Tags:
create new tag
view all tags

Mode Parameters Deployment Results

First trial: 02/02/2007

Cases 1a, 1b, 2a, 2b, 3a and 3b ran successfully

Timing

Note: obtaining output targets for cases 1a, 2a and 3a were obtained by executing % optimal_sph_coeff 1 1 20

obtaining output targets for cases 1b, 2b and 3b were obtained by executing % optimal_sph_coeff 10 0 15

1. 1a, 2a, 3a. Optimal Time Series for Target l=1 m=1 n=20 Fri Feb 2 17:26:59 GMT 2007 11.48user 1.69system 0:13.17elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+25103minor)pagefaults 0swaps

2. 1b, 2b, 3b. Optimal Time Series for Target l=10 m=0 n=15 Fri Feb 2 17:29:50 GMT 2007 20.91user 3.70system 0:24.62elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+25526minor)pagefaults 0swaps

Output

See attached timeseries, profiles and screen output text files.

Discussion

  1. Output filenames are hardcoded into optimal_sph_coeff.c, so changing input or output filename requires code recompilation. Not optimal for running program in batch mode for output filenames, but it should be discussed whether this will be a requirement in production (JSOC pipeline, AstroGrid CEA and SolarSoft)
  2. Input filenames (for leakage coefficients) are hardcoded into optimal_constants.h, as are directories for MDI_SHT files. Highest degree (l_max) and leakage width (delta_lm) are hardcoded into constants.h. Changing any of these names or values requires code recompilation. Similar problems with running in batch mode as above, but should discuss whether problem is relevant to production mode.
  3. Code assumes that input files will be in same directory as executable. This is not suitable for AstroGrid CEA execution; it should be discussed if it is suitable for JSOC pipeline.
  4. Ideally, parameters, input filenames and output filenames should be entered by the user either as 1) commandline parameters or 2) keyword / value pairs in a config.txt file that is read at runtime.
  5. Code was fixed to look in two directories for MDI_SHT frequencies files. These directory names are hardcoded into optimal_constants.h, so changing directory names requires code recompilation. (Probably not a problem in production as files are static.) The need for two directories was specific to msslxx since one disk partition was not large enough to hold all 42 GB files; however, problem may be extrapolated to requirement for n directories in production; should address how this will best be handled.

-- ElizabethAuden - 01 Feb 2007

Topic attachments
I Attachment History Action Size Date Who Comment
Unknown file formatdat l10_m0_n15_profile.dat r1 manage 29.6 K 2007-02-02 - 17:54 ElizabethAuden Case 2b: profile for l=10, m=0, n=15
Texttxt l10_m0_n15_screen_output.txt r1 manage 186.8 K 2007-02-02 - 17:56 ElizabethAuden Case 3b: screen output for l=10, m=0, n=15
Unknown file formatdat l10_m0_n15_timeseries.dat r1 manage 19837.1 K 2007-02-02 - 17:51 ElizabethAuden Case 1b: timeseries for l=10, m=0, n=15
Unknown file formatdat l1_m1_n20_profile.dat r1 manage 29.7 K 2007-02-02 - 17:50 ElizabethAuden Case 2a: profile for l=1, m=1, n=20
Texttxt l1_m1_n20_screen_output.txt r1 manage 266.8 K 2007-02-02 - 17:51 ElizabethAuden Case 3a: screen output for l=1, m=1, n=20
Unknown file formatdat l1_m1_n20_timeseries.dat r1 manage 19833.6 K 2007-02-02 - 17:49 ElizabethAuden Case 1a: timeseries for l=1, m=1, n=20
Texttxt testModeParameters130207.txt r1 manage 146.3 K 2007-02-14 - 13:37 ElizabethAuden test results from Mode Parameters Unit Tests
Edit | Attach | Watch | Print version | History: r3 < r2 < r1 | Backlinks | Raw View | More topic actions
Topic revision: r3 - 2007-02-14 - ElizabethAuden
 
This site is powered by the TWiki collaboration platform Powered by PerlCopyright © 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback