Added examples for 3D space charge benchmarking. #141

cemitch99 · 2022-06-09T00:29:19Z

depends on Space Charge Solver #162
adjust resolution via amr.n_cell and geometry.prob_relative
add Python equivalent inputs

The following examples were added to test 3D space charge:

A constant focusing channel with an intense waterbag (6D) beam - beam is rms-matched and second moments should be well-preserved.
A constant focusing channel with an intense Kurth (6D) beam - beam is rms-matched and distribution should be fully stationary.
A periodic focusing channel with an intense Kurth (6D) beam - beam is rms-matched and distribution should vary with the periodicity of the lattice.

Corresponding to 3), an example was added for the same lattice, but without space charge:

A periodic focusing channel with an zero-charge Kurth (6D) beam - beam is rms-matched and distribution should vary with the periodicity of the lattice.

Follow-ups commits:

Add example documentation (partially complete).
Add example using a FODO lattice + short rf cavities.

examples/CMakeLists.txt

ax3l · 2022-06-15T18:40:32Z

Thanks a lot, @cemitch99 !

Do we need to increase tolerances for these tests? I think they should fail as long as space charge is not enabled, right?

cemitch99 · 2022-06-15T21:33:17Z

Thanks a lot, @cemitch99 !

Do we need to increase tolerances for these tests? I think they should fail as long as space charge is not enabled, right?

Yes, they should fail if space charge is not enabled. I will modify the tolerances now.

ax3l · 2022-06-16T04:21:48Z

Thanks. Looks like we need to be even stricter :)

ax3l · 2022-09-16T00:34:05Z

Hi @cemitch99, can you please check if the updates make sense?

I think there is a 4. th test (another Kurth addition) that is not mentioned in the PR description yet. Can you also check all 4 Kurth tests are also described in the README.rst of Kurth? I think that section is not too clear to me yet.

- Modified the initial beam size in the IOTA lens benchmark example. - Added 2 benchmarks of 3D space charge for initial testing. - Add documentation for 2 benchmarks with space charge. - Add a benchmark example with space charge and periodic s-dependent focusing. - Added an s-dependent example using a Kurth beam without space charge. - Modified tolerance for IOTA lens benchmark example. Reduced tolerance to account for smaller initial beam size and improved preservation of invariants of motion. - Modified tolerances of space charge examples to allow CI tests to pass when space charge is not active. - Modified tolerance for space charge examples. These should fail unless space charge is turned on.

examples/kurth/input_kurth_10nC.in

Modified tolerance slightly

Selected amr.n_cell, lattice.nslice parameters, and geometry.prob_relative parameters.

Added nonzero charge. However, this is not a space charge test. It is, rather, another distribution/lattice test. It should be run algo.space_charge = false.

Modified tolerance slightly.

Added appropriate numerical parameters for amr.n_cell, lattice.nslice, and geometry.prob_relative.

1) Improved the accuracy of the target values used for analysis, 2) modified the tolerance slightly

Improved the accuracy of the target values.

Selected numerical values for amr.n_cell, lattice.nslice, and geometry.prob_relative.

Modified the tolerance slightly.

…rge_benchmarks

ax3l · 2022-12-20T18:41:25Z

Hi @cemitch99,

I prepared the branch so far and we just need to make the tolerances a bit more strict I think.
These tests pass even if space-charge is off

cfchannel_spacecharge.analysis
kurth_periodic.analysis (update: ah, has no space charge)

The other two are good: fail if space charge is off and pass if turned on :)

Tightened the tolerance. This test should pass if space charge is OFF, and fail is space charge is ON.

Increased the default charge to match the value in input_kurth_10nC_periodic.in. This test should be run with space charge OFF. If it is run with space charge ON, then it will fail with large error.

examples/kurth/README.rst

cemitch99 · 2022-12-21T18:35:17Z

examples/kurth/input_kurth.in

@@ -35,3 +32,6 @@ constf1.kt = 1.0
 ###############################################################################
 algo.particle_shape = 2
 algo.space_charge = false
+
+amr.n_cell = 40 40 32
+geometry.prob_relative = 1.0


The above test, "input_kurth.in", uses only a Kurth distribution and a constant focusing element (without space charge), features that are already tested elsewhere. I suggest that this test can be removed to save time. The two important Kurth tests are "input_kurth_periodic.in" (without space charge) and "input_kurth_10nC_periodic.in" (with space charge).

I agree - thanks for the update, I will add one more push that removes input_kurth.in :)

cemitch99 · 2022-12-21T18:36:22Z

examples/kurth/input_kurth_10nC.in

+
+amr.n_cell = 48 48 40
+#amr.n_cell = 72 72 72 #optional for increased precision
+geometry.prob_relative = 1.0


The above test, "input_kurth_10nC.in", uses only a Kurth distribution and a constant focusing element (with space charge), features that are already tested elsewhere. I suggest that this test can be removed to save time. The two important Kurth tests are "input_kurth_periodic.in" (without space charge) and "input_kurth_10nC_periodic.in" (with space charge).

I agree - thanks for the update, I will add one more push that removes input_kurth_10nC.in :)

Modified documentation of the relevant benchmarks tests using the Kurth distribution, "input_kurth_periodic.in" and "input_kurth_10nC_periodic.in". (One without space charge, and one with space charge.)

cemitch99 · 2022-12-21T18:56:19Z

examples/CMakeLists.txt

+
+# w/ space charge
+add_impactx_test(kurth_10nC
+    examples/kurth/input_kurth_10nC.in


Suggest that the test here, corresponding to "input_kurth_10nC.in", can be removed since the previous test "input_kurth_10nC_periodic.in" is more comprehensive.

Thanks - removed :)

See Chad's GitHub comments for context.

cemitch99 requested review from ax3l and qianglbl June 9, 2022 00:29

cemitch99 changed the title ~~Modified the initial beam size in the IOTA lens benchmark example.~~ Added 2 examples for initial 3D space charge benchmarking. Jun 9, 2022

cemitch99 changed the title ~~Added 2 examples for initial 3D space charge benchmarking.~~ Added examples for initial 3D space charge benchmarking. Jun 14, 2022

cemitch99 changed the title ~~Added examples for initial 3D space charge benchmarking.~~ Added examples for 3D space charge benchmarking. Jun 14, 2022

ax3l reviewed Jun 15, 2022

View reviewed changes

examples/CMakeLists.txt Outdated Show resolved Hide resolved

ax3l added the component: tests examples, tests and benchmarks label Jun 15, 2022

ax3l mentioned this pull request Aug 10, 2022

Space Charge Solver #162

Merged

20 tasks

ax3l force-pushed the add_spacecharge_benchmarks branch from 27f79c5 to b73defc Compare August 31, 2022 22:44

ax3l force-pushed the add_spacecharge_benchmarks branch 4 times, most recently from 207c193 to 051a4e8 Compare September 16, 2022 00:28

ax3l force-pushed the add_spacecharge_benchmarks branch 5 times, most recently from 365b45d to 261f32e Compare September 16, 2022 01:49

ax3l added the component: space charge Space charge & potential solver label Sep 19, 2022

ax3l force-pushed the add_spacecharge_benchmarks branch 2 times, most recently from b7285a9 to 20b6b65 Compare September 20, 2022 17:07

ax3l force-pushed the add_spacecharge_benchmarks branch from 20b6b65 to cd84186 Compare September 20, 2022 17:10

ax3l reviewed Sep 20, 2022

View reviewed changes

examples/kurth/input_kurth_10nC.in Outdated Show resolved Hide resolved

cemitch99 added 2 commits December 5, 2022 14:03

Update analysis_cfchannel_10nC.py

a3394f2

Modified tolerance slightly

Update input_cfchannel_10nC.in

db77bb9

Selected amr.n_cell, lattice.nslice parameters, and geometry.prob_relative parameters.

cemitch99 added 7 commits December 5, 2022 14:16

Update input_kurth_periodic.in

6bdaa41

Added nonzero charge. However, this is not a space charge test. It is, rather, another distribution/lattice test. It should be run algo.space_charge = false.

Update analysis_kurth_periodic.py

4fa0d5e

Modified tolerance slightly.

Update input_kurth_10nC_periodic.in

688461d

Added appropriate numerical parameters for amr.n_cell, lattice.nslice, and geometry.prob_relative.

Update analysis_kurth_10nC_periodic.py

2459e5a

1) Improved the accuracy of the target values used for analysis, 2) modified the tolerance slightly

Update analysis_kurth_periodic.py

9dd7bea

Improved the accuracy of the target values.

Update input_kurth_10nC.in

095bd66

Selected numerical values for amr.n_cell, lattice.nslice, and geometry.prob_relative.

Update analysis_kurth_10nC.py

cedbfb8

Modified the tolerance slightly.

ax3l force-pushed the development branch from 0de9439 to e8378ee Compare December 12, 2022 19:49

Merge remote-tracking branch 'mainline/development' into add_spacecha…

710eee1

…rge_benchmarks

cemitch99 added 2 commits December 20, 2022 11:39

Update analysis_kurth_periodic.py

48c9d93

Tightened the tolerance. This test should pass if space charge is OFF, and fail is space charge is ON.

Update input_kurth_periodic.in

5c502db

Increased the default charge to match the value in input_kurth_10nC_periodic.in. This test should be run with space charge OFF. If it is run with space charge ON, then it will fail with large error.

ax3l approved these changes Dec 21, 2022

View reviewed changes

ConstF Channel: Python 10nC Example

2968c59

ax3l reviewed Dec 21, 2022

View reviewed changes

examples/kurth/README.rst Outdated Show resolved Hide resolved

Python: kurth periodic example

0d75550

cemitch99 commented Dec 21, 2022

View reviewed changes

Update README.rst

af13356

Modified documentation of the relevant benchmarks tests using the Kurth distribution, "input_kurth_periodic.in" and "input_kurth_10nC_periodic.in". (One without space charge, and one with space charge.)

cemitch99 commented Dec 21, 2022

View reviewed changes

ax3l added 2 commits December 22, 2022 17:32

Remove Already Covered Scenarios

88f6035

See Chad's GitHub comments for context.

Python: kurth 10nC periodic

d231348

ax3l self-assigned this Dec 22, 2022

ax3l approved these changes Dec 22, 2022

View reviewed changes

ax3l merged commit 1e7cddc into BLAST-ImpactX:development Dec 22, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Added examples for 3D space charge benchmarking. #141

Added examples for 3D space charge benchmarking. #141

cemitch99 commented Jun 9, 2022 •

edited by ax3l

Loading

ax3l commented Jun 15, 2022

cemitch99 commented Jun 15, 2022

ax3l commented Jun 16, 2022

ax3l commented Sep 16, 2022

ax3l commented Dec 20, 2022 •

edited

Loading

cemitch99 Dec 21, 2022 •

edited

Loading

ax3l Dec 22, 2022 •

edited

Loading

cemitch99 Dec 21, 2022 •

edited

Loading

ax3l Dec 22, 2022

cemitch99 Dec 21, 2022

ax3l Dec 22, 2022

Added examples for 3D space charge benchmarking. #141

Added examples for 3D space charge benchmarking. #141

Conversation

cemitch99 commented Jun 9, 2022 • edited by ax3l Loading

ax3l commented Jun 15, 2022

cemitch99 commented Jun 15, 2022

ax3l commented Jun 16, 2022

ax3l commented Sep 16, 2022

ax3l commented Dec 20, 2022 • edited Loading

cemitch99 Dec 21, 2022 • edited Loading

Choose a reason for hiding this comment

ax3l Dec 22, 2022 • edited Loading

Choose a reason for hiding this comment

cemitch99 Dec 21, 2022 • edited Loading

Choose a reason for hiding this comment

ax3l Dec 22, 2022

Choose a reason for hiding this comment

cemitch99 Dec 21, 2022

Choose a reason for hiding this comment

ax3l Dec 22, 2022

Choose a reason for hiding this comment

cemitch99 commented Jun 9, 2022 •

edited by ax3l

Loading

ax3l commented Dec 20, 2022 •

edited

Loading

cemitch99 Dec 21, 2022 •

edited

Loading

ax3l Dec 22, 2022 •

edited

Loading

cemitch99 Dec 21, 2022 •

edited

Loading