Small bug fix. This was not a bit changing bug, but could cause crashes if we compile with floating point trapping in a multicolumn run.
(details / githubweb)
Small refactoring to mfl_xm_lhs, just removing need to pass in gr.
(details / githubweb)
Final loop pushes needed before the grid change.
(details / githubweb)
added -heap-arrays tag to FFlags on Makefile to fix jenkins testing error
(details / githubweb)
Makefile came with a few other accidental changes, these have been corrected.
(details / githubweb)
The big grid change. Converting gr from being an array of types containing 1D arrays, to a type containing 2D arrays. All cases BFB, cam multicolumn+silhs BFB, and cam multicolumn (no silhs) with backwards compatible settings BFB.
(details / githubweb)
The big grid change. Converting gr from being an array of types containing 1D arrays, to a type containing 2D arrays. All cases BFB, cam multicolumn+silhs BFB, and cam multicolumn (no silhs) with backwards compatible settings BFB.
(details / githubweb)
Adding column index to gr arrays, hard coding to 1 since sam only uses single column calls.
(details / githubweb)
Removing the zt2zm interface from clubb_api and making the api calls just redirect to the grid class interface for it.
(details / githubweb)
Pushing column loop into mean advection procedures.
(details / githubweb)
adding line to fix mpi configuration in micro drizzle test
(details / githubweb)
I altered the w_up_in_cloud and w_down_in_cloud code so that a
(details / githubweb)
I have optimized the new w_up_in_cloud and w_down_in_cloud code by
(details / githubweb)
Refactored fill_holes_vertical to make GPUization simple. This is BIT_CHANGING, but results are bit-for-bit when using -O0 optimization, thus it is not answer changing. The first pass over each grid column will not parallelize well, the k-loop needs to be done in serial. Maximum parallelization has been exposed for the global hole-filling though, at the cost of occasionally doing unneccesary calculations. larson-group/clubb#972.
(details / githubweb)
Removing fill_holes_multiplicative and replacing magic numbers with parameters from constants_clubb. larson-group/clubb#972
(details / githubweb)
Moving vertical_avg and vertical_integral to advance_helper_module. larson-group/clubb#972
(details / githubweb)
Moving vertical_avg and vertical_integral to advance_helper_module. larson-group/clubb#972
(details / githubweb)
Making compatible with latest clubb change.
(details / githubweb)
Removing elementalness from sat_vapor_press_liq and making interal procedures subroutines rather than functions to prevent unneccesary data copies. Doing the same for thlm2T_in_K since it is often used in conjunction with sat_vapor_press_liq. Bit-for-bit confirmed with O0 using all single column cases with or without l_diag_Lscale_from_tau, and with cam_coarse_res. larson-group/clubb#972
(details / githubweb)
Cleaning up new subroutine calc_liquid_cloud_frac_component, and making sat_mixrat_ice a subroutine that works the same way as sat_mixrat_liq. larson-group/clubb#972
(details / githubweb)
Removing these paratheses is BIT_CHANGING since it modifies the order of operations, but allows for the multiplication and subraction to be done in parallel for complex pipelines.
(details / githubweb)
Improvements. Netcdf output is now functional and we can detect errors with multiple columns even when the standard output is identical.
(details / githubweb)
Updating column num input in config flags. larson-group/clubb#1033
(details / githubweb)
Adding comments, setting to num_standalone_columns=1 by default, and preventing multicolumn output when l_stats=false. larson-group/clubb#1033
(details / githubweb)
Improvements. Netcdf output is now functional and we can detect errors with multiple columns even when the standard output is identical.
(details / githubweb)
Namelists need to be closed with a '/', and omitting it will break many tests.
(details / githubweb)
Making save variables in the multicol code threadprivate, this fixes the openmp test. Also changing the way we name and save the multicol output, this allows the parallel run to run with multiple columns. I've reflected the changes in comments where relevant. larson-group/clubb#1033
(details / githubweb)
Adds commented-out line that prevents the monotonic flux
(details / githubweb)
Making compatible with latest clubb changes.
(details / githubweb)
Adding capability to change matrix solving method via clubb_config_flags.
(details / githubweb)
Fixing small bug, need to pass _copy arrays to prevent lapack mangling the real ones.
(details / githubweb)
Adding capability to change matrix solving method via clubb_config_flags.
(details / githubweb)
Making GPU and CPU versions of the penta_lu solver the same as discussed in larson-group/clubb#1024.
(details / githubweb)
Replacing $ with & in namelist definition to make consistent.
(details / githubweb)
I added "smooth" max clipping for invrs_tau_shear, which is a variable
(details / githubweb)
ADG1_pdf_driver subroutine port with OpenACC
(details / githubweb)
Replacing old elemental ADG1_w_closure with new GPUized one. Making G_unit tests work with new version. Also making mixt_frac_max_mag a scalar since it was only being used as such. larson-group/clubb#1049
(details / githubweb)
I can now safely remove all the "ifdef E3SM" statements from CLUBB's
(details / githubweb)
Restructuring and Porting of Compute_mixing_length subroutine(Phase 1) (#1052)
(details / githubweb)
Restructuring and Porting of Compute_mixing_length subroutine(Phase 2) (#1054)
(details / githubweb)
Breaking up column loop in mono_flux_limiter. This may not be the final form for GPUization, but it's definitely a start, no longer do we have to copy single column variables to multicolumn ones anywhere. (#1051)
(details / githubweb)
Adding OpenACC data directives for mixing length and adg routines
(details / githubweb)
Fixing error causing GPU code not to run. Some variables to be copied were labelled as (ngrdcol,nz) when it should be (:ngrdcol,:nz). I've just removed the data length specifiers completely since they are not neccesary in general.
(details / githubweb)
Removing update_pressure since it is no longer called anywhere in clubb or host models. The addition of this subroutine was discussed in larson-group/e3sm#6 and the removal of the call to it was discussed in larson-group/clubb#926.
(details / githubweb)
Removing update_pressure from public list. This was causing compilation crashes. RESOLVED:8c7230fecb877d04fb129ef5e143e0993b4b29b1
(details / githubweb)
Removing sigma_sqd_w from the acc data copyout statement. This is a bug which was causing the code to crash when not using managed memory.
(details / githubweb)
Porting pdf_closure subroutine with OpenACC (#1059)
(details / githubweb)
Adding wp3_on_wp2_cfl_num to CLUBB's output stats.
(details / githubweb)
Removing usage of gr from pdf_closure. It was only ever used for nz, which is now fed in directly.
(details / githubweb)
Clubb ticket #1025: Implemented way to make esa tuner reproducible, h… (#1068)
(details / githubweb)
Moving compute_cloud_cover outside of if ( l_use_cloud_cover ) then statement, the cloud_cover and rcm_in_layer variables they compute aren't output in clubb_standalone, but are in cam, causing cam bit diff tests to break.
(details / githubweb)
Adding Skthl_zm to the update host list, I missed this in the last PR. I noticed this by comparing results with and without managed memory, now I've checked BFBness with arm, mpace_b, mc3e, and gabls2.
(details / githubweb)
Small tweaks to fix some GPU bugs. Some variables were uninitialized on the CPU while we were saving them. This could only have been caught by comparing consecutive runs and checking _zt and _zm files, even then few cases were having problems.
(details / githubweb)
Fixing a labelling error in redirect_interpolated_azt_2D and similar procedures, since this interpolates to zt the input should be zm. I think this was my fault, so I cleaned all the zt2zm and zm2zt things up to make it a little nicer. Also ordered the routines _k _1D _2D to make it easier to jump around, it was a bit confusing as they were out of order and the typo really made it hard.
(details / githubweb)
Making it so sclr_tol is set to 0 before the specified sclr_tol_in. This is so that it is initialized to 0 in the case that sclr_dim = 0, since now we are setting it to have a minimum allocation size of 1 and would otherwise have a garbage value. This is what broke the clubb_openmp_gfortran_test.
(details / githubweb)
This commit is a commit that changes absolutely nothing. It is meant to trigger a change in the git update scripts, so that I can start the commit message logging in the autocommit updates larson-group/sys_admin#797
(details / githubweb)
This commit is a commit that changes absolutely nothing. It is meant to trigger a change in the git update scripts, so that I can start the commit message logging in the autocommit updates larson-group/sys_admin#797
(details / githubweb)
this is another commit that cahnges nothing that will trigger the gitUpdate scripts
(details / githubweb)
adding an update that changes nothing and is just a test for the autoupdate script
(details / githubweb)
Making CLUBB's splatting scheme implicit and smoother (#1075)
(details / githubweb)
change to calc pressure to trigger autoupdate
(details / githubweb)
GPUizing Lscale_width_vert_avg. Loops have been restructured for simplicity, and algorithm has a different starting value to avoid k dependency. Results are BFB. (#1083)
(details / githubweb)
GPUizing most of advance_clubb_core (#1084)
(details / githubweb)
advance_wp2_wp3 with explicitly managed memory (#1085)
(details / githubweb)
advance_xp2_xpyp with explicitly managed memory (#1086)
(details / githubweb)
advance_windm_edsclrm with explicitly managed memory (#1087)
(details / githubweb)
Moving data statements to outter most parts of clubb and little fix in advance_wp2_wp3 (#1088)
(details / githubweb)
Fixed some oversights for the bv_efold commits and increased string length for tuning field names for invrs_tau variables - CLUBB ticket #1069 (#1106)
(details / githubweb)
BIT_CHANGING! brunt_vaisala_freq_sqd_mixed now replaces brunt_vaisala_freq_sqd for later calculations - CLUBB ticket #1102
(details / githubweb)
BIT_CHANGING for lscale setup. Updated configurable_model_flags.in and model_flags.F90 - CLUBB ticket #1102 (#1108)
(details / githubweb)
Likely BIT_CHANGING because of added output fields. Added code outputting C6_term and different Brunt Vaisala frequencies to disk - CLUBB ticket #1102 (#1109)
(details / githubweb)
Changed sqrt_Ri_zm to Ri_zm in preparation of changing the exponent for calculation of invrs_tau_wpxp_zm in mixing_length.F90 - CLUBB ticket #1102 (#1110)
(details / githubweb)
BIT_CHANGING! Removed Richardson_num and all corresponding code - CLUBB ticket #1102 (#1111)
(details / githubweb)
Changing order of sqrt and **3 to improve stability, see clubb#1126. Since this is a change to the order of operations, it is BIT_CHANGING.
(details / githubweb)
Setting the minimum tolerance to be epsilon, this is BIT_CHANGING but only for single precision. See larson-group/clubb#1126
(details / githubweb)
Making compatible with latest clubb change.
(details / githubweb)
Removing awkward nan check in favor of a default for Ncnp2_on_Ncnm2
(details / githubweb)
Fixing bug by changing saturation_equation to match namelist argument, also adding it to namelist.
(details / githubweb)
Adding new argument list variables to openacc copyin statements
(details / githubweb)
Setting l_smooth_Heaviside_tau_wpxp to .true. by default
(details / githubweb)
I am altering the mean advection code in preparation for ghost point
(details / githubweb)
Fixing tridiag solver bug. See https://github.com/larson-group/clubb/issues/1139
(details / githubweb)
Slight rework to acc data statements. This allows us to use return calls in advance_clubb_core again. See https://github.com/larson-group/clubb/issues/1139
(details / githubweb)
Changing xm to out in mono flux solver. There is no need for it to be inout, and it confused me for a bit so I'm fixing it.
(details / githubweb)
Adding clubb specific compiler flag -DCLUBB_GPU, intended for use with openmp or openacc running on the GPU. Also adding a compiler flags, -DCUDA, and case statment for generating random numbers in silhs.
(details / githubweb)
Adding clubb specific compiler flag -DCLUBB_GPU, intended for use with openmp or openacc running on the GPU. Also adding a compiler flags, -DCUDA, and case statment for generating random numbers in silhs.
(details / githubweb)
Now that the value of wp3 is being set to a boundary value of 0
(details / githubweb)
This is (amazingly) BIT_CHANGING. The differences are due to some compiler optimizations, see https://github.com/larson-group/clubb/pull/1133\#issuecomment-1865156979.
(details / githubweb)
Removing backslashes. These should be slashes, but gnu,intel,nvidia generate code that treats slashes as whitespace and backslashes as a character, while cray treats backslashes as whitespace as well. This is easily fixed by using quotes, and in the meantime we will change the backslashes to forward slashes. See https://github.com/larson-group/clubb/issues/1138\#issuecomment-1910844013
(details / githubweb)
Fixing a bug where the Lscale used in calculating the damping
(details / githubweb)
adding equals sign cuz I've had it up to here with this typo!
(details / githubweb)
fixing bug, name of wm_zt was left as "wm"
(details / githubweb)
Add blank line in order to trigger Jenkins tests for CAM
(details / githubweb)
Fixing bug with C6 parameters, see https://github.com/larson-group/clubb/pull/1147. This is technically BIT_CHANGING, but only when C6thl != C6thlb and l_diag_Lscale_from_tau=.false.
(details / githubweb)
Making compatible with latest clubb changes.
(details / githubweb)
At the ghost point below the surface, which does not matter in
(details / githubweb)
The code that advances um and vm when l_predict_upwp_vpwp is turned off
(details / githubweb)
It was discovered that the location where the ghost point values
(details / githubweb)
It was discovered that the location where the ghost point values
(details / githubweb)
I accidentally committed a couple things in the previous commit
(details / githubweb)
While it doesn't affect results, I fixed what appears to be an
(details / githubweb)
I am updating the loop in subroutine diagnose_xpwp so that it doesn't
(details / githubweb)
This commit is being made to keep the stats zt output from the
(details / githubweb)
I am committing a bug fix to the code that calculates mixed Brunt
(details / githubweb)
I am committing a stats change to the ghost point values of zt
(details / githubweb)
Committing some thresholding around positive-definite fields after
(details / githubweb)
Changing the linear interpolation in the zt2zm function at momentum
(details / githubweb)
I'm changing the way ddzt handles the derivative over momentum level 1.
(details / githubweb)
Moved the ghost-point setting of xm(1) = xm(2) to the last line
(details / githubweb)
Optional arguments are just best sometimes (#1150)
(details / githubweb)
Fixing bug. RESOLVED:556370c6406e6b943b88affbe7728a5036b1ba22, this was catchable by using debugging flags, but seems to otherwise cause no issues. Investigating I found that it only clips the below ground level, so that must not have been affecting other levels. This is technically BIT_CHANGING since it resolves an issue.
(details / githubweb)
Making num_draw_points in fill_holes a constant. We were already using this value as a constant everywhere, but passing a constant by argument list makes it difficult/impossible for a compiler to optimize using that constant, unless it does inlining. Now, rather than passing the constant num_hf_draw_points (or sometimes a hardcoded 2) we just use num_hf_draw_points directly from constants_clubb. This massively improves the performance of a loop in fill holes when using nvhpc+omp, which was the motivation for this, but should also improve the performance on CPUs. Everything is BFB.
(details / githubweb)
Making compatible with latest clubb changes.
(details / githubweb)
I am committing code that effectively ignores the values of certain
(details / githubweb)
I am moving the advance_xm_wpxp setting of xm(1) = xm(2) from the
(details / githubweb)
Adding extra exclamation to acc statement to comment it out.
(details / githubweb)
gr%zm should be used here instead of gr%zt, since these are
(details / githubweb)
Thermodynamic-level (zt) values should be used with zt-variable
(details / githubweb)
Fixing small GPU bug in advance_xm, also fixing a bug with sclrpthvp that's only detectable when were running 1 column on GPUs.
(details / githubweb)
Fixing a bug. Does not change default results.
(details / githubweb)
The variable rc_coef_zm needed to be passed in and saved from
(details / githubweb)
Making the f in f90 capital for consistency.
(details / githubweb)
Making the only change required to make clubb run with openacc in cam.
(details / githubweb)
Ri zm calc (Part 1) - Combine and move call to calc_brunt_vaisala_freq_sqd (#1154)
(details / githubweb)
This code in adj_low_res_nu was using the ghost point when
(details / githubweb)
I am updating the SILHS code that calculates k_lh_start to work
(details / githubweb)
I am committing changes to stats output for cloud_cover, rcm_in_layer,
(details / githubweb)
There were furhter issues found with the k_lh_start code.
(details / githubweb)
Ri_zm calc (Part 3) - Moving calculations for Ri_zm out of diagnose_Lscale_from_taus (#1157)
(details / githubweb)
In my previous commit, I changed the domain of the maxloc commands
(details / githubweb)
This commit brings the stats output from the all cases run with
(details / githubweb)
Committing a change that will initialize this set of variables
(details / githubweb)
Making a change to the l_C2_cloud_frac option with advance_xp2_xpyp.
(details / githubweb)
Fixing bug. qclvar copying is handled by the api, but we were deleting it with an unstructured data statement. I added this due as a fix before thinking qclvar was a local variable, which happened because I mixed up the CLUBB_CAM and CLUBBND_CAM flags. https://github.com/larson-group/cam/issues/175
(details / githubweb)
removing cheyenne-related content and adding new content for
(details / githubweb)
Fixing GPU bug. There is a vertical dependency with wpxp. This is what was breaking the ECT test, see larson-group/cam#175
(details / githubweb)
Fixing bugs that are only caught when not outputting w_[up/down]_in_cloud stats or if l_host_applies_sfc_fluxes=.true. The l_host_applies_sfc_fluxes=.true. isn't testable (at least not in clubb_standalone), and I just caught that visually. The cloudy_updraft_frac and cloudy_downdraft_frac fields not being set were only noticable in cam, because we always run with iw_up_in_cloud>0 and iw_down_in_cloud>0 to output stats.
(details / githubweb)
Moving pdf_params copying to copyin. This is the data structure that contains pointers, and these pointers will be different on the host and device, so it is bad in theory to copy the structure back to the CPU as it might overwrite cpu memory pointers with gpu memory pointers. In practice though I've seen no problems caused by this, I'm just making this commit preemptively, and it has already been tested with the ECT test.
(details / githubweb)
Fixing mpirun command in the derecho and casper batch scripts.
(details / githubweb)
Editing the LES setup script to give instructions for UCAR computers.
(details / githubweb)
Added zt2zm2zt-smoothing to remove noise in MPACE_B winds (#1161)
(details / githubweb)
Added zt2zm2zt-smoothing to remove noise in MPACE_B winds (#1161)
(details / githubweb)
Removing the target attributes wherever possible. These are artifacts from when gr and the the stats types were globals, and got naively copied when these were pushed through the call stack.
(details / githubweb)
Removing the target attributes wherever possible. These are artifacts from when gr and the the stats types were globals, and got naively copied when these were pushed through the call stack.
(details / githubweb)
![[Jenkins]](/jenkins/static/57c8a776/images/svgs/logo.svg){kind=logo}
