/core > 72 cores = 191 Gflops = 2.7 Gflops/core > 96 cores = 245 Gflops = 2.6 Gflops/core >144 cores = 510 Gflops = 3.5 Gflops/core >288 cores = 959 Gflops = 3.3 Gflops/core >We use a custom assignment of tasks to nodes in order to minimize >communications. Earlier recoding reduced the number of global reductions >in the routines, which use most of the CPU time, by a factor of 2. The >node codes are transparently vectorizable, to allow use of SSE >instructions. > > > Requested: 80000 > >A specific reason has been given: >Needed to complete the runs at the current set of parameters. This set of >runs is 90% completed > >This needs to be approved and the final allocation amount decided upon. > >Thank You, >The LCRC Accounts System >_______________________________________________ >allocations-admins mailing list >allocations-admins(a)lcrc.anl.gov >https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 22 Jun '15

22 Jun '15

Hello, A change in allocation has been requested: Requester: sinclair (Donald Sinclair) Project: Lattice-QCD Title: Lattice simulations of Conformal and Walking Technicolor. Description: We perform simulations to evaluate the functional integrals of QCD-like theories formulated on a discrete space-time lattice, to enable determination of the non-perturbative aspects these theories. These include the properties of these theories at non-zero temperature, including the scales of confinement and chiral symmetry breaking, and such zero temperature properties as spectra (including the Higgs mass), decay constants and the running of the gauge coupling constant. We are particularly interested in those theories where the coupling constant evolves very slowly, since these are candidate 'Walking Technicolor' theories. Related to these are theories with an infrared fixed point (conformal field theories). Our first goal is to differentiate between these two different types of behaviour for candidate theories. We are performing simulations of QCD-like theories which are models for Walking or Conformal Technicolor. We have been studying theories which are essentially QCD but with colour-sextet rather than colour-triplet quarks. We hope to measure the running of the QCD coupling constant. For 2 or 3 flavours, 2-loop perturbation theory predicts an infrared fixed point. For 2 flavours, it is possible that a chiral condensate forms before this fixed point is reached. If so, the fixed point is avoided, the theory is confining, and chiral symmetry breaks spontaneously. However, there is a region where the coupling constant evolves very slowly. These are the properties required for a walking technicolor theory. Simulations we have performed so far at finite temperature suggest that this theory might walk. So far the results are inconclusive. For 3 flavours we know that the theory should be conformal. Simulations at N_t=6 (12^3 X 6 lattice) and N_t=8 (16^3 X 8 lattice), some of which used Fusion, did not yet show evidence of conformality. We are thus moving to N_t=12 (24^3 X 12 lattice). We started these simulations on Blues, and intend to continue them in FY2015 on Blues, while performing lower mass simulations on Edison at NERSC. Our sextet quark codes are based on our earlier triplet quark codes and use the RHMC simulation method. The Rational Hybrid Monte Carlo (RHMC) is a stochastic molecular dynamics algorithm. The functional integral of QCD is written as a partition function of a classical field theory evolving in a fictitious time. The determinant of the Dirac operator raised to a fractional power is calculated by introducing bosonic fields (pseudofermions), and sandwiching this Dirac operator raised to minus said fractional power between them. This fractional power of the Dirac operator is approximated to machine accuracy by a rational approximation. After defining this this theory on a discrete space-time lattice, the inversions required by the partial-fraction expansion of the rational approximation are performed using Krylov space methods, in particular a multi-shift extension of the conjugate gradient algorithm. A global Metropolis Monte-Carlo accept/reject step applied at the end of each trajectory removes discretization errors introduced by the numerical integration of these stochastic equations of motion. We parallelize the code by assigning a fixed number of adjacent lattice sites to each MPI task. Network bandwidth ultimately limits how small a chunk of the lattice can be assigned to each task. Our 24^3 X 12 runs will be performed on 288 cores (18 nodes) of Blues. A short benchmark run of the 24^3 X 12 code yielded 667 Gflops or 2.3 Gflops/core. Since we have seen performances of > 3 Gflops/core on machines with similar processors, we suspect that at least 3 Gflops/core should be achievable on this machine. Since we have exhausted our allocation on Blues, we cannot perform more extensive tests. Scaling tests on Fusion and on Edison at NERSC for numbers of cores from 24 up to 288 cores indicate that the per core performance increases up to 144 or 288 cores, which we believe to be related to cache usage. Our new project, simulating QCD at finite mu using Complex Langevin with Gauge Cooling, will require some code development. Our old codes do not as yet have gauge cooling and our parallel code uses SHMEM rather than MPI for communication. Since the real Langevin is the limit of our Hybrid Molecular Dynamics code (the forerunner of the RHMC method) in which there is a single update per trajectory, most of what we have discussed for the RHMC applies, and will not be repeated here. We plan to use Blues and Carver at NERSC for the early, small-lattice simulations, 8^3 X 4 on 32 cores and 8^4 on 64 cores. Because of the smaller memory requirements, allowing more efficient cache usage, we expect to get a minimum performance of 3-4 Gflops/core for this code. Our requested allocation is based on running 1 288 core job half of the time at times when the NERSC machines are experiencing slowdowns for amall jobs. Current: undetermined amount Justification: We have not yet performed a detailed scaling analysis on Blues, however, we have such analyses performed on Fusion as well as Edison at NERSC. For Fusion, for our QCD with sextet quarks code on a 24^3 X 12 lattice running on Fusion we observed the following performances: 24 cores = 45 Gflops = 1.9 Gflops/core 48 cores = 95 Gflops = 2.0 Gflops/core 72 cores = 148 Gflops = 2.1 Gflops/core 96 cores = 220 Gflops = 2.3 Gflops/core 144 cores = 369 Gflops = 2.6 Gflops/core 288 cores = 784 Gflops = 2.7 Gflops/core For the same code and lattice size running on Edison at NERSC we observed the following performances. 24 cores = 53 Gflops = 2.2 Gflops/core 48 cores = 115 Gflops = 2.4 Gflops/core 72 cores = 191 Gflops = 2.7 Gflops/core 96 cores = 245 Gflops = 2.6 Gflops/core 144 cores = 510 Gflops = 3.5 Gflops/core 288 cores = 959 Gflops = 3.3 Gflops/core We use a custom assignment of tasks to nodes in order to minimize communications. Earlier recoding reduced the number of global reductions in the routines, which use most of the CPU time, by a factor of 2. The node codes are transparently vectorizable, to allow use of SSE instructions. Requested: 80000 A specific reason has been given: Needed to complete the runs at the current set of parameters. This set of runs is 90% completed This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

Re: [allocations-admins] Project Allocation Request - NeutrinoSim
by John Blaas 19 Jun '15

19 Jun '15

Allocation of 100k core hours granted. - JB On Fri, Jun 19, 2015 at 10:04 AM, Bair, Raymond A. <rbair(a)anl.gov> wrote: > NeutrinoSim seems to have a plan for their time, so the allocation request > is reasonable. However they will likely need our help to bundle up 1-node > jobs so they can get enough throughput to use the time. > > Ray > > > > From: Guang Yang <gyang9(a)hawk.iit.edu> > Date: Wednesday, June 17, 2015 at 4:05 PM > To: Ray Bair <rbair(a)anl.gov> > Subject: Re: Project Allocation Request - NeutrinoSim > > Hi Raymond, > What code do you run? For the runs you still need > to make in FY2015, how many nodes will each job use? Do you use all the > cores on each node? How long will each job typically take? How many jobs > are needed? > > There are two kinds of codes I will run: 1. simulation to generate neutrino > MC. 2. minimization code to fit the data to MC. > But both of them are parallel jobs. As John told me, there are 16 cores in > each node, so I usually use the 16 cores in parallel. For each job in one > core, it takes roughly 2 hours. I will submit roughly 100 jobs(1,600 cores > needed) at one time and for one day, maybe 2 job submissions. So I might > need 6,400 core hours one day and this will take for few weeks. > > Thanks, > Guang

1 0

Re: [allocations-admins] Project Allocation Request - NeutrinoSim
by Bair, Raymond A. 19 Jun '15

19 Jun '15

NeutrinoSim seems to have a plan for their time, so the allocation request is reasonable. However they will likely need our help to bundle up 1-node jobs so they can get enough throughput to use the time. Ray From: Guang Yang <gyang9(a)hawk.iit.edu<mailto:[email protected]>> Date: Wednesday, June 17, 2015 at 4:05 PM To: Ray Bair <rbair(a)anl.gov<mailto:[email protected]>> Subject: Re: Project Allocation Request - NeutrinoSim Hi Raymond, What code do you run? For the runs you still need to make in FY2015, how many nodes will each job use? Do you use all the cores on each node? How long will each job typically take? How many jobs are needed? There are two kinds of codes I will run: 1. simulation to generate neutrino MC. 2. minimization code to fit the data to MC. But both of them are parallel jobs. As John told me, there are 16 cores in each node, so I usually use the 16 cores in parallel. For each job in one core, it takes roughly 2 hours. I will submit roughly 100 jobs(1,600 cores needed) at one time and for one day, maybe 2 job submissions. So I might need 6,400 core hours one day and this will take for few weeks. Thanks, Guang On Wed, Jun 17, 2015 at 4:34 PM, Bair, Raymond A. <rbair(a)anl.gov<mailto:[email protected]>> wrote: Dear Guang Yang, Please provide some more details about this project that are missing from the project request. What code do you run? For the runs you still need to make in FY2015, how many nodes will each job use? Do you use all the cores on each node? How long will each job typically take? How many jobs are needed? Best regards, Ray On 6/16/15, 11:45 AM, "accounts(a)lcrc.anl.gov<mailto:[email protected]>" <accounts(a)lcrc.anl.gov<mailto:[email protected]>> wrote: >Hello, > >A change in allocation has been requested: > > Requester: yangg (Guang Yang) > Project: NeutrinoSim > Title: Neutrino Simulation > Description: Submitting simulation jobs to simulate reactor >antineutrino detector response by using offline softwares. > Current: undetermined amount >Justification: > > Requested: 100000 > >A specific reason has been given: >Massive calculation work will come soon > >This needs to be approved and the final allocation amount decided upon. > >Thank You, >The LCRC Accounts System >_______________________________________________ >allocations-admins mailing list >allocations-admins(a)lcrc.anl.gov<mailto:[email protected]> >https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 16 Jun '15

16 Jun '15

Hello, A change in allocation has been requested: Requester: yangg (Guang Yang) Project: NeutrinoSim Title: Neutrino Simulation Description: Submitting simulation jobs to simulate reactor antineutrino detector response by using offline softwares. Current: undetermined amount Justification: Requested: 100000 A specific reason has been given: Massive calculation work will come soon This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by John Blaas 08 Jun '15

08 Jun '15

Granted 50k to AIN - JB On Mon, Jun 8, 2015 at 1:52 PM, Bair, Raymond A. <rbair(a)anl.gov> wrote: > A small request, so lets approve it. > > Ray > > > > > > > > > On 6/4/15, 10:04 AM, "accounts(a)lcrc.anl.gov" <accounts(a)lcrc.anl.gov> wrote: > >>Hello, >> >>A change in allocation has been requested: >> >> Requester: zapol (Peter Zapol) >> Project: AlN >> Title: A DFT study of reaction pathways of NH3 decomposition on >>AlN (0001) surface >> Description: Structure optimizations of the adsorbate-free (0001) >>surface of AlN were already investigated using (2 Ã‹ 2) surface unit >>cells. We will perform density functional theory (DFT) calculations to >>determine structure, bonding, and reaction energies of the adsorbed >>species. Periodic boundary condition (PBC) will be imposed in order to >>maintain periodicity of the solid support. Gradient corrected PBE density >>functional will be used for these calculations. The Climbing Nudged >>Elastic Band (cNEB) method of Henkelman and coworkers will be used to >>calculate transition states and activation barriers for formation of >>different surface species. >>Each structure in the supported (0001) surface of AlN typically involves >>~100 atoms. To obtain accurate reaction energies and barriers, rigorous >>density functional theory (DFT) calculations are needed. We intend to >>carry out the DFT calculations using the VASP package. VASP is currently >>available on Blues cluster and scaling with number of processors is well >>established. For the MPI version of VASP with the system size of ~100 >>atoms, a good performance was observed up to 32 processors. >>Energy optimization in VASP takes approximately 750 CPU hours and we plan >>approximately 100 of such calculations for different adsorbate >>geometries. The Climbing Nudged Elastic Band (cNEB) calculation (with 5 >>to 7 images) takes approximately 1600 CPU hours and we plan to run 60 of >>these calculations to find reaction pathways. >>So our total request is 750*150 + 60*1600 = 208,500 core hours >> >> Current: undetermined amount >>Justification: >> >> Requested: 50000 >> >>A specific reason has been given: >>we have received considerably smaller amount than requested and as a >>result we have a negative balance. Please add time or at least advance >>the remaining balance in order to continue the project. >> >>This needs to be approved and the final allocation amount decided upon. >> >>Thank You, >>The LCRC Accounts System

1 0

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by John Blaas 08 Jun '15

08 Jun '15

Granted 100k to entropy - JB On Mon, Jun 8, 2015 at 1:51 PM, Bair, Raymond A. <rbair(a)anl.gov> wrote: > A small request, so lets approve it. > > Ray > > > > > > > > > On 6/5/15, 9:10 PM, "accounts(a)lcrc.anl.gov" <accounts(a)lcrc.anl.gov> wrote: > >>Hello, >> >>A change in allocation has been requested: >> >> Requester: sunhwan (Sunhwan Jo) >> Project: entropy >> Title: Quantifying protein-protein binding with greatly scalable >>multiple copy algorithms of NAMD >> Description: In PMF-based framework, a set of geometric and >>conformational restraints is applied to the proteins to reduce sampling >>noise and accelerate the convergence of the resulting free energy >>profile. These biases are rigorously accounted for to produce the >>absolute binding free energy. To enhance the PMF convergence, we will >>combine the umbrella sampling with Hamiltonian and temperature (ℋ and T) >>replica-exchange molecular dynamics (REM) algorithm. In the US/(ℋ,T)-REMD >>simulation approach, multiple copies of the molecular systems are >>simulated concurrently, but all with different conditions, e.g., >>umbrella potential (Hamiltonians) or temperature. Attempts are >>periodically made to exchange configurations between different systems >>using a Metropolis-type Monte Carlo acceptance criterion, insuring >>Boltzmann-weighted statistics. Such REMD methodology is extremely >>scalable, yet also very effective to sample a rugged free energy surface. >> >>We anticipate the simulation systems will be comprised of ~50,000 atoms >>and using at least 255 replicas concurrently. Because of its large size, >>we plan to use primarily Mira for production and Fusion will be used for >>initial system preparation and development of enhanced methodology in >>PMF-based framework. All planned simulations will be carried out by the >>open-source NAMD molecular dynamics software. The software engine used >>for the planned simulations is NAMD 2.10 with Charm++ 6.4.0. We have >>tested t he performance on BG/Q Mira. The scaling of parallel performance >>was linear up to 512 cores for our test system comprised of 100,000 atoms >>without using restraint potential and achieved the speed of about 0.025 >>s/step. Using restraint potential reduces the parallel performance >>significantly (~0.2 s/step). >> >>The expected number of project member is 3. I, Sunhwan Jo, recently >>joined LCF as a postdoctoral appointee and will be the main person >>performing code improvement and algorithm development under the >>supervision of Wei Jiang (LCF) and Benoit Roux (UChicago). >> >> >> Current: undetermined amount >>Justification: >> >> Requested: 100000 >> >>A specific reason has been given: >>A paper is published from this calculation >>(http://pubs.acs.org/doi/abs/10.1021/ct501034w) and we are extending the >>project to perform more sophisticated calculation. We have completed ~50% >>of the calculation needed but need more time to finish the calculations. >> >>This needs to be approved and the final allocation amount decided upon. >> >>Thank You, >>The LCRC Accounts System > > _______________________________________________ > allocations-admins mailing list > allocations-admins(a)lcrc.anl.gov > https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins

1 0

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by John Blaas 08 Jun '15

08 Jun '15

Granted 50k to Atmos_Chem - JB On Mon, Jun 8, 2015 at 1:51 PM, Bair, Raymond A. <rbair(a)anl.gov> wrote: > A small request, so lets approve it. > > Ray > > > > > > > > On 6/5/15, 2:52 PM, "accounts(a)lcrc.anl.gov" <accounts(a)lcrc.anl.gov> wrote: > >>Hello, >> >>A change in allocation has been requested: >> >> Requester: raok (Rao Kotamarthi) >> Project: Atmos_Chem >> Title: Process Scale Modeling of Atmospheric Chemistry and >>Transport >> Description: We continue to investigate the regional scale climate >>modeling and data assimilation of aerosols in chemical transport models. >>The models employed for these studies include (a) WRF-CHEM (b) NRCM and >>(3) GEOS-CHEM. Each of these models has been ported to the fusion luster >>and uses to perform calculations by our group. All the codes use a mix of >>MPI and OpenMP programming models and can scale for as few as 4 nodes to >>several hundred. The GEOS-CHEM model doesn't scale beyond 32 nodes due >>to gridding constraints. The NRCM (Nested Regional Scale Climate Model) >>and the WRF-CHEM models are highly scalable and limited by job turnover >>times and queue limitations. >> >>The project will have approximately 4 members. >> Current: undetermined amount >>Justification: >> >> Requested: 50000 >> >>A specific reason has been given: >>We have doing pre-processing of regional climate model simulations using >>Blues allocation. This has increased the usage of this account by our >>group >> >>This needs to be approved and the final allocation amount decided upon. >> >>Thank You, >>The LCRC Accounts System >>_______________________________________________ >>allocations-admins mailing list >>allocations-admins(a)lcrc.anl.gov >>https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins > > _______________________________________________ > allocations-admins mailing list > allocations-admins(a)lcrc.anl.gov > https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins

1 0

Re: [allocations-admins] [LCRC Accounts] Project Allocation Request
by Bair, Raymond A. 08 Jun '15

08 Jun '15

A small request, so lets approve it. Ray On 6/4/15, 10:04 AM, "accounts(a)lcrc.anl.gov" <accounts(a)lcrc.anl.gov> wrote: >Hello, > >A change in allocation has been requested: > > Requester: zapol (Peter Zapol) > Project: AlN > Title: A DFT study of reaction pathways of NH3 decomposition on >AlN (0001) surface > Description: Structure optimizations of the adsorbate-free (0001) >surface of AlN were already investigated using (2 Ã‹ 2) surface unit >cells. We will perform density functional theory (DFT) calculations to >determine structure, bonding, and reaction energies of the adsorbed >species. Periodic boundary condition (PBC) will be imposed in order to >maintain periodicity of the solid support. Gradient corrected PBE density >functional will be used for these calculations. The Climbing Nudged >Elastic Band (cNEB) method of Henkelman and coworkers will be used to >calculate transition states and activation barriers for formation of >different surface species. >Each structure in the supported (0001) surface of AlN typically involves >~100 atoms. To obtain accurate reaction energies and barriers, rigorous >density functional theory (DFT) calculations are needed. We intend to >carry out the DFT calculations using the VASP package. VASP is currently >available on Blues cluster and scaling with number of processors is well >established. For the MPI version of VASP with the system size of ~100 >atoms, a good performance was observed up to 32 processors. >Energy optimization in VASP takes approximately 750 CPU hours and we plan >approximately 100 of such calculations for different adsorbate >geometries. The Climbing Nudged Elastic Band (cNEB) calculation (with 5 >to 7 images) takes approximately 1600 CPU hours and we plan to run 60 of >these calculations to find reaction pathways. >So our total request is 750*150 + 60*1600 = 208,500 core hours > > Current: undetermined amount >Justification: > > Requested: 50000 > >A specific reason has been given: >we have received considerably smaller amount than requested and as a >result we have a negative balance. Please add time or at least advance >the remaining balance in order to continue the project. > >This needs to be approved and the final allocation amount decided upon. > >Thank You, >The LCRC Accounts System

1 0