From: David Wells Date: Fri, 13 Apr 2018 23:13:05 +0000 (-0400) Subject: Update the step-32 results. X-Git-Tag: v9.0.0-rc1~142^2 X-Git-Url: https://gitweb.dealii.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=d7073315f5368aec3e268373e1f55b735af78585;p=dealii.git Update the step-32 results. Use a lovely coffee color scheme for the new plots to emphasize the color gradient. --- diff --git a/contrib/utilities/makeofflinedoc.sh b/contrib/utilities/makeofflinedoc.sh index 9f582afeb7..a2b98dd0d0 100755 --- a/contrib/utilities/makeofflinedoc.sh +++ b/contrib/utilities/makeofflinedoc.sh @@ -30,7 +30,7 @@ echo "Downloading images (press ctrl-c to cancel) ..." cd images { trap "echo \"(skipping)\"" SIGINT -wget -q -nd -A svg,png,gif -m -l 1 -np https://www.dealii.org/images/steps/developer/ +wget -q -nd -A svg,png,gif,webm -m -l 1 -np https://www.dealii.org/images/steps/developer/ } rm robots.txt* cd .. diff --git a/examples/step-32/doc/results.dox b/examples/step-32/doc/results.dox index 70e802fff5..6a00095610 100644 --- a/examples/step-32/doc/results.dox +++ b/examples/step-32/doc/results.dox @@ -38,13 +38,13 @@ cluster at Texas A&M University.

Results for a 2d circular shell testcase

-Next, we will run step-32 with the parameter file in the directory. Here we -are using 50 processors. The command to launch is (note that step-32.prm is -the default): +Next, we will run step-32 with the parameter file in the directory with one +change: we increase the final time to 1e9. Here we are using 16 processors. The +command to launch is (note that step-32.prm is the default):
-\$ mpirun -np 50 ./step-32
+\$ mpirun -np 16 ./step-32
 
@@ -54,260 +54,202 @@ this:
+\$ mpirun -np 16 ./step-32
 Number of active cells: 12,288 (on 6 levels)
 Number of degrees of freedom: 186,624 (99,840+36,864+49,920)
 
 Timestep 0:  t=0 years
 
    Rebuilding Stokes preconditioner...
-   Solving Stokes system... 40 iterations.
+   Solving Stokes system... 41 iterations.
    Maximal velocity: 60.4935 cm/year
    Time step: 18166.9 years
    17 CG iterations for temperature
    Temperature range: 973 4273.16
 
-Number of active cells: 16,389 (on 7 levels)
-Number of degrees of freedom: 259,617 (140,300+49,167+70,150)
+Number of active cells: 15,921 (on 7 levels)
+Number of degrees of freedom: 252,723 (136,640+47,763+68,320)
 
 Timestep 0:  t=0 years
 
    Rebuilding Stokes preconditioner...
-   Solving Stokes system... 49 iterations.
-   Maximal velocity: 60.5202 cm/year
-   Time step: 10619.2 years
+   Solving Stokes system... 50 iterations.
+   Maximal velocity: 60.3223 cm/year
+   Time step: 10557.6 years
    19 CG iterations for temperature
    Temperature range: 973 4273.16
 
-Number of active cells: 20,124 (on 8 levels)
-Number of degrees of freedom: 323,145 (175,182+60,372+87,591)
+Number of active cells: 19,926 (on 8 levels)
+Number of degrees of freedom: 321,246 (174,312+59,778+87,156)
 
 Timestep 0:  t=0 years
 
    Rebuilding Stokes preconditioner...
-   Solving Stokes system... 51 iterations.
-   Maximal velocity: 59.9101 cm/year
-   Time step: 5334.14 years
+   Solving Stokes system... 50 iterations.
+   Maximal velocity: 57.8396 cm/year
+   Time step: 5453.78 years
    18 CG iterations for temperature
-   Temperature range: 973 4273.43
+   Temperature range: 973 4273.16
 
-Timestep 1:  t=5334.14 years
+Timestep 1:  t=5453.78 years
 
    Solving Stokes system... 49 iterations.
-   Maximal velocity: 61.0837 cm/year
-   Time step: 5232.24 years
+   Maximal velocity: 59.0231 cm/year
+   Time step: 5345.86 years
    18 CG iterations for temperature
-   Temperature range: 973 4273.43
+   Temperature range: 973 4273.16
 
-Timestep 2:  t=10566.4 years
+Timestep 2:  t=10799.6 years
 
    Solving Stokes system... 24 iterations.
-   Maximal velocity: 62.265 cm/year
-   Time step: 5133.56 years
-   18 CG iterations for temperature
-   Temperature range: 973 4273.43
+   Maximal velocity: 60.2139 cm/year
+   Time step: 5241.51 years
+   17 CG iterations for temperature
+   Temperature range: 973 4273.16
 
 [...]
 
-Timestep 100:  t=270918 years
+Timestep 100:  t=272151 years
 
-   Solving Stokes system... 22 iterations.
-   Maximal velocity: 161.391 cm/year
-   Time step: 1683.55 years
+   Solving Stokes system... 21 iterations.
+   Maximal velocity: 161.546 cm/year
+   Time step: 1672.96 years
    17 CG iterations for temperature
-   Temperature range: 973 4282.45
+   Temperature range: 973 4282.57
 
-Number of active cells: 54,717 (on 8 levels)
-Number of degrees of freedom: 880,626 (477,650+164,151+238,825)
+Number of active cells: 56,085 (on 8 levels)
+Number of degrees of freedom: 903,408 (490,102+168,255+245,051)
 
 
 
 +---------------------------------------------+------------+------------+
-| Total wallclock time elapsed since start    |       170s |            |
+| Total wallclock time elapsed since start    |       115s |            |
 |                                             |            |            |
 | Section                         | no. calls |  wall time | % of total |
 +---------------------------------+-----------+------------+------------+
-| Assemble Stokes system          |       103 |      2.25s |       1.3% |
-| Assemble temperature matrices   |        12 |     0.296s |      0.17% |
-| Assemble temperature rhs        |       103 |      9.66s |       5.7% |
-| Build Stokes preconditioner     |        12 |      1.78s |         1% |
-| Solve Stokes system             |       103 |       149s |        88% |
-| Solve temperature system        |       103 |      1.84s |       1.1% |
-| Postprocessing                  |         3 |     0.398s |      0.23% |
-| Refine mesh structure, part 1   |        12 |     0.599s |      0.35% |
-| Refine mesh structure, part 2   |        12 |     0.216s |      0.13% |
-| Setup dof systems               |        13 |      2.15s |       1.3% |
+| Assemble Stokes system          |       103 |      2.82s |       2.5% |
+| Assemble temperature matrices   |        12 |     0.452s |      0.39% |
+| Assemble temperature rhs        |       103 |      11.5s |        10% |
+| Build Stokes preconditioner     |        12 |      2.09s |       1.8% |
+| Solve Stokes system             |       103 |      90.4s |        79% |
+| Solve temperature system        |       103 |      1.53s |       1.3% |
+| Postprocessing                  |         3 |     0.532s |      0.46% |
+| Refine mesh structure, part 1   |        12 |      0.93s |      0.81% |
+| Refine mesh structure, part 2   |        12 |     0.384s |      0.33% |
+| Setup dof systems               |        13 |      2.96s |       2.6% |
 +---------------------------------+-----------+------------+------------+
 
 [...]
 
-Timestep 6946:  t=1.00001e+08 years
-
-   Solving Stokes system... 12 iterations.
-   Maximal velocity: 8.94391 cm/year
-   Time step: 31505.4 years
-   18 CG iterations for temperature
-   Temperature range: 973 4273.04
-
-
-
 +---------------------------------------------+------------+------------+
-| Total wallclock time elapsed since start    |  8.86e+03s |            |
+| Total wallclock time elapsed since start    |  9.14e+04s |            |
 |                                             |            |            |
 | Section                         | no. calls |  wall time | % of total |
 +---------------------------------+-----------+------------+------------+
-| Assemble Stokes system          |      6949 |       153s |       1.7% |
-| Assemble temperature matrices   |       697 |      18.2s |       0.2% |
-| Assemble temperature rhs        |      6949 |       600s |       6.8% |
-| Build Stokes preconditioner     |       697 |       115s |       1.3% |
-| Solve Stokes system             |      6949 |  7.56e+03s |        85% |
-| Solve temperature system        |      6949 |       123s |       1.4% |
-| Postprocessing                  |       140 |      18.1s |       0.2% |
-| Refine mesh structure, part 1   |       696 |      34.6s |      0.39% |
-| Refine mesh structure, part 2   |       696 |      12.6s |      0.14% |
-| Setup dof systems               |       697 |       130s |       1.5% |
+| Assemble Stokes system          |     47045 |  2.05e+03s |       2.2% |
+| Assemble temperature matrices   |      4707 |       310s |      0.34% |
+| Assemble temperature rhs        |     47045 |   8.7e+03s |       9.5% |
+| Build Stokes preconditioner     |      4707 |  1.48e+03s |       1.6% |
+| Solve Stokes system             |     47045 |  7.34e+04s |        80% |
+| Solve temperature system        |     47045 |  1.46e+03s |       1.6% |
+| Postprocessing                  |      1883 |       222s |      0.24% |
+| Refine mesh structure, part 1   |      4706 |       641s |       0.7% |
+| Refine mesh structure, part 2   |      4706 |       259s |      0.28% |
+| Setup dof systems               |      4707 |  1.86e+03s |         2% |
 +---------------------------------+-----------+------------+------------+
 
-The simulation terminates when the time reaches the 100 million years +The simulation terminates when the time reaches the 1 billion years selected in the input file. You can extrapolate from this how long a simulation would take for a different final time (the time step size ultimately settles on somewhere around 20,000 years, so computing for -one billion years will take 50,000 time steps, give or take 20%). As +two billion years will take 100,000 time steps, give or take 20%). As can be seen here, we spend most of the compute time in assembling linear systems and — above all — in solving Stokes systems. -We can clearly not show all output files produced by this program when -run for a longer period of time, so let us -only show the output from every 2500th time step here (these figures -are from a version of this tutorial program before the introduction of -manifolds, when it was not yet possible to arrange vertices along -circles; the figures may therefore look slightly different if you run -the current version): - - - - - - - - - - - - - - - - - - - - - - - - +To demonstrate the output we show the output from every 1250th time step here: +
- - - - - -
- - - - - -
- - - - - -
-
- + - + - +
- + - + - +
- + - + - +
- + - + - +
- + - + - +
-The last two images show the grid as well as the partitioning of the -mesh for a computation with 10 subdomains on 10 processors. The full dynamics -of this simulation are really only visible by looking at -an animation, for example the one shown on -this site. Beware that this animation is -about 20MB large, though it is well worth watching due to its almost -artistic quality. +The last two images show the grid as well as the partitioning of the mesh for +the same computation with 16 subdomains and 16 processors. The full dynamics of +this simulation are really only visible by looking at an animation, for example +the one shown +on this site. This image is well worth watching due to its artistic quality +and entrancing depiction of the evolution of the magma plumes. If you watch the movie, you'll see that the convection pattern goes through several stages: First, it gets rid of the instable temperature