| Subject: | STATUS OF SMA MIRIAD PIPELINE -
|
|
for handling new format data taken in various SMA correlator modes |
| during the
science operations
|
| Date: | June 9, 2015[1] |
From: | Jun-Hui Zhao & Glen R. Petitpas |
[1]Updated from the versions issued on Arpil 7, April 13, April 28, 2015 |
Since November 15, 2014 when SMA announced that the storage of SMA data taken
with all the observing modes in the SMA operation is switched to the new format
system, a few updated versions for SMA Miriad have been exported to a CfA public computer
in Hilo (hilodr2) at the SMA site in Hilo, Hawaii for SMA data reduction. Mac Cooper
has been working on a dupilication system to fetch the Miriad on hilodr2 to Cambridge.
An early report on the status of handling the SMA data stored in the new format was
issued in December 2014. In this SMA Technical Memo, we are reporting the status of
the Miriad pipeline in transferring SMA data into Miriad. We have examined all the
data taken for the science tracks in the period between December 22, 2014 and March 31,
2015 including a total of 128 data sets. The observations have covered numerious
observing modes including the ASIC correlator of 4 GHz band, the hybrid correlators
of the ASIC[2] 4 GHz band plus 2 SWARM bands, dual receivers with 2 GHz band for
polarization or non-polarization, ASIC 4 GHz polarization and ASIC 2 GHz polarization.
The new Miriad pipeline successfully transfers 114 SMA data sets to Miriad formats.
We note the origins of the problems caused by correlator configuring errors, initializing
SWARM correlator, hardware overheating, resetting RM errors, recording errors with online
software. The relevant error messages reported on screen from the Miriad pipeline for
those problematic data sets are provided with further diagnosis of the data errors and
information. The test was done with Miriad SWARM3.0.0. The problems in six data sets
that were initially failed in reading a bad data string in an ancillary data table have
been fixed with special patches implemented in the updated version Miriad SWARM3.0.1.
[2]The SMA ASIC correlator was referred as Legacy correlator.
- 1. Miriad software for SMA -
Since the SMA-Miriad version (Miriad SMA-WB2.0.0) for the pipeline in handling
SMA data taken in the new format, the version released on the CfA public computers
has been updated and new versions of pipeline will have been exported via the duplication
system on hilodr2 at the SMA Hilo site. Before the duplication system is fully function,
a temporary service has been provided on rglinux12. The version Miriad SWARM3.0.0
was used for the pipeline status tests described in this memo.
As we have discussed in the SMA Technical Memo (December 2015) for the
Status of SMA Miriad WB for handling SMA wideband data in new format,
one of the ASCII files under the subdirectory for each of the raw SMA data sets
provides two integers to code the observing mode concerning number of receivers
(1 for single and 2 for dual) and the correlator mode (4 for 4 GHz or doubleband
with 48 spectral chunks, and 2 for 2 GHz usually with 24 spectral chunks) that
are used in observations. So far, we have detected three general modes, i.e.
(1,4), (2,2), and (2,1). However, unlike the other modes, the two integers in
the (2,1) mode appear to correspond to the size of band in GHz and receiver codes,
respectively, i.e. single receiver with 2 GHz band; this mode seems to be
particularly used for polization. Under each of the three general observing modes,
there are specific correlator combinations to be used. For example, the mode (1,4)
may contain either only the ASIC correlator of 4 GHz band data with various correlator
configurations applied containing one pseudo continuum channel at the first spectral
window following by number of ≤48 spectral windows or a hybrid configuration
with the ASIC correlator plus two additional broadband windows from the SWARM
correlator. Occationally, the correlators are configured for four polarization
correlations with either one or two receivers.
Table 1 lists the results for the data sets observed in the new data format in the
period between December 22, 2014 and April 1, 2015. Column 1 gives the name of all
the data files from the science category under the SMA archival system in Hawaii
(/sma/data/science/~) for the period of the observations applied to this report.
Column 2 gives the handling status of data sets for the Miriad pipeline SMALOD: "OK"
means readable, and "??" marks the problematic data sets that may need further
information to fix. Column 3 lists the two integer code for modeInfo; the observing
modes of the data appear to be separated into three in general, i.e.[1,4], [2,2] and
[2,1]. Additional codes are given for futher information on the specific observing
modes:
+2SW for two broadband SWARM spectral windows;
Pol for polarization;
in case of dual receiver modes, the relevant receiver code are given,
rx=0 for 230 band receiver;
rx=1 for 340 band receiver;
rx=2 for 400 band receiver;
rx=3 for 690 band receiver;
The byte size (GB) for each of the converted SMA visibility data sets is given in Column 4.
Column 5 lists the error messages reported in the Miriad pipeline for those problematic data
sets. We also add superscript numbers to table 1 to note the origins of the errors and
possibility for repair. The origins of the problems were caused by correlator configuring
error, initializing SWARM correlator, hardware overheating, resetting RM errors, recording
errors with online software and etc..
Table 1: Observing mode
(1) (2) (3) (4) (5)
File name Status modeInfo Size Report errors & test notes
2014Dec
001- 141222_03:50:32 OK 1 4 0.84G
002- 141227_03:52:23 OK 1 4 0.82G
003- 141228_10:00:25 OK 1 4 0.21G
004- 141229_05:39:16 OK 1 4 0.30G
005- 141229_14:47:54 OK 1 4 0.29G
006- 141230_03:49:11 OK 1 4 0.02G
007- 141230_04:19:43 ?? 1 4 ≤0.01G Segmentation fault1
008- 141230_04:31:35 OK 1 4 0.44G
2015Jan
009- 150101_03:34:22 OK 1 4 0.50G
010- 150111_03:22:27 OK 1 4 0.37G
011- 150112_03:09:00 OK 1 4 0.82G
012- 150113_03:26:19 OK 1 4 0.15G
013- 150113_07:09:10 OK 1 4 0.40G
014- 150114_09:24:35 OK 1 4+2SW 7.33G
015- 150115_11:01:10 OK 1 4+2SW 4.58G
016- 150116_18:54:14 OK 1 4 0.06G Segmentation fault (sour)6
017- 150117_03:23:11 OK 1 4+2SW 10.95G
018- 150118_05:47:30 OK 1 4 1.24G
019- 150119_03:35:52 OK 1 4+2SW 14.72G
020- 150120_10:28:51 OK 1 4 0.47G
021- 150122_03:48:30 OK 1 4+2SW 15.84G
022- 150123_03:35:39 OK 1 4+2SW 5.13G Observing Date: 1959 Jan 022
023- 150123_15:01:14 OK 1 4+2SW 3.34G
024- 150123_21:15:35 ?? 1 4 0.01G Miriad needs updating3
025- 150123_23:28:57 OK 1 4 0.08G
026- 150126_03:51:31 OK 1 4+2SW 7.50G
027- 150127_03:50:01 OK 1 4+2SW 11.08G
028- 150128_05:21:11 OK 1 4 0.39G
029- 150130_03:42:21 OK 1 4 0.02G
030- 150130_08:46:40 OK 1 4 0.01G
2015Feb
031- 150201_03:36:56 OK 1 4 0.55G
032- 150201_12:56:44 OK 1 4 0.15G
033- 150202_03:25:12 OK 1 4 0.47G
034- 150202_13:45:43 OK 1 4 0.02G
035- 150203_03:40:05 OK 1 4 1.75G
036- 150203_13:48:47 OK 1 4 0.46G
037- 150204_03:06:38 ?? 2 2 0.55G Fatal: number of antennas = 3.7015624
038- 150205_14:04:24 ?? 2 1 Pol 0.44G Fatal: number of antennas = 6.8442894
039- 150206_04:59:40 OK 1 4 0.43G
040- 150207_05:42:29 ?? 2 1 Pol 1.73G Fatal: number of antennas = 8.2620875
041- 150208_04:29:54 OK 1 4+2SW 4.07G
042- 150208_12:18:48 OK 1 4+2SW 2.16G
043- 150209_05:18:28 OK 1 4 0.16G Segmentation fault (Sour)6
044- 150209_08:14:05 OK 1 4 0.63G
045- 150210_04:23:33 OK 1 4 0.46G
046- 150212_05:21:54 ?? 2 2 1.82G Fatal: number of antennas = 4.2749177
047- 150212_16:29:48 OK 1 4 0.12G
048- 150213_02:32:15 OK 1 4 0.84G
049- 150213_11:01:12 OK 1 4 0.31G
050- 150214_03:58:01 OK 1 4 0.49G
051- 150216_03:34:25 OK 1 4 0.73G Segmentation fault (Sour)8
052- 150216_10:55:54 OK 1 4 0.34G
053- 150217_04:10:12 OK 1 4 1.56G
054- 150218_03:46:42 OK(10,950) 1 4+2SW 4.35G
055- 150218_12:04:09 OK 1 4 0.01G Pre-SW track9
056- 150218_12:20:40 ?? 1 4 ≤0.01G Segmentation fault9
057- 150218_12:26:18 OK 1 4 0.03G Pre-SW track9
058- 150218_12:39:13 OK 1 4 0.05G Pre-SW track9
059- 150218_12:47:38 OK 1 4+2SW 4.06G
060- 150219_03:14:04 ?? 1 4 0.02G Segmentation9
061- 150219_03:33:28 ?? 1 4 0.01G Segmentation9
062- 150219_03:45:59 OK 1 4 0.01G Pre-SW track9
063- 150219_03:55:56 OK 1 4+2W 0.35G Segmentation fault (Sour)10
064- 150220_04:07:43 OK 1 4 0.20G
065- 150220_05:47:05 OK 1 4 0.33G
066- 150220_09:33:40 OK 1 4 0.98G
067- 150221_02:50:36 OK 1 4 0.64G
068- 150222_03:50:13 OK 1 4 1.11G
069- 150223_04:06:17 OK 1 4+2SW 0.05G
070- 150223_04:33:24 OK 1 4+2SW 4.14G
071- 150223_11:23:32 OK 1 4+2SW 5.31G
072- 150224_04:39:08 ?? 1 4+2SW 0.02G Pre-SW track
073- 150224_04:47:03 OK 1 4+2SW 0.19G
074- 150224_05:33:03 OK(1,695) 1 4+2SW 2.05G
075- 150224_10:58:07 OK(1,477) 1 4+2SW 1.41G
076- 150224_15:14:52 OK(1,535) 1 4+2SW 1.57G
077- 150225_03:21:49 OK(1,1895) 1 4+2SW 8.34G
078- 150225_13:48:07 OK 1 4 0.09G
079- 150226_12:24:44 OK 1 4 0.01G
080- 150227_03:14:20 OK(1,955) 1 4+2SW 4.23G
081- 150227_12:00:28 OK 1 4+2SW 0.04G
082- 150227_12:06:54 OK 1 4+2SW 5.85G
083- 150228_05:34:15 OK 1 4 0.35G
2015Mar
084- 150302_17:32:24 OK 1 4 0.06G
085- 150306_04:50:05 ?? 2 2 0.51G Fatal: number of antennas = 4.27491711
086- 150306_13:47:44 OK 1 4 0.50G
087- 150307_02:24:55 OK 1 4 0.20G
088- 150307_05:37:45 OK 1 4 0.04G
089- 150307_05:45:40 OK 1 4 0.29G
090- 150307_11:02:49 OK 1 4 0.09G
091- 150308_04:38:50 OK 1 4 1.09G
092- 150308_11:35:41 OK 1 4 0.09G
093- 150317_04:23:41 OK 1 4 0.12G
094- 150319_04:13:53 OK 1 4 0.29G
095- 150320_04:02:29 OK 1 4 0.01G Polarization track12
096- 150320_04:16:54 OK 1 4 0.07G Polarization track12
097- 150320_05:11:15 OK 1 4 0.06G Polarization track12
098- 150320_05:38:14 OK 1 4 0.01G Polarization track12
099- 150320_05:40:07 OK 1 4 0.01G Polarization track12
100- 150320_05:42:01 OK 1 4 0.05G Polarization track12
101- 150320_05:49:27 OK 1 4 0.11G Polarization track12
102- 150320_06:38:29 ?? 1 4 ≤0.01G Segmentation fault12
103- 150320_06:47:49 OK 1 4 0.49G Polarization track12
104- 150320_10:36:12 OK 1 4 0.01G Polarization track12
105- 150320_10:46:58 OK 1 4 0.22G Polarization track12
106- 150320_12:50:32 ?? 1 4 ≤0.01G Segmentation fault12
107- 150320_13:03:39 OK 1 4 Pol 0.41G Polarization track12
108- 150322_06:24:51 OK 2 2 rx=0 9.25G
109- 150322_06:24:51 OK 2 2 rx=1 9.25G
110- 150323_05:14:40 OK 1 4 0.29G
111- 150324_03:31:48 OK 1 4 0.13G
112- 150324_04:12:02 OK 1 4 1.35G
113- 150325_03:03:28 OK 1 4 0.15G
114- 150325_11:04:08 OK 1 4 1.57G
115- 150326_04:58:27 OK 2 2 rx=0 12.09G
116- 150326_04:58:27 OK 2 2 rx=2 12.09G
117- 150327_05:15:59 ?? 2 2 11.11G Segmentation fault (Tsys)13
118- 150328_05:20:03 OK 2 2 rx=0 10.93G
119- 150328_05:20:03 OK 2 2 rx=2 10.93G
120- 150329_04:51:46 OK 1 4 0.27G
121- 150330_00:09:42 OK 2 2 rx=0 4.25G
123- 150330_00:09:42 OK 2 2 rx=2 4.25G
124- 150330_04:53:54 OK 2 2 rx=0 11.90G
125- 150330_04:53:54 OK 2 2 rx=2 11.90G
126- 150330_19:40:24 OK 1 4+2SW 1.28G
127- 150331_02:46:49 OK 1 4 0.75G Segmentation fault (Sour)14
128- 150331_10:46:11 OK 1 4 0.08G
Notes: (updated April 28, 2015)
1Correlator overheated and needed to shut down some racks and restarted the correlator.
2The wrong observing date reported in the version of of Miriad SWARM 3.0.0 has been fixed in the updated version (3.0.1 or higher).
Data error: the cdh file of the data set provides
two reference times 1) at cd_count=61 ref_time corresponds to
Jan 23, 2015; 2) at cd_count=5113 ref_time corresponds to Jan 2, 1959
3Started setting RM errors so the observing program stopped shortly.
4Polarization tests not a source track.
5Daul receiver polarization, PI = Hirano.
6PI: Glen Petitpas; the source coordinates in inh file contains a bug which has been fixed in the version of Miriad SWARM 3.0.1 or higher.
Data error:The source coordinate string in inh at inhid=342 for 150116_18:54:14 and inhid=348 for 150209_05:18:28
7Daul receiver polarization, PI = Hirano, the data stored incorrectly.
8PI: Srinath; the source coordinates in inh file contains a bug which has been fixed in the version of Miriad SWARM 3.0.1 or higher.
Data error:The source coordinate string in inh at inhid=993
9Useless tracks prior to setting up SWARM correlator.
10PI: Kalfountzov, many sources in the track.
Data error:The source coordinate string in inh at inhid=1882
11PI: Dunham
12Major correlator problems during the polarization track.
13VLBI track, non-standard configuration in the backend.
14PI: Dunham; the source coordinates in inh file contains a bug which has been fixed in the version of Miriad SWARM 3.0.1 or higher.
Data error:The source coordinate string in inh at inhid=1015
For the total of 128 data sets taken between December 22, 2014 and March 31, 2015, the
Miriad pipeline (version Miriad SWARM3.0.0) successfully transfers 108 data sets to Miriad,
covering the modes with the ASIC correlator 4 GHz or 2 GHz bands, hybrid correlators
of the ASIC 4 GHz band with 2 SWARM bands (4GHz), dual receivers for polarization
or non-polarization, ASIC 4 GHz polarization, and/or ASIC 2 GHz polarization. Among the
readable 108 data sets, six files containing bad visibility data in the beginning or
ending of the tracks can cause Segmentation fault; the problem can be resolved simply
with the function nscans in SMALOD; the datasets can be processed by dropping out
the integrations that may contain bad data strings. The readable range of the integrations
are given with two integers in the parenthesis in Column 2 of Table 1 for the easy-repairable
data sets. The usage of SMALOD for 150227_03:14:20 with status = OK(1,955),
for example:
smalod in=150227_03:14:20 out=150227tst rxif=0 \
options=circular sideband=0 nscans=1,955
- 5. Diagnosis of data errors -
For those data sets that are failed to read, SMALOD usually provides error reports on screen.
The Column 5 contains the errors reported on the screen from the exam with
Miriad SWARM3.0.0. Further hints for the cause of the data problem in handling the
problematic files are discussed below, and further information is needed in patching
the pipeline program to read them:
Segmentation fault -
Among the problematic data sets (status=??), the screen reports from the Miriad
pipeline program indicate "Segmentation fault", which usually tells one that those
data contain bad strings in the visibility data files. The seven files with small
byte size (≤0.02G) contains only a few integrations of visibility data, which are
usually produced in the initialization process to get the SWARM correlator online.
"Segmentation fault" may also caused by ancillary bad strings or wrong information in
an ancillary data table in those files. Further information is given in the parenthesis:
"Sour" indicates that bad string occurs in parsing the source names; "Tsys" indicates
that bad data strings contain in the system temperature data.
Fatal: number of antennas -
"Fatal: number of antennas" with a non integer number indicates that the baseline
codes in the five data sets have been packed specially, not following the convention.
They can be fixed if additional information on the format of baseline structure is provided.
Observing Date: 1959 Jan 02 -
For 150123_03:35:39, the Miriad pipeline interprets worngly the observing date.
This indicates that a wrong data string with incorrect observing date
was placed in the entry of refernce time for observation. Usually such a problem
can be repairable with a special patch.
Miriad needs updating -
This usually indicates that a new version of the online code has been
used in the production of the data. A special patch is needed to repair this data.
However, according to the daily operation logs (Glen Petitpas), most of problematic
data sets were casued by initializing SWARM correlator (pre-SW), which contains small size
of bites; a few of them were produced by special testing tracks, such as VLBI with
non-standard configuration in the backend, non-source or dual-receiver polarization tests;
and the rest of them were due to correlator configuring error, hardware overheating,
resetting RM errors, recording errors with online software and etc,;
- 6. Patches (updated May 30, 2015) -
The data recording errors often occur in the beginning or ending of the operation for
an observing track. For data errors containing in the visibility file, SMALOD
prvodes a function to select a range of integrations with no data errors to load into
Miriad. For those errors occurring in ancillary data files, special patches are needed.
For the five data sets 150116_18:54:14, 150209_05:18:28, 150216_03:34:25, 150219_03:55:56
and 150331_02:46:49 with Segmentation fault (Sour) reported in Miriad SWARM3.0.0,
the problems have been fixed with patches implemented in Miriad SWARM3.0.1. For the
data set 150123_03:35:39, an error occurs in the entry ref_time. A patch has been
implemented Miriad SWARM3.0.1. for correcting the error.