Testing log for EVLA Test1 on AXAF field data. Search through this report for "NOTE" for some encountered problems and comments. running on imager-b: /home/imager-b/smyers/2005-06-22/ aipsrc system.resources.memory: 2048 _______________________________________________________________________________ Summary: The use of imager for wide-field imaging of the AXAF deep field was straightforward, and the w-projection algorithm does indeed provide an efficiency gain over standard faceting. I did not try a direct comparison with AIPS, but did try to explore the effect of changing the imager control parameters. There were a few quirks in the operation of imager (in particular the need to re-invoke setimage if you want to changed from masked to unmasked cleaning) but no serious problems. The setting of masks (clean "boxes") for cleaning was easy to do interactively. It is clear that the next key development is in improving the calibration capability as is was apparent that the image quality in both AIPS and aips++ was now limited by various effects. At various points in the report I make comments or suggestions which are indicated by **** NOTE: **** _______________________________________________________________________________ 2005-06-23 _______________________________________________________________________________ Start axaf.g at 13:49:37 Ended 17:13:00 (elapsed 12203s) Uses around 500M memory for 3000x3000 images ALGO := 'mfclark'; # The algorithm to use for IMSIZE := [3000,3000]; # The image size CELLSIZE := ['2arcsec','2arcsec']; # The cellsize SPWID := [1,2]; # Spectral windows to image NFACETS := 1; # No. of facets WPLANES := 128; # No. of w-projection planes NITER := 50000; # No. of Clean components THRESHOLD:= '50e-3mJy' # The Cleaning threshold. ROBUST := 0.0; # Briggs robustness parameter GAIN := 0.1; # Clean gain ** Starting major cycle 7 ... Clean used 50000 iterations to approach a threshold of 4.75e-05 Finalizing residual images for all fields Final maximum residual = 5.98841e-05 Model 1: max, min residuals = 5.93786e-05, -5.98841e-05 Clean did not reach threshold Restoring beam = 11.4" x 8.6" @ 59.3deg Max in image = 43 mJy/beam Box 1368,1407,1541,1551 RMS = 26 uJy/beam ==> 1654:1 Is there an issue because the provided axaf.g used mfclark and not wfclark? It looks like it was doing W-projection from the log. (Kumar says these are the same for single fields.) _______________________________________________________________________________ 2005-06-24 _______________________________________________________________________________ test1 Emulated Ed's imaging: (see efomalon_axaf.g). IF1 only (Ed claims interference in IF2). No clean boxing. PREFIX := 'test1.'; # This will prepend all output filenames ALGO := 'wfclark'; # The algorithm to use for IMSIZE := [3000,3000]; # The image size CELLSIZE := ['2.5arcsec','2.5arcsec']; # The cellsize SPWID := [1]; # Spectral windows to image NFACETS := 1; # No. of facets WPLANES := 128; # No. of w-projection planes NITER := 20000; # No. of Clean components THRESHOLD:= '70e-3mJy' # The Cleaning threshold. ROBUST := 1; # Briggs robustness parameter GAIN := 0.1; # Clean gain Note: Ed had threshold = '0.06e-4Jy' or 6uJy, but I'm sure that threshold was never reached. Ed also restored with a circular beam 11" x 11". Start stm-axaf.g at 10:40:20 Finished 12:36:46 (elapsed 6986s) *** Starting major cycle 6 Clean used 20000 iterations to approach a threshold of 6.89355e-05 Finalizing residual images for all fields Final maximum residual = 9.76111e-05 Model 1: max, min residuals = 9.53381e-05, -9.76111e-05 Wide-field clean failed Clean did not reach threshold Fitted beam used in restoration: 13.056 by 9.917 (arcsec) at pa 45.8701 (deg) Max in image = 45.3 mJy/beam Box 1377,1408,1542,1554 RMS = 34.1 uJy/beam _______________________________________________________________________________ test1 continued Try cleaning deeper... NITER := 50000; # No. of Clean components THRESHOLD:= '70e-3mJy' # The Cleaning threshold. Start stm-axaf.g at 13:57:38 Finished 17:29:49 (elapsed 12731s) Using single-field SkyEquation for single image Created Sky Equation Clean gain = 0.1, Niter = 50000, Threshold = 0.07 mJy, Algorithm = wfclark Starting deconvolution Starting wide-field clean with 1 facets and 0 outliers W projection using 128 planes Estimating maximum possible W = 20626.5 (wavelengths) Typical wavelength = 0.221147 (m) Scaling in W (at maximum W) = 1.27884 wavelengths per pixel ... Clean used 22475 iterations to approach a threshold of 6.65e-05 ** Starting major cycle 13 Making residual images for all fields Model 1: max, min (weighted) residuals = 6.84092e-05, -6.80468e-05 Reached stopping peak residual = 6.84092e-05 Residual images for all fields are up-to-date Successfully deconvolved image Fitted beam used in restoration: 18.569 by 14.021 (arcsec) at pa 26.8534 (deg) Finished imager::clean 12653.9 real 11562.3 user 425.99 system **** NOTE: Its sort of annoying that clean does not report the final number of iterations reached when it cleans down to a threshold. **** test1.axaf.im.clean : Max= 43.0 mJy test1.axaf.im.clean : RMS= 26.2 uJy (Box 1377,1408,1542,1554) I cannot reproduce the above numbers. On 2005-06-28 I get: test1.axaf.im.clean : Max= 49.3 mJy test1.axaf.im.clean : RMS= 36.3 uJy ______________________________________________________________________________ 2005-06-27 _______________________________________________________________________________ test2 GAIN := 0.05; # Clean gain Start stm-axaf.g at 11:15:00 Finished 14:49:18 (elapsed 12858s) ... Clean used 42516 iterations to approach a threshold of 6.65e-05 *** Starting major cycle 13 Making residual images for all fields Model 1: max, min (weighted) residuals = 6.93657e-05, -6.98299e-05 Reached stopping peak residual = 6.98299e-05 Residual images for all fields are up-to-date Successfully deconvolved image Fitted beam used in restoration: 18.569 by 14.021 (arcsec) at pa 26.8534 (deg) Finished imager::clean 12780.6 real 11629.1 user 408.91 system test2.axaf.im.clean : Max= 49.3 mJy test2.axaf.im.clean : RMS= 37.1 uJy Not as good as gain = 0.1 it seems _______________________________________________________________________________ test3 WPLANES := 64; # No. of w-projection planes Start stm-axaf.g at 15:36:15 Finished 19:21:29 (elapsed 13504s) ... Clean used 22554 iterations to approach a threshold of 6.65e-05 *** Starting major cycle 14 Making residual images for all fields Model 1: max, min (weighted) residuals = 6.65328e-05, -6.649e-05 Reached stopping peak residual = 6.65328e-05 Residual images for all fields are up-to-date Successfully deconvolved image Fitted beam used in restoration: 18.516 by 14.000 (arcsec) at pa 27.1878 (deg) Finished imager::clean 13443 real 12198.8 user 421.95 system test3.axaf.im.clean : Max= 49.3 mJy test3.axaf.im.clean : RMS= 36.4 uJy I do not understand this, it took longer than test1. Not surprisingly did worse. _______________________________________________________________________________ 2005-06-28 _______________________________________________________________________________ test4 ALGO := 'wfclark'; # The algorithm to use for IMSIZE := [3000,3000]; # The image size CELLSIZE := ['2.5arcsec','2.5arcsec']; # The cellsize SPWID := [1]; # Spectral windows to image NFACETS := 1; # No. of facets WPLANES := 128; # No. of w-projection planes NITER := 50000; # No. of Clean components THRESHOLD:= '70e-3mJy' # The Cleaning threshold. ROBUST := 1; # Briggs robustness parameter GAIN := 0.10; # Clean gain CYCFAC := 0.10; # Major cycle speedup factor Try the cyclefactor for major cycle speedup. Start stm-axaf.g at 09:08:38 Finished 10:05:14 (elapsed 3396s) ==> about 3x faster! ... Clean used 436 iterations to approach a threshold of 0.00179607 *** Starting major cycle 2 Making residual images for all fields Model 1: max, min (weighted) residuals = 0.00192668, -0.000544485 Maximum residual = 0.00192668, cleaning down to 7.05451e-05 Processing model 1 Initial maximum residual: 0.00192668 Iteration: 3665, Maximum residual=0.000343751 Iteration: 13665, Maximum residual=0.000183101 Iteration: 23665, Maximum residual=0.00013834 Iteration: 33665, Maximum residual=0.000126034 Iteration: 43665, Maximum residual=0.000109629 Iteration: 50000, Maximum residual=9.72729e-05 Finished Clark clean inner cycle Clean used 50000 iterations to approach a threshold of 7.05451e-05 Finalizing residual images for all fields Final maximum residual = 0.000227096 Model 1: max, min residuals = 0.000164797, -0.000227096 Wide-field clean failed Clean did not reach threshold Fitted beam used in restoration: 18.569 by 14.021 (arcsec) at pa 26.8534 (deg) Finished imager::clean 3312.6 real 2894.07 user 97.59 system Didn't quite reach 70uJy test4.axaf.im.clean : Max= 49.2 mJy test4.axaf.im.clean : RMS= 34.8 uJy test1 and test4 look very similar _______________________________________________________________________________ Remade test1 image using: ALGO := 'wfclark'; # The algorithm to use for IMSIZE := [3000,3000]; # The image size CELLSIZE := ['2.5arcsec','2.5arcsec']; # The cellsize SPWID := [1]; # Spectral windows to image NFACETS := 1; # No. of facets WPLANES := 128; # No. of w-projection planes NITER := 50000; # No. of Clean components THRESHOLD:= '70e-3mJy' # The Cleaning threshold. ROBUST := 1; # Briggs robustness parameter GAIN := 0.10; # Clean gain CYCFAC := 1.5; # Major cycle speedup factor Start stm-axaf.g at 10:34:00 Finished 14:08:24 (elapsed 12864s) Image polarization = Stokes I Using wide-field algorithm with Clark Clean Performing w-plane projection Using single-field SkyEquation for single image Created Sky Equation Clean gain = 0.1, Niter = 50000, Threshold = 0.07 mJy, Algorithm = wfclark Starting deconvolution Starting wide-field clean with 1 facets and 0 outliers W projection using 128 planes Estimating maximum possible W = 20626.5 (wavelengths) Typical wavelength = 0.221147 (m) Scaling in W (at maximum W) = 1.27884 wavelengths per pixel ... Clean used 22475 iterations to approach a threshold of 6.65e-05 *** Starting major cycle 13 Making residual images for all fields Model 1: max, min (weighted) residuals = 6.84092e-05, -6.80468e-05 Reached stopping peak residual = 6.84092e-05 Residual images for all fields are up-to-date Successfully deconvolved image Fitted beam used in restoration: 18.569 by 14.021 (arcsec) at pa 26.8534 (deg) Finished imager::clean 12781.1 real 11586.6 user 395.4 system test1.axaf.im.clean : Max= 49.3 mJy test1.axaf.im.clean : RMS= 36.3 uJy So the speedup in test4 looks to be reasonable. _______________________________________________________________________________ test5 Try different speedup CYCFAC := 0.5; # Major cycle speedup factor Start stm-axaf.g at 15:24:14 Finished 17:23:12 (elapsed 7138) Making empty image: test5.axaf.im Frequency = 1.365 GHz, synthesized continuum bandwidth = 0.025 GHz Preferred polarization representation is circular Image polarization = Stokes I Using wide-field algorithm with Clark Clean Performing w-plane projection Using single-field SkyEquation for single image Created Sky Equation Clean gain = 0.1, Niter = 50000, Threshold = 0.07 mJy, Algorithm = wfclark Starting deconvolution Starting wide-field clean with 1 facets and 0 outliers W projection using 128 planes Estimating maximum possible W = 20626.5 (wavelengths) Typical wavelength = 0.221147 (m) Scaling in W (at maximum W) = 1.27884 wavelengths per pixel ... Clean used 31114 iterations to approach a threshold of 6.65e-05 *** Starting major cycle 7 Making residual images for all fields Model 1: max, min (weighted) residuals = 6.6471e-05, -6.6486e-05 Reached stopping peak residual = 6.6486e-05 Residual images for all fields are up-to-date Successfully deconvolved image Fitted beam used in restoration: 18.569 by 14.021 (arcsec) at pa 26.8534 (deg) Finished imager::clean test5.axaf.im.clean : Max= 49.3 mJy test5.axaf.im.clean : RMS= 33.9 uJy _______________________________________________________________________________ 2005-06-29 _______________________________________________________________________________ test6 Now use test5 to make mask MASK_NAME := 'test5.axaf.im.clean.mask'; Used interactivemask to construct mask image. This worked pretty well, though because you have to refresh the mask manually using the REFRESH button to see the boxes you laid down, which takes a couple of seconds and causes the image to annoyingly blink a few times, its not ideal. **** NOTE: the viewer region creation Tool (under the Image Analysis) works much better, leaving the box from each region on the screen. I think it would nice if that tool had the option to convert the current constructed region to an image mask (a "CREATE MASK" button, with a filename entry window). Also, it would be nice if a circluar region selection was available in addition to the rectangular region tool (the standard viewer button) and the polygon. This is a feature in AIPS that is used fairly often for cleaning. **** Clean using mask first, then finish without mask NITER := 10000; # No. of Clean components THRESHOLD:= '90e-3mJy' # The Cleaning threshold. im.clean(algorithm=ALGO, gain=GAIN, niter=NITER,threshold=THRESHOLD, interactive=DOINTERACTIVE, model=MODEL_IMAGE_NAME, image=CLEAN_IMAGE_NAME, mask=MASK_NAME, residual=RESID_IMAGE_NAME); # Final unmasked deconvolution NITER2 := 50000; # No. of Clean components THRESHOLD2 := '60e-3mJy' # The Cleaning threshold. im.clean(algorithm=ALGO, gain=GAIN, niter=NITER2,threshold=THRESHOLD2, interactive=DOINTERACTIVE, model=MODEL_IMAGE_NAME, image=CLEAN_IMAGE_NAME, mask='', residual=RESID_IMAGE_NAME); Start stm-axaf-mask.g at 17:59:45 Finished 20:00:03 (elapsed 7218s) Starting imager::clean ... Clean used 7677 iterations to approach a threshold of 7.6e-05 *** Starting major cycle 6 Making residual images for all fields Model 1: max, min (weighted) residuals = 7.62335e-05, -7.6029e-05 Reached stopping peak residual = 7.62335e-05 Residual images for all fields are up-to-date Successfully deconvolved image Fitted beam used in restoration: 18.569 by 14.021 (arcsec) at pa 26.8534 (deg) Finished imager::clean 6076.64 real 5399.27 user 209.29 system Starting imager::clean ... Clean used 2130 iterations to approach a threshold of 5.7e-05 *** Starting major cycle 2 Making residual images for all fields Model 1: max, min (weighted) residuals = 5.86411e-05, -5.83172e-05 Reached stopping peak residual = 5.86411e-05 Residual images for all fields are up-to-date Successfully deconvolved image Beam used in restoration: 18.5693 by 14.0217 (arcsec) at pa 26.8534 (deg) Finished imager::clean 1036.06 real 948.83 user 36.93 system test6.axaf.im.clean : Max= 49.3 mJy test6.axaf.im.clean : RMS= 47.7 uJy **** NOTE: It looks like it kept using the mask even in the second run. Evidently it just keeps on using the mask. Kumar says to rerun setimage. **** _______________________________________________________________________________ 2005-06-30 _______________________________________________________________________________ Rerun test6 Start stm-axaf-mask.g at 14:28:00 Finished 17:22:33 (elapsed 10473s) ... *** Starting major cycle 1 ... Initial maximum residual: 0.0490531 Iteration: 77, Maximum residual=0.00865942 Finished Clark clean inner cycle Clean used 77 iterations to approach a threshold of 0.00898035 Iteration: 77, Maximum residual=0.00865942 *** Starting major cycle 2 ... Initial maximum residual: 0.00874609 Iteration: 490, Maximum residual=0.00176092 Iteration: 499, Maximum residual=0.00158552 Finished Clark clean inner cycle Clean used 499 iterations to approach a threshold of 0.00160118 *** Starting major cycle 3 ... Initial maximum residual: 0.00164912 Iteration: 2130, Maximum residual=0.000511785 Iteration: 2194, Maximum residual=0.00039241 Iteration: 2201, Maximum residual=0.00032108 Iteration: 2205, Maximum residual=0.000310646 Iteration: 2208, Maximum residual=0.000307141 Iteration: 2210, Maximum residual=0.000301194 Finished Clark clean inner cycle Clean used 2210 iterations to approach a threshold of 0.00030191 *** Starting major cycle 4 ... Initial maximum residual: 0.00034854 Iteration: 6173, Maximum residual=0.000165811 Iteration: 6447, Maximum residual=0.000103788 Iteration: 6481, Maximum residual=9.0372e-05 Iteration: 6492, Maximum residual=9.16416e-05 Iteration: 6496, Maximum residual=8.55872e-05 Iteration: 6497, Maximum residual=8.53882e-05 Finished Clark clean inner cycle Clean used 6497 iterations to approach a threshold of 8.55e-05 *** Starting major cycle 5 ... Initial maximum residual: 0.000101732 Iteration: 6573, Maximum residual=8.90517e-05 Iteration: 6585, Maximum residual=8.53969e-05 Finished Clark clean inner cycle Clean used 6585 iterations to approach a threshold of 8.55e-05 *** Starting major cycle 6 ... Reached stopping peak residual = 8.56245e-05 Residual images for all fields are up-to-date Successfully deconvolved image Fitted beam used in restoration: 18.569 by 14.021 (arcsec) at pa 26.8534 (deg) Finished imager::clean 5935.41 real 5374.6 user 196.72 system ... *** Starting major cycle 1 Initial maximum residual: 0.000364534 Iteration: 120, Maximum residual=0.000203585 Iteration: 1679, Maximum residual=0.000125622 Iteration: 10229, Maximum residual=0.000111558 Iteration: 19761, Maximum residual=8.58978e-05 Iteration: 20475, Maximum residual=7.09142e-05 Iteration: 20599, Maximum residual=6.89061e-05 Iteration: 20625, Maximum residual=6.69666e-05 Iteration: 20637, Maximum residual=6.6874e-05 Iteration: 20643, Maximum residual=6.68099e-05 Iteration: 20652, Maximum residual=6.68314e-05 Finished Clark clean inner cycle Clean used 20652 iterations to approach a threshold of 6.67367e-05 *** Starting major cycle 2 ... Initial maximum residual: 0.000103349 Iteration: 29177, Maximum residual=6.95917e-05 Iteration: 29917, Maximum residual=6.28429e-05 Iteration: 30036, Maximum residual=5.98937e-05 Iteration: 30072, Maximum residual=5.92645e-05 Iteration: 30078, Maximum residual=5.81069e-05 Iteration: 30082, Maximum residual=5.80182e-05 Iteration: 30089, Maximum residual=5.70562e-05 Iteration: 30093, Maximum residual=5.70014e-05 Iteration: 30094, Maximum residual=5.7011e-05 Iteration: 30097, Maximum residual=5.69938e-05 Finished Clark clean inner cycle Clean used 30097 iterations to approach a threshold of 5.7e-05 *** Starting major cycle 3 ... Initial maximum residual: 6.42151e-05 Iteration: 30285, Maximum residual=5.76807e-05 Iteration: 30305, Maximum residual=5.71334e-05 Iteration: 30310, Maximum residual=5.70046e-05 Iteration: 30311, Maximum residual=5.69943e-05 Finished Clark clean inner cycle Clean used 30311 iterations to approach a threshold of 5.7e-05 *** Starting major cycle 4 ... Reached stopping peak residual = 5.72242e-05 Residual images for all fields are up-to-date Successfully deconvolved image Fitted beam used in restoration: 18.569 by 14.021 (arcsec) at pa 26.8534 (deg) Finished imager::clean 4476.86 real 4155.46 user 112.54 system test6.axaf.im.clean : Max= 49.3 mJy test6.axaf.im.clean : RMS= 33.6 uJy _______________________________________________________________________________ 2005-07-01 _______________________________________________________________________________ test7 Try larger images for test7, unmasked 5000x5000 ALGO := 'wfclark'; # The algorithm to use for IMSIZE := [5000,5000]; # The image size CELLSIZE := ['2.5arcsec','2.5arcsec']; # The cellsize SPWID := [1]; # Spectral windows to image NFACETS := 1; # No. of facets WPLANES := 128; # No. of w-projection planes NITER := 50000; # No. of Clean components THRESHOLD:= '70e-3mJy' # The Cleaning threshold. ROBUST := 1; # Briggs robustness parameter GAIN := 0.10; # Clean gain CYCFAC := 0.1; # Major cycle speedup factor Start stm-axaf.g at 16:27:54 Finished 18:59:45 ... *** Starting major cycle 1 Making residual images for all fields ... Initial maximum residual: 0.0490529 Iteration: 97, Maximum residual=0.00708676 Iteration: 435, Maximum residual=0.00186919 Iteration: 441, Maximum residual=0.00179435 Finished Clark clean inner cycle Clean used 441 iterations to approach a threshold of 0.00179621 *** Starting major cycle 2 ... Initial maximum residual: 0.00192913 Iteration: 3676, Maximum residual=0.000326795 Iteration: 13676, Maximum residual=0.000162765 Iteration: 23676, Maximum residual=0.000113067 Iteration: 33676, Maximum residual=9.96205e-05 Iteration: 43676, Maximum residual=9.70957e-05 Iteration: 50000, Maximum residual=8.75341e-05 Finished Clark clean inner cycle Clean used 50000 iterations to approach a threshold of 7.06406e-05 Finalizing residual images for all fields Final maximum residual = 0.000295352 Model 1: max, min residuals = 0.000140742, -0.000295352 Wide-field clean failed Clean did not reach threshold Fitted beam used in restoration: 18.553 by 14.003 (arcsec) at pa 26.8356 (deg) Finished imager::clean 9038.88 real 8306.57 user 137.79 system test7.axaf.im.clean : Max= 49.2 mJy test7.axaf.im.clean : RMS= 36.0 uJy There are only a few sources outside the original 3000x3000 image. The residual errors are calibration errors of some sort, and the next stage would be to improve the image through better calibration. I also suspect the R vs. L squint imaging correction needs to be applied (I could not find this discussed in the manual or cookbook). If I get a chance I will continue testing incorporating calibration... _______________________________________________________________________________ 2005-07-02 _______________________________________________________________________________ test8 Make interactivemask from test7 => test7.axaf.im.clean.mask ALGO := 'wfclark'; # The algorithm to use for IMSIZE := [5000,5000]; # The image size CELLSIZE := ['2.5arcsec','2.5arcsec']; # The cellsize SPWID := [1]; # Spectral windows to image NFACETS := 1; # No. of facets WPLANES := 128; # No. of w-projection planes ROBUST := 1; # Briggs robustness parameter GAIN := 0.10; # Clean gain CYCFAC := 0.1; # Major cycle speedup factor # First masked deconvolution MASK_NAME := 'test7.axaf.im.clean.mask'; NITER := 40000; # No. of Clean components THRESHOLD:= '80e-3mJy' # The Cleaning threshold. # Final unmasked deconvolution NITER2 := 80000; # No. of Clean components THRESHOLD2 := '60e-3mJy' # The Cleaning threshold. Start stm-axaf-mask.g at 13:56:12 Finished 22:42:40 test8.axaf.im.clean : Max= 6371.0 mJy test8.axaf.im.clean : RMS= 3055.3 uJy **** NOTE: Looking at the logger output (in /home/imager-b/smyers/2005-06-22) it looks like something went wrong on major cycles 4 and 5 of the second unmasked clean. I have no idea what happened here, but it is reproducible with the script stm-axaf-mask.g. Is this a bug? **** _______________________________________________________________________________ 2005-07-29 _______________________________________________________________________________ test9 Final testing using uv compressed data ms, presumably derived from Ed's compressed AIPS data, that I got from Kumar. Otherwise same as test4 MS := 'axaf_avg.ms'; # Name of the MS ALGO := 'wfclark'; # The algorithm to use for IMSIZE := [3000,3000]; # The image size CELLSIZE := ['2.5arcsec','2.5arcsec']; # The cellsize SPWID := [1]; # Spectral windows to image NFACETS := 1; # No. of facets WPLANES := 128; # No. of w-projection planes NITER := 50000; # No. of Clean components THRESHOLD:= '70e-3mJy' # The Cleaning threshold. ROBUST := 1; # Briggs robustness parameter GAIN := 0.10; # Clean gain CYCFAC := 0.10; # Major cycle speedup factor Start stm-axaf-avg.g at 16:39:17 Finished 17:09:01 ( elapsed 1784s ) ... *** Starting major cycle 1 ... Initial maximum residual: 0.0451148 Iteration: 97, Maximum residual=0.00662271 Iteration: 426, Maximum residual=0.00172677 Iteration: 430, Maximum residual=0.00167065 Finished Clark clean inner cycle Clean used 430 iterations to approach a threshold of 0.00167104 *** Starting major cycle 2 Making residual images for all fields Model 1: max, min (weighted) residuals = 0.00177907, -0.000381231 Maximum residual = 0.00177907, cleaning down to 6.65e-05 Processing model 1 Initial maximum residual: 0.00177907 Iteration: 3413, Maximum residual=0.00031658 Iteration: 13413, Maximum residual=0.00018116 Iteration: 23413, Maximum residual=0.000122421 Iteration: 33413, Maximum residual=0.000114782 Iteration: 43413, Maximum residual=9.79205e-05 Iteration: 50000, Maximum residual=8.68345e-05 Finished Clark clean inner cycle Clean used 50000 iterations to approach a threshold of 6.65e-05 Finalizing residual images for all fields Final maximum residual = 0.000206786 Model 1: max, min residuals = 0.000188881, -0.000206786 Wide-field clean failed Clean did not reach threshold Fitted beam used in restoration: 18.781 by 13.973 (arcsec) at pa 26.3632 (deg) Finished imager::clean test9.axaf.im.clean : Max= 45.1 mJy test9.axaf.im.clean : RMS= 24.9 uJy Compare to test4: ( elapsed 3396s ) ==> about 1.9x faster than uncompressed test4.axaf.im.clean : Max= 49.2 mJy test4.axaf.im.clean : RMS= 34.8 uJy **** NOTE: The RMS are different (better in compressed) and the test9 image looks way better. There is clearly something more happening than just reducing the size of the vis set (maybe self-cal and/or extra editing (Ed - was something extra done?] ). BEWARE COMPARISONS BETWEEN aips++ ON THE UNCOMPRESSED ms and AIPS ON THE COMPRESSED!!!! And not just on timing... **** Done with testing for now...