Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Michael Borland Project Name: ApsRenewalLattice Division: ASD Project title: Lattice Development and Beam Dynamics for APS Upgrade Associated funding: Department of Energy,Office of Science, Office of Basic Energy Sciences Other Systems: ASD "weed" cluster, operated by ASD/AOP group, has 1112 cores with a lustre filesystem and infiniband network; we can use approximately 80% of this cluster full time. We also have full-time use of a 960-core subcluster that is part of Blues. The allocation request assumes the use of these systems at the levels described. Science: As part of the APS Upgrade, the accelerator lattice (the location and powering of the magnets) may be completely replaced with a multi-bend achromat (MBA) configuration. This configuration promises dramatically lower emittance and dramatically higher x-ray brightness. If successful, it would make APS the brightest storage ring hard x-ray source in the world. The goal of this project is to design and validate an MBA lattice suitable for incorporation into the APS Upgrade. FY18 tasks will emphasize refinement of the design, including iteration with engineering constraints, as well as increased detail in the simulations. Project description: APS accelerator simulation activities center on the code Pelegant, which is capable of optimization, single-particle dynamics, and collective dynamics simulations. Some of the relevant uses include multi-objective genetic algorithms (MOGA) for optimization of accelerator lattices, evaluation of robustness of optimized lattices in the presence of errors, computation of particle loss distributions, and simulation of single- and multi-bunch collective instabilities. Pelegant and related codes have been used for years on Blues to successfully develop and understand the present preliminary design for the APS Upgrade. Tests of Pelegant on Bebop in the early user period showed even better performance. In FY18, we will continue refinement and analysis of APS upgrade lattice. Among the important issues we will address are the following: (1) Continue MOGA optimization of the lattice to adapt to changes originating in evolving engineering constraints. (2) Perform optimization and evaluation for different chromaticity values, from 3 to 7, in addition to the present value of 5. (3) Perform optimization and evaluation for various sextupole terms in the dipole magnets. (4) Determine whether less-performant cases from ensemble evaluation can be improved by reapplication of MOGA optimization. (5) Evaluate performance with field errors derived from simulations of mechanical stack-up in the magnets. (6) Continue optimization of collimation scheme for localization of beam losses. Perform simulations of beam dumps with various collimator shapes. (7) Continue simulations of longitudinal feedback system options and situations, including effects of various sets of higher-order modes, 324 bunches with gaps, phase- and energy-based detection, and noise effects. (8) Simulate the effect of canted insertion devices. (9) Refine and verify field quality requirements for insertion devices, particularly for polarized devices. (10) Perform Touschek scattering simulations for many error ensembles, to understand the loss distribution more fully. (11) Include soft-edge effects on the multipole magnets. Requires reoptimization of the nonlinear dynamics, then reevaluation. (12) Perform simulations using generalized gradient expansion of the fields of the dipole magnets. (13) Compute the beam-stay-clear envelope using injection simulations. (14) Simulate injection with measured kicker waveforms and field non-uniformity. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 40000000 Q1: 10000000 Q2: 10000000 Q3: 10000000 Q4: 10000000 Justification: Pelegant has been used for years on Blues, Mira, and other computers with good parallel efficiency. The number of cores that a computation can utilize efficiently depends on the type of simulation and the problem size. Many of the computations are embarrassingly parallel and scale well up to hundreds of cores; a recent example is the newly-added simulation of elastic gas scattering, which shows no significant variation in efficiency between 16 and 640 cores (the largest number tested). Less common types of simulations involve collective effects; with typical problem parameters, these show greater than 50% efficiency up to 144 cores, limited by the need to frequently share particle distribution information. Most of our simulations make use of multiple jobs to explore parameter space or different error seeds, with each job using on the order of a hundred cores. For example, MOGA optimization will typically involve 30-60 simultaneous jobs using two nodes each; th e efficiency in this case is very high. Storage requirements: 3 TB, needed for storage of large data files generated by collective effects simulation. Thank You, The LCRC Accounts System