Parent Directory
| 
Revision Log
| 
Revision Graph
| Links to HEAD: | (view) (download) (annotate) |
| Links to checkpoint57p_post: | (view) (download) (annotate) |
| Sticky Tag: |
updated after changing g77 default testing option
Changed the order of operations for global sums of arrays.
o this *DOES* change output
- the following results are updated by this change:
exp4 front_relax global_ocean.cs32x15 global_ocean_pressure
ideal_2D_oce internal_wave natl_box plume_on_slope
- changes are primarily due to different cg2d residuals but in some
the Eta field changes and this changes other fields
o the global sum used to accumulate sequentially over all tiles on a single
process and then the result be passed to _GLOBAL_SUM()
o now we create a sum per tile and accumulate the sums
- this should fix differences in sum for two tiles on 1 or 2 processors
- this may help when using more than2 tiles but is not guaranteed
- introducing a gather operation in GLOBAL_SUM() will fix differences
for multiple tiles but only for fixed tiles configurations
- an array based sum function is needed to make this independent of
tile configuration
Re-tiled many experiments so as to allow valid multi-processor testing
o single processor configurations now have multiple tiles
o multi-processor configurations now have the same number of tiles as
the single processor config. but distrubuted
o all output for T,S,U,V were unchanged but the cg2d output was sensitive
to the re-tiling for some expts. Namely:
adjustment.128x64x1, exp4, exp5 front_relax, natl_box and plume_on_slope
Updated (affected by changes in vertical advect. of momentum).
Regenerated results/output.txt for (output of) experiments affected by changes
made in pkg/monitor on Nov 8, '01.
**This did not change the solutions but just the way in which the statistics
are calculated**
The following expts are affected:
adjust_nlfs.cs-32x32x1, adjustment.128x64x1, adjustment.cs-32x32x1,
advect_cs, advect_xz, aim.5l_Equatorial_Channel, aim.5l_LatLon,
aim.5l_cs, exp1, exp2, exp4, exp5, front_relax, hs94.128x64x5,
hs94.1x64x5, hs94.cs-32x32x5, internal_wave, natl_box, solid-body.cs-32x32x1
Output from testscript after the pkg/monitor changes:
T S U V
C D M c m s m s m s m s
n p a R g m m e . m m e . m m e . m m e .
f n k u 2 i a a d i a a d i a a d i a a d
g d e n d n x n . n x n . n x n . n x n .
Y Y Y Y 16 16 16 16 0 16 16 16 16 16 16 22 0 16 16 22 0 pass adjust_nlfs.cs-32x32x1
Y Y Y Y 16 16 16 16 0 16 16 16 16 16 16 16 16 16 22 0 0 pass adjustment.128x64x1
Y Y Y Y 16 16 16 16 0 16 16 16 16 16 16 0 0 16 16 0 0 pass adjustment.cs-32x32x1
Y Y Y Y 16 16 13 2 0 16 16 2 0 16 16 3 2 16 16 3 2 pass advect_cs
Y Y Y Y 16 22 16 16 16 16 16 16 13 16 16 16 16 16 16 16 16 pass advect_xy
Y Y Y Y 16 22 16 1 1 16 16 16 16 16 16 0 1 16 16 16 16 pass advect_xz
Y Y Y Y 16 16 16 2 0 16 16 1 1 16 16 0 1 16 16 0 2 pass aim.5l_Equatorial_Channel
Y Y Y Y 16 16 16 2 0 16 16 0 1 16 16 1 1 16 16 0 1 pass aim.5l_LatLon
Y Y Y Y 16 16 16 1 0 0 16 0 0 16 16 0 1 16 16 1 1 pass aim.5l_cs
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 12 13 16 16 13 16 pass exp0
Y Y Y Y 16 16 16 8 8 16 16 12 22 16 16 1 1 16 16 1 1 pass exp1
Y Y Y Y 16 16 16 0 0 16 16 3 0 16 16 0 0 16 16 0 0 pass exp2
Y Y Y Y 16 16 16 0 2 16 16 16 22 0 16 0 0 16 16 0 1 pass exp4
Y Y Y Y 16 16 16 12 11 16 16 11 0 16 16 6 12 16 16 6 12 pass exp5
Y Y Y Y 16 16 16 1 1 16 16 1 3 16 16 0 0 16 16 0 0 pass front_relax
Y Y Y Y 4 5 7 7 5 4 9 8 5 3 3 4 5 3 3 4 5 FAIL global_ocean.90x40x15
Y Y Y Y 16 16 16 2 0 16 16 16 16 16 16 1 1 16 16 0 1 pass hs94.128x64x5
Y Y Y Y 16 16 16 2 0 16 16 16 16 16 16 1 1 16 16 22 1 pass hs94.1x64x5
Y Y Y Y 14 16 16 1 1 16 16 16 16 13 16 0 1 16 16 0 1 pass hs94.cs-32x32x5
Y Y Y Y 16 16 16 10 13 0 16 1 0 16 16 1 2 16 16 16 16 pass internal_wave
Y Y Y Y 16 16 16 0 0 16 16 1 0 16 16 0 0 16 16 0 0 pass natl_box
Y Y Y Y 16 16 16 16 16 16 16 16 22 16 16 16 16 16 16 16 16 pass plume_on_slope
Y Y Y Y 16 16 16 16 0 0 16 0 0 16 16 1 2 16 16 1 2 pass solid-body.cs-32x32x1
Output after regenerating output:
T S U V
C D M c m s m s m s m s
n p a R g m m e . m m e . m m e . m m e .
f n k u 2 i a a d i a a d i a a d i a a d
g d e n d n x n . n x n . n x n . n x n .
Y Y Y Y 16 16 16 16 0 16 16 16 16 16 16 22 0 16 16 22 0 pass adjust_nlfs.cs-32x32x1
Y Y Y Y 16 16 16 16 0 16 16 16 16 16 16 16 16 16 22 0 0 pass adjustment.128x64x1
Y Y Y Y 16 16 16 16 0 16 16 16 16 16 16 0 0 16 16 0 0 pass adjustment.cs-32x32x1
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass advect_cs
Y Y Y Y 16 22 16 16 16 16 16 16 13 16 16 16 16 16 16 16 16 pass advect_xy
Y Y Y Y 16 22 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass advect_xz
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass aim.5l_Equatorial_Channel
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass aim.5l_LatLon
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass aim.5l_cs
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 12 13 16 16 13 16 pass exp0
Y Y Y Y 16 16 16 16 16 16 16 16 22 16 16 16 16 16 16 22 16 pass exp1
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass exp2
Y Y Y Y 16 16 16 16 16 16 16 16 22 16 16 16 16 16 16 16 16 pass exp4
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 22 16 16 16 22 16 pass exp5
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass front_relax
Y Y Y Y 4 5 7 7 5 4 9 8 5 3 3 4 5 3 3 4 5 FAIL global_ocean.90x40x15
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 22 16 16 16 16 16 pass hs94.128x64x5
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 22 16 pass hs94.1x64x5
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 22 16 16 16 22 16 pass hs94.cs-32x32x5
Y Y Y Y 16 16 16 22 16 16 16 16 16 16 16 22 16 16 16 16 16 pass internal_wave
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass natl_box
Y Y Y Y 16 16 16 16 16 16 16 16 22 16 16 16 16 16 16 16 16 pass plume_on_slope
Y Y Y Y 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 pass solid-body.cs-32x32x1
Note: global_ocean.90x40x15 was not regenerated since it's output was generated
later when changes were made to pkg/gmredi (Dec 4,'01).
Updated results after changing timestep.F & timestep_tracer.F
Since the model results are sensitive to small changes in the code when compiling with the older g77, it seems safer to use the -ffloat-store option when making tests. testscript now calls genmake with the -ieee option by default. To disable this, use testscript -noieee. If changing between with and without -noieee, it is useful to use -clean the first time you use testscript to forc a make clean.
Changed initialization of Adams-Bashforth time stepping which consequently changes most output: exp0, exp1, exp2, exp4, exp5, front_relax, hs94.128x64x5, internal_wave and natl_box had substantial changes in numerical output adjustment.128x64x1, adjustment.cs-32x32x1, aim.5l_LatLon, hs94.1x64x5 and hs94.cs-32x32x5 were not affected but output checked in anyway
Across the board regeneration of results/output.txt o format of output differs o some numbers in some expts were inaccurate/out of date o now can test model using more than cg2d_init_res
Generate new output for OBCs fixes.
Made length of output.txt correspond to "data".
Restructuring verification experiments.
This form allows you to request diffs between any two revisions of this file. For each of the two "sides" of the diff, select a symbolic revision name using the selection box, or choose 'Use Text Field' and enter a numeric revision.
| ViewVC Help | |
| Powered by ViewVC 1.1.22 |