I tried for about a week to figure out the right format file for the CTIO 4m Blanco MOSAIC2 that would support both full-FPA processing and individual amp- or chip-based processing. In the end I just gave up and treated each reduced amplifier as an FPA in the configuration files. This was easy and simple, but a guide with examples on how the configurations are supposed to be set up for various different detectors would be quite helpful to be able to self-consistently use the IPP for a full processing of images. But for now the first goal is just to see if we can get subtractions from already-processed ESSENCE images and do object detection, lightcurves, and classification.

On our IIC cluster here at Harvard, I ran

ppSub /gpfs1/Panstarrs/smsn/data/v9.0/W/workspace/sm061227/2/wdd8.061227_0404.033_2.sw.fits \ /gpfs1/Panstarrs/smsn/data/v9.0/W/workspace/wtempl_05/2/wdd8.041114_0358.078_2.sw.fits \ -inweight /gpfs1/Panstarrs/smsn/data/v9.0/W/workspace/sm061227/2/wdd8.061227_0404.033_2.sw.noise.fits.gz \ -refweight /gpfs1/Panstarrs/smsn/data/v9.0/W/workspace/wtempl_05/2/wdd8.041114_0358.078_2.sw.noise.fits.gz \ Wsub

ESSENCE and the IPP use a slightly different convention for masking so I didn't use the ESSENCE masks. However, the weight images do have very high variances set for the masked-out regions so it's OK for now not to have the mask. I ran with a very simple, and theoretically inadequate, convolution kernel size. Here's the "ppSub.config" I used. This was with Paul Price's ipp-2.2b ppSub modifications.

### Recipe file for ppSub

KERNEL.TYPE STR ISIS # Kernel type to use (POIS|ISIS|SPAM|FRIES) KERNEL.SIZE S32 5 # Kernel half-size (pixels) SPATIAL.ORDER S32 0 # Spatial polynomial order STAMP.SPACING F32 200 # Typical spacing between stamps (pixels) STAMP.FOOTPRINT S32 5 # Size of stamps (pixels) STAMP.THRESHOLD F32 0 # Flux threshold for stamps (ADU) ITER S32 10 # Number of rejection iterations REJ F32 1.5 # Rejection level (std dev) MASK.BAD STR BLANK,BAD,SAT # Mask value for bad pixels MASK.BLANK STR BLANK # Mask value to give blank pixels @ISIS.WIDTHS F32 3 6 # Gaussian widths for ISIS kernels @ISIS.ORDERS S32 2 2 # Polynomial orders for ISIS kernels SPAM.BINNING S32 2 # Binning in outer region for SPAM kernels INNER S32 5 # Inner half-size for SPAM and FRIES kernels RINGS.ORDER S32 1 # Polynomial order for RING kernels

REGION.SIZE F32 NAN # Size of region over which to solve for one variable convolution; 0 or NAN means one regio