View source

The week-of-code branch has multiple heads

week-of-code
change destination
week-of-code (949a0436ab40)
  • Commit status
  • Contributors
    1. Loading...
Author Commit Message Date Builds
250+ commits behind week-of-code.
John Wise
Removing SNII field from Larrue's fixes for my testing purposes.
james_larrue
Fixed conditions on the field type to include ExtraType0, ExtraType1, MetalSNIaDensity, and MetalSNIIDensity.
james_larrue
Added comments to explain why BoundaryFluxesThisTimeStep does not need to be undone when Density, Total Energy, or Internal Energy are <= 0.
james_larrue
Modified so that the results are no longer dependent on the order of the fields in the BaryonField array. Previously, if the corrected density or energy (total or internal) was <= 0 in a cell, then the correction for that field would be undone and all further corrections would be set to zero. However, any past corrections would remain as they were. In a simulation where the corrections were to be undone because, for example,the total energy became negative, the fields that were undone would depend on where in the BaryonField array the offending field (e.g. total energy) was located. The modified version will undo corrections for all fields in an offending cell. Also, in the scenario I was investigating, I found that the correction for velocity in offending cells might be far larger than the pre-correction value, resulting in v_pre + v_correction - v_correction != v_pre, when the correction was undone. While velocity was the issue for me, this could theoretically occur for an…
John Wise
Fixing a bug in the RT restart. Previously tried to access NumberOfPhotonPackages before being calculated. Just count them instead.
John Wise
Fixing a bug in the ray merging with threading. (transplanted from c612b0267cadf9c6f2c549b7a91f2f63988fde5f)
John Wise
Don't thread the ray tracing if there are less photon packages than cores. (transplanted from ee71c407d16da1b83890867410d1b34dbd82073c)
John Wise
Finally threaded the ray tracing!! (transplanted from e3aa465da3ca0bf840787f036878678695298c94)
John Wise
Added basic configuration for OpenMP. No threading yet. (transplanted from 2554f3d5a0f3fe7ad6a7a19e4f438253583a0258)
John Wise
Minor updates to output in the star cluster formation routines.
John Wise
Updates to color field in the collapse test. Renaming it SN_Colour and default values to 1e-6 of solar metallicity.
John Wise
Turning off inline analysis stdout for collapse test.
John Wise
Hard-coding an option to output just after a Pop III star forms. Current default setting is off.
John Wise
Place a floor on the color Riemann solver fluxes that is consistent with the density floors. Also added more debugging info in crashes.
John Wise
Use hydro timestep at maximum level for the RT timestep when there is no radiation in anticipation for any sources that might form.
John Wise
Fixing a crash when running an RT simulation with the RT load balancing parameter on but running in serial.
John Wise
Adding timers to Moray.
John Wise
Don't calculate radiation pressure prefactor if not using it.
John Wise
Storing HI/HeI/HeII cross-sections in rays instead of recalculating them for every trace.
John Wise
Attempt at partially vectorizing the photo-updates in the ray tracer.
John Forbes
Alterations to two-shock Riemann solver and flux correction to keep initially uniform metallicity fields uniform.
John Wise
Adjustments to the timestep limited by cooling time.
John Wise
Forgot to take into account metals when increasing the species fields for supernova feedback.
John Wise
Supernova feedback bubble doesn't assume the medium is fully ionized anymore. Let the solver take care of that.
John Wise
Split min(dt) for Cray compiler (internal bug).
John Wise
Reformulating the minimum photon package flux to be dependent on the refinement mass (i.e. total absorbers in a cell) instead of the mean density.
John Wise
Including all radiation sources that are active at the restart during RestartPhotons. Minor adjustment to the initial RT time, which affects extreme zoom-in simulations.
John Wise
Adding some optimization flags to the Blue Waters machine file.
John Wise
When creating a radiating particle, an initial radiation field in the optically thin limit is placed in the "accretion sphere". It was already set to be ionized, but then it quickly recombined and cooled.
John Wise
Uninitialized maxLevel for RT timestep calculation. More debugging statements for NaN temperatures.
John Wise
Calculate RT timestep from the maximum level with radiation, not overall.
John Wise
Making indentation more clear.
John Wise
Merging
John Wise
Fixing a line with >72 characters when macros are included.
John Wise
Merging.
John Wise
Adding Britton's changes to consider a lower (1e-7) H2 fraction for Pop III star formation if the cell has a high LW flux, using Omukai & Yoshii (2003) as motivation.
John Wise
Don't require the OmegaDarkMatter parameter anymore since it's defined during initialization.
John Wise
Commenting out debugging statement.
John Wise
Reverting formation sphere to have uniform density.
John Wise
Implementing Wolcott-Green H2 self-shielding (Sobolev-like, spherically averaged).
John Wise
Fixed the non-cosmology initial H2I density.
John Wise
Adding the ability to set species fraction in the sphere and the ambient medium in the collapse test.
John Wise
Including a lower critical H2 fraction if in the presence of a strong LW field, and lowering the ramp time of Pop III stars to 10 kyr (Susa 2013).
John Wise
Re-enabling initial HII region after a radiating star particle is created. Changed the mass removal to be a constant fraction of mass loss instead of replacing the sphere with a uniform density.
John Wise
Increasing chemistry max iteration and decreasing LW field min radius to 1 AU.
John Wise
Adjustments to RT timestep and load balancing. Should not interpolate RaySegments field.
John Wise
Fix for simulations with over 2 billion cells on a level.
John Wise
Fix to newly added multi-mrp refinement restriction that considers all refinement criteria to work when there are only particle based refinement.
John Wise
Adding option to neglect all Pop III supernovae.
Britton Smith
Adding cooling timestep limiter.
Britton Smith
Adding function to stop when reaching a maximum density and a maximum metal-enriched density.
Britton Smith
Adding option to use core collapse supernova yields instead of hypernova for 20-40 Msun Pop III stars.
Nathan Goldbaum
Ensure the total energy field is consistent with the internal energy field after flux correction.
Nathan Goldbaum
Only make energy fields conservative if cooling is turned off.
dcollins4096
fixing the internal energy flux.
John Wise
Error check if Omega_DM is specified when using multi-mrp.
John Wise
Changes to allow AMR only where the must-refine particles exist in gas+DM simulations. Previously, this method was restricted to particle overdensity refinement.
John Wise
With must-refine particles in cosmology simulations, regions will only be refined to levels higher than RefineToLevel if there is a must-refine particle there. Important: For all cosmological simulations that use must-refine particles from initialization to follow Lagrangian regions, the RefineToLevel must be set to the level of the finest initial grid.
John Wise
cic_flag.F had the particle mass defined to have P_PREC, i.e. FLOAT. It should have been R_PREC (float).
John Wise
Swapping slab dimensions for reading integer fields for cosmological ICs.
John Wise
Missed a bracket during the transplant of the multi-mrp routines.
Christine Simpson
converted critical grid ratio to run time parameter
Christine Simpson
Reversed MRP buffer change; simplified changes to IdentifyNewSubgridsBySignature
Christine Simpson
Testing larger MRP buffer
Christine Simpson
removed/added some print statements and made a correction for multi-MRP method
Christine Simpson
Added check of subgrid size for new grids created by inflection point splitting
Christine Simpson
Fixed some typos
Greg Bryan
Change to grid identification routine: we now check for large axis ratios before doing check of second-derivative and if the grid axis ratio is too large, we split it along the long axis.
Christine Simpson
First stab at implementing multi-MRP
Christine Simpson
Removed debugging print statements
Christine Simpson
Reversed incorrect change to SubCellWidth calculation in nested cosmology initializer
Christine Simpson
Inserted temporary error reporting; will be removed later
Christine Simpson
Fixed slab dimensions in ReadIntFile
Christine Simpson
Implemented changes to read in Music masking field and to convert dm particles within mask to type must refine
Christine Simpson
Changed method for reading in list of MRPs
csim...@trestles-login1.sdsc.edu
Fixed ParticleCount
csim...@trestles-login2.sdsc.edu
Only report if must refine particles are flagged on a grid
Christine Simpson
Corrected range over which grid slab is read in
Christine Simpson
Removed unnessary instance of ParticleTempCount in NestedCosmologySimulationInitialize.C
Christine Simpson
Added changes to utilize must refine particles from initial conditions.
John Wise
Minor changes to remove sources that weren't formed yet during restarts.
John Wise
In ray merging, don't walk up the clustering tree before tracing it.
John Wise
Forgot to remove a debugging statement
John Wise
Merging RT heads.
John Wise
Adjusting RT restart timesteps and deleting any photons that still exist on the grids after the restart.
John Wise
Adjusting debugging statements.
John Wise
Fixed a bug where static sources weren't found in a dynamic AMR hierarchy.
John Wise
Fixing a bug where the super sources weren't located when rebuilding the radiation source tree. Also fixed a couple of memory leaks.
John Wise
Merging RT changes (again)
John Wise
Reverting changes to the source clustering tree.
John Wise
Merging RT bookmark heads
John Wise
Merging latest RT changes.
John Wise
New star particle - Moray interface compiles with photon-no.
John Wise
Bugfix for the ray merging when distant two leafs are associated. Also fixes some duplication of single leaf nodes.
John Wise
Adjusting star object masses to account for mass loss that occurs during SN feedback. Missed a factor of 4pi in the luminosity. Fixed a bug where a fixed dtPhoton was changed during restart with no radiating particles.
John Wise
Small bugfixes in the RT / star particle coupling.
John Wise
Accounting for SNe mass loss in star objects.
John Wise
Multiple minor fixes to get the RT working with normal star particles.
John Wise
Fix restart if there are no sources.
John Wise
First shot at converting normal star particles into radiation sources. Not tested, but it compiles.
John Wise
More optimizations for feedback spheres for better scalability.
John Wise
Optimizations in feedback sphere checking.
John Wise
Fixed a bug where inactive sources were being considered during RT restarts.
John Wise
Fixing a bug where photons were not transferred to there correct grid when they split near the source in a tightly packed grid structure.
John Wise
Adding #define LW_TREE around the tree calculation of the LW radiation field because it can cause crashes in some (unknown) cases.
John Wise
Use RadiationFieldRedshift for Compton cooling/heating.
John Wise
Removing RaySegments if RT is turned off.
John Wise
Adding a new radiative transfer parameter that limits the timestep based on a sound crossing time on a user-specified level.
John Wise
RaySegments should not be interpolated and should be zeroed out in regions with child grids, i.e. rays are only traced on the finest level.
John Wise
Fixing a long-standing bug where a photon was orphaned when the MoveToGrid was computed incorrectly.
John Wise
Merging from mainline.
John Wise
Merging from mainline.
John Wise
Merging.
John Wise
Fixing some mixed precision issues.
John Wise
Fixing LW luminosities for Pop II stars.
John Wise
Merging from mainline
John Wise
Merging
John Wise
Removing debugging statement in FOF
John Wise
Merging in mainline changes again
John Wise
Merging
John Wise
More refinements to the feedback sphere.
John Wise
Merging
John Wise
Merging changes from mainline
John Wise
Merging
John Wise
Merging.
John Wise
Merging.
John Wise
Merging
John Wise
Merging
John Wise
Deleting the LevelArray linked list in the non-grouped datasets. Setting the default hierarchy output back to HDF5 and ASCII.
John Wise
Deleting LevelArray linked list.
John Wise
Merging
John Wise
Merging
John Wise
Merging
John Wise
Fixed bug when the Pop III SN remnant volume isn't calculated correctly when the prescribed radius is below the resolution.
John Wise
Merging from mainline
John Wise
Merging with mainline
John Wise
Should consider fractional coverage of the ray's solid angle on the cell face in the minimum photon flux for deletion.
John Wise
Implemented a user-defined parameter that controls the minimum flux of a ray before deletion. New Parameter: RadiativeTransferHubbleTimeFraction The minimum flux is determined by considering the extreme case where all of the photons are absorbed in one cell, where the cell will become fully ionized within t = (this parameter) * (Hubble time).
John Wise
Correcting some merging errors.
John Wise
Merging.
John Wise
Merging.
John Wise
Don't add NULL subgrid markers from parent grids. Fixes a segfault.
John Wise
Merging in latest changes from mainline into my fork.
John Wise
Merging Dave's fix
John Wise
Merging
John Wise
Fixed some bugs in the inline halo finder in calculating the critical density at the current time. Changed default linking length to 0.2.
John Wise
Halo centers are now calculated with an iterative method that calculates the center of mass with in a shrinking sphere.
John Wise
Changing Jeans cold temperature to a floor instead of a constant.
John Wise
Reverting accidental commit of hg path in Makefile
John Wise
Grid_RegridPausedPhotonPackage.C
John Wise
Only consider the dt increase limit at high density. Caused problems when photo-ionization rates are used.
John Wise
enzo_anyl compiles with photon-yes.
John Wise
Adding performance timers to hydro_rk and turning off STAR_LOG in star_maker_h2reg.F.
John Wise
Merging another head.
John Wise
Merging heads.
John Wise
Forgot about these conflicts in SGSizeExtrema.
John Wise
Changes to compile with open64.
John Wise
Merged changes from mainline
John Wise
Merging
John Wise
Fixes to compile with PGI.
John Wise
Decreasing critical photon flux for deletion.
John Wise
Merging
John Wise
Enforcing a density floor for initial density field
John Wise
When using optically-thin X-rays, must solve for chem/energy for all grids. Changed looping order for better efficiency.
John Wise
Commented out tree-accelerated optically-thin x-rays.
John Wise
Extending optically-thin X-rays to stars.
John Wise
1. Ability to calculate X-ray flux in the optically thin limit (no support for stars or tree acceleration yet). 2. Mark subgrid marker for rays to NULL for ghost zones outside the domain for isolated boundary conditions.
John Wise
Cleaning up rate solver calls in EvolvePhotons
John Wise
Place a floor for external density fields.
John Wise
Turning off ray boundary warning.
John Wise
Calculate mu directly from data for initial temperature.
John Wise
Merging
John Wise
Adding ability to read external species fields for PhotonTest.
John Wise
Merging
John Wise
10% speedup to MHD by reordering some loops
John Wise
Merging.
John Wise
Turn off star lifetime timestep limiter
John Wise
Reducing the number of MPI calls during cosmology initializations. Updated some calls in RK2. Removing star lifetime limiter on the timestep.
John Wise
Adding cell flaggin method to refine by metal mass.
John Wise
When MultiSpecies is on, the colors are considered as fractional species. So we need to advect them as well in the hydro_rk solvers.
John Wise
Setting the upper limit for Type II SN to 40.1, so 40 (plus a little) Msun is included.
John Wise
Level loop in RH over all levels and then check if grids exist.
John Wise
LW tree calculation now works with star particles; optimized the tree search; debugging
John Wise
Creating a wrapper routine to choose the tree calculation if there is ray merging and >10 sources.
John Wise
Lyman-Werner radiation calculation now done in a binary tree!
John Wise
Merging in LW (incomplete) tree calculation. Saved from April 2011.
John Wise
First attempt at calculating the Lyman-Werner flux using the binary tree built for ray merging. Not working.
John Wise
Minor optimizations to ray merging. The radius was a float instead of FLOAT. Maybe this was causing some "misplaced" photons when ray merging is used?
John Wise
Making my old projection routines to have the fields weighted by density instead of density^2.
John Wise
Consolidating an MPI call. Avoiding an Allreduce if star_maker9 is not used.
John Wise
Merging.
John Wise
Merging
John Wise
Adding a command line option to write the dust temperature field to the dataset.
John Wise
Precision changes for ray merging.
John Wise
Fixing a bug that removed gas from the wrong cell after BH accretion. Also removing density from any existing colour field.
John Wise
Deleting the LevelArray linked list in the non-grouped datasets. Setting the default hierarchy output back to HDF5 and ASCII.
John Wise
Deleting LevelArray linked list.
John Wise
Adding a command line option to write the dust temperature field to the dataset.
John Wise
Precision changes for ray merging.