Table of Contents

Beamforming on B0031-07

VCSBeam version

Required: >= v3.2.1

Code Block

language	bash
theme	Midnight

# On Garrawarla, test vcsbeam version
$ module use /pawsey/mwa/software/python3/modulefiles
$ module load vcsbeam
$ make_mwa_tied_array_beam -V
MWA Beamformer v2.18.8_99df2a4

...

Calibration (Hyperdrive, 1320412440)

Code Block

language	bash
theme	Midnight

$ cd /astro/mwavcs/asvo/252057
$ wget -O 1320499816.fits http://ws.mwatelescope.org/metadata/fits?obs_id=1320499816
$ cd vcs/1320499816/cal/1320412440/hyperdrive
$ sbatch hyperdrive.sbatch

/combined $ for freq in {109..132} do sed "s/FREQ/${freq}/" < make_legacy.sbatch > make_legacy_ch${freq}.sbatch sbatch make_legacy_ch${freq}.sbatch done

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1320499816

Code Block

language	bash
theme	Midnight
title	make_legacy.sbatch
collapse	true

#!/bin/bash -l

#SBATCH --nodes=1
#SBATCH --cpus-per-task=1
#SBATCH --ntasks-per-node=1
#SBATCH --mem=370gb
#SBATCH --partition=gpuq
#SBATCH --gres=gpu:1
#SBATCH --time=01:00:00
#SBATCH --account=mwavcs
#SBATCH --export=NONE

module use /pawsey/mwa/software/python3/modulefiles
module load vcsbeam
module load openmpi-ucx-gpu

DATAPATH=/astro/mwavcs/asvo/252057

srun -N 1 -n 1 fine_pfb_offline -m $DATAPATH/1320499816.metafits -b 1320499816 -T 600 -f FREQ -d $DATAPATH

Incoherent Beam

(Currently not tested since v2.18. Results not guaranteed.)

Code Block

language	bash
theme	Midnight

$ cd /astro/mwavcs/vcs/1320499816/vcsbeam/incoh
$ sbatch vcsbeam_incoh.sbatch

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1318345216/vcsbeam/incoh/vcsbeam_incoh.sbatch
collapse	true

#!/bin/bash -l

#SBATCH --nodes=24
#SBATCH --ntasks-per-node=1
#SBATCH --cpus-per-task=1
#SBATCH --mem=370gb
#SBATCH --partition=gpuq
#SBATCH --gres=gpu:1
#SBATCH --time=00:30:00
#SBATCH --account=mwavcs
#SBATCH --nice=0
#SBATCH --export=NONE
 
module use /pawsey/mwa/software/python3/modulefiles
module load vcsbeam
 
srun -N 24 -n 24 make_mwa_incoh_beam \
    -m /astro/mwavcs/vcs/1320499816/vcsbeam/1320499816_legacy.metafits \
    -b 1320499816 \
    -T 600 \
    -f 109 \
    -d /astro/mwavcs/vcs/1320499816/combined \
    -o incoh

After beamforming finishes, fold on B0031-07:

Code Block

language	bash
theme	Midnight

$ module load singularity
$ /pawsey/mwa/singularity/presto/presto.sif prepfold -psr J0034-0721 -nosearch -nodmsearch -start 0.01 incoh*.fits

Image Removed

Calibration

Offline Correlator

/cal/1320412440/hyperdrive/hyperdrive.sbatch
collapse	true

#!/bin/bash -l
#SBATCH --job-name=hyp-$1
#SBATCH --output=hyperdrive.out
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=40
#SBATCH --time=01:00:00
#SBATCH --clusters=garrawarla
#SBATCH --partition=gpuq
#SBATCH --account=mwavcs
#SBATCH --export=NONE
#SBATCH --gres=tmp:50g
#SBATCH --gres=gpu:1
 
module use /pawsey/mwa/software/python3/modulefiles
module load hyperdrive
 
set -eux
which hyperdrive
 
#cd /astro/mwaeor/MWA/data/1090008640
METAFITS=/astro/mwavcs/asvo/252007/1320412440.metafits
 
if [[ ! -r srclist_1000.yaml ]]; then
   hyperdrive srclist-by-beam -n 1000 -m ${METAFITS} --tile-flags 57 /pawsey/mwa/software/python3/srclists/master/srclist_pumav3_EoR0aegean_fixedEoR1pietro+ForA_phase1+2.txt srclist_1000.yaml
fi
 
hyperdrive di-calibrate -s srclist_1000.yaml -d /astro/mwavcs/asvo/252007/*.fits ${METAFITS}

This outputs hyperdrive_solutions.bin. The --tile-flags 57 corresponds to TileName HexE2. It is used because doing it without shows the amplitudes for that tile are much higher than the others.

Using these solutions in vcsbeam requires >= v2.19.13.

Code Block

language	bash
theme	Midnight
title	Plot Hyperdrive solutions

$ mwa_plot_calibration -m 1320499816_legacy.metafits -c 1320412440.fits -C hyperdrive_solutions.bin -R NONE -U 0,0 -O -X -z nan > solution.txt
$ plot_calibration.py --phases_png phases.png --amps_png amps.png solution.txt

Image AddedImage Added

Weirdly, even though I flagged HexE2 in Hyperdrive (corresponding to Antenna #57), it still shows up in these solution plots. This might be a bug in Hyperdrive. Fornow, I'll just also flag it when I try to beamform.

Beamforming

All 24 channels

Beamforming directly from MWAX (which does the fine PFB on GPU memory) requires >= 3.2.1.

Code Block

language	bash
theme	Midnight
title	Run beamforming job

$ cd /astro/mwavcs/vcs/1320499816/vcsbeam/coh
$ wget -O 1320412440_metafits_ppds.fits http://ws.mwatelescope.org/metadata/fits?obs_id=1320412440&include_ppds=1
$ sbatch vcsbeam.sbatch

$ cd cal1320499816/vis $ for freq in {109..132} do sed "s/FREQ/${freq}/" < offline_correlate.sbatch > offline_correlate_${freq}.sbatch sbatch offline_correlate_${freq}.sbatch done

Code Block

language	bash
theme	Midnight

title	/astro/mwavcs/vcs/1320499816/

vcsbeam/

coh/pointings.txt
collapse	true

00:34:08.8703 -07:21:53.409

#!/bin/bash

Code Block

language	bash
theme	Midnight
title	offline_correlate.sbatch
collapse	true

/astro/mwavcs/vcs/1320499816/vcsbeam/coh/flagged_tilenames.txt
collapse	true

HexE2

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1320499816/vcsbeam/coh/vcsbeam.sbatch
collapse	true

#!/bin/bash -l

#SBATCH --nodes=124
#SBATCH --ntasks-memper-node=370gb1
#SBATCH ---partitioncpus-per-task=1
#SBATCH --mem=370gb
#SBATCH --partition=gpuq
#SBATCH --gres=gpu:1
#SBATCH --time=01:00:00
#SBATCH --account=mwavcs
#SBATCH --nice=0
#SBATCH --export=NONE
 
module use /pawsey/mwa/software/python3/modulefiles
module load offline_correlatorvcsbeam
 for
tsrun in {8..599}
do

    gpssecond=$(echo "$t + 1320499816" | bc)
    gpubox=$(echo "FREQ - 108" | bc)-N 24 -n 24 make_mwa_tied_array_beam \
    -m /astro/mwavcs/asvo/252057/1320499816.fits \
    -b 1320499824 \
    -T 592 \
    -f 109 \
     INPUT_DATA_FILE=-d /astro/mwavcs/vcs/1320499816/combined/1320499816_${gpssecond}_chFREQ.dat
    START_SECOND=${gpssecond}asvo/252057 \
    -P pointings.txt \
    -F DUMPS_PER_SECOND=1 # This sets the output time resolution
flagged_tilenames.txt \
    -c 1320412440_metafits_ppds.fits \
              -C /astro/mwavcs/vcs/1320499816/cal/1320412440/hyperdrive/hyperdrive_solutions.bin \
    # (e.g. 20 --> 1/20 = 0.05s = 50 ms)
                    # Minimum allowed resolution is 2 ms
    CHANS_TO_AVERAGE=1 # This sets the output frequency resolution
                   # (e.g. 1 --> 1x10 kHz = 10 kHz)
    GPUBOX_CHANNEL_NUMBER=${gpubox} # This should be chosen to "match" the input channel
                         # This is not easy! (mwalib handles this, but at the
                         # moment, offline_correlator is not using mwalib)
    OUTPUT_PREFIX=1320499816 # Output files begin with this

    srun -N 1 -n 1 offline_correlator \
        -d ${INPUT_DATA_FILE} \
        -s ${START_SECOND} \
        -r ${DUMPS_PER_SECOND} \
        -n ${CHANS_TO_AVERAGE} \
        -c ${GPUBOX_CHANNEL_NUMBER} \
        -o ${OUTPUT_PREFIX}

done

Hyperdrive

Attempt to run Hyperdrive for in-beam calibration (failed!)

Code Block

language	bash
theme	Midnight

$ cd /astro/mwavcs/vcs/1320499816/cal/1320499816/hyperdrive
$ sbatch hyperdrive-p -R NONE -U 0,0 -O -X

Code Block

language	bash
theme	Midnight

$ module load singularity
$ /pawsey/mwa/singularity/presto/presto.sif prepfold -psr J0034-0721 -nosearch -nodmsearch C001_1320499816_00:34:08.87_-07:21:53.41_ch109-132_000?.fits

Image Added

***(Needs to be updated)***

Single Channel

Changing all the number of nodes "24" to "1" in the sbatch script above produces the following:

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1320499816/vcsbeam/coh/vcsbeam.sbatch
collapse	true

#!/bin/bash -l

#SBATCH --nodes=1
#SBATCH --ntasks-per-node=1
#SBATCH --cpus-per-task=1
#SBATCH --mem=370gb
#SBATCH --partition=gpuq
#SBATCH --gres=gpu:1
#SBATCH --time=01:00:00
#SBATCH --account=mwavcs
#SBATCH --nice=0
#SBATCH --export=NONE
 
module use /pawsey/mwa/software/python3/modulefiles
module load vcsbeam
 
srun -N 1 -n 1 make_mwa_tied_array_beam \
    -m /astro/mwavcs/asvo/252057/1320499816.fits \
    -b 1320499824 \
    -T 592 \
    -f 109 \
    -d /astro/mwavcs/asvo/252057 \
    -P pointings.txt \
    -F flagged_tilenames.txt \
    -c 1320412440_metafits_ppds.fits \
    -C /astro/mwavcs/vcs/1320499816/cal/1320412440/hyperdrive/hyperdrive_solutions.bin \
    -p -R NONE -U 0,0 -O -X

Code Block

language	bash
theme	Midnight

$ module load singularity
$ /pawsey/mwa/singularity/presto/presto.sif prepfold -psr J0034-0721 -nosearch -nodmsearch C001_1320499816_00:34:08.87_-07:21:53.41_ch109_000?.fits

Image Added

A good single pulse for testing purposes

Code Block

language	bash
theme	Midnight
title	Run beamforming job

$ cd /astro/mwavcs/vcs/1320499816/vcsbeam/coh
$ sbatch vcsbeam_single_pulse.sbatch

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1320499816/calvcsbeam/1320499816/hyperdrive/hyperdrivecoh/vcsbeam_single_pulse.sbatch
collapse	true

#!/bin/bash -l

#SBATCH --job-name=hyp-$1nodes=24
#SBATCH --output=hyperdrive.out
#SBATCH --nodesntasks-per-node=1
#SBATCH --ntaskscpus-per-nodetask=401
#SBATCH --time=01:00:00mem=370gb
#SBATCH --clusterspartition=garrawarlagpuq
#SBATCH --partitiongres=gpuqgpu:1
#SBATCH --account=mwavcstime=01:00:00
#SBATCH --exportaccount=NONEmwavcs
#SBATCH --gresnice=tmp:50g0
#SBATCH --gres=gpu:1export=NONE
 
module use /pawsey/mwa/software/python3/modulefiles
module load hyperdrive

set -eux
which hyperdrive

#cdvcsbeam
 
srun -N 24 -n 24 make_mwa_tied_array_beam \
    -m /astro/mwaeormwavcs/vcs/MWA1320499816/datavcsbeam/1090008640
METAFITS=1320499816_callegacy.metafits \
if  [[ ! -r srclist_1000.yaml ]]; thenb 1320500016 \
    hyperdrive srclist-by-beam -n 1000 -m ${METAFITS} /pawsey/mwa/software/python3/srclists/master/srclist_pumav3_EoR0aegean_fixedEoR1pietro+ForA_phase1+2.txt srclist_1000.yaml
fi

hyperdrive di-calibrate -s srclist_1000.yaml -d ../vis/*gpubox*.fits ${METAFITS}

Unfortunately, this doesn't work on the gpubox files as written out by the legacy offline_correlator, which doesn't write out most of the metadata info that mwalib needs to correctly parse those files. I'll work on this route later.

On MWAX_CORRELATOR observation 1320412440 (SUCCESS!)

At the moment, Harry is still helping me obtain the files. As it turned out, this obs was already downloaded for Cath and the EoR team, so Harry was able to just copy them to a place I can see them on Garrawarla: /astro/mwavcs/asvo/252007. Then, running them through Hyperdrive is straightfoward:

Code Block

language	bash
theme	Midnight

$ cd /astro/mwavcs/vcs/1320499816/cal/1320412440/hyperdrive
$ sbatch hyperdrive.sbatch-T 1 \
    -f 109 \
    -d /astro/mwavcs/vcs/1320499816/combined \
    -P pointings.txt \
    -F flagged_tilenames.txt \
    -c 1320412440_metafits_ppds.fits \
    -C /astro/mwavcs/vcs/1320499816/cal/1320412440/hyperdrive/hyperdrive_solutions.bin \
    -p -R NONE -U 0,0 -O -X

Code Block

language	bash
theme	Midnight

$ module load singularity
$ /pawsey/mwa/singularity/dspsr/dspsr.sif dspsr -c 1 -D 0 -b 1024 -turns 1 -U 600 -A C001_1320499816_00:34:08.87_-07:21:53.41_ch109-132_0001.fits
$ singularity run -B ~/.Xauthority /pawsey/mwa/singularity/psrchive_tempo2/psrchive_tempo2.sif pav -G -f 4 2021-11-09-13\:29\:57.ar

Image Added

Other processes

Offline FinePFB

vcs/1320499816/cal/1320412440/hyperdrive/hyperdrive.sbatch

Code Block

language	bash
theme	Midnight	title

$ cd /astro/mwavcs/

collapse

true

#!/bin/bash -l
#SBATCH --job-name=hyp-$1
#SBATCH --output=hyperdrive.out
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=40
#SBATCH --time=01:00:00
#SBATCH --clusters=garrawarla
#SBATCH --partition=gpuq
#SBATCH --account=mwavcs
#SBATCH --export=NONE
#SBATCH --gres=tmp:50g
#SBATCH --gres=gpu:1
 
module use /asvo/252057
$ wget -O 1320499816.fits http://ws.mwatelescope.org/metadata/fits?obs_id=1320499816
$ cd /astro/mwavcs/vcs/1320499816/combined
$ for freq in {109..132}
do
    sed "s/FREQ/${freq}/" < make_legacy.sbatch > make_legacy_ch${freq}.sbatch
    sbatch make_legacy_ch${freq}.sbatch
done

Code Block

language	bash
theme	Midnight
title	make_legacy.sbatch
collapse	true

#!/bin/bash -l

#SBATCH --nodes=1
#SBATCH --cpus-per-task=1
#SBATCH --ntasks-per-node=1
#SBATCH --mem=370gb
#SBATCH --partition=gpuq
#SBATCH --gres=gpu:1
#SBATCH --time=01:00:00
#SBATCH --account=mwavcs
#SBATCH --export=NONE

module use /pawsey/mwa/software/python3/modulefiles
module load hyperdrivevcsbeam
module load set -eux
which hyperdrive
 
#cd /astro/mwaeor/MWA/data/1090008640
METAFITSopenmpi-ucx-gpu

DATAPATH=/astro/mwavcs/asvo/252007/1320412440.metafits252057

srun if-N [[1 !-n -r srclist_1000.yaml ]]; then
   hyperdrive srclist-by-beam -n 1000 -m ${METAFITS} --tile-flags 57 /pawsey/mwa/software/python3/srclists/master/srclist_pumav3_EoR0aegean_fixedEoR1pietro+ForA_phase1+2.txt srclist_1000.yaml
fi
 
hyperdrive di-calibrate -s srclist_1000.yaml -d /astro/mwavcs/asvo/252007/*.fits ${METAFITS}

This outputs hyperdrive_solutions.bin. The --tile-flags 57 corresponds to TileName HexE2. It is used because doing it without shows the amplitudes for that tile are much

Using these solutions in vcsbeam requires >= v2.19.13.

1 fine_pfb_offline -m $DATAPATH/1320499816.metafits -b 1320499816 -T 600 -f FREQ -d $DATAPATH

Legacy Offline Correlator

The legacy offline correlator can be run on recombined files created with Offline Fine PFB.

Code Block

language	bash
theme	Midnight

$ cd /astro/mwavcs/vcs/1320499816/cal/1320499816/vis
$ for freq in {109..132}
do
    sed "s/FREQ/${freq}/" < offline_correlate.sbatch > offline_correlate_${freq}.sbatch
    sbatch offline_correlate_${freq}.sbatch
done

#!/bin/bash -l #SBATCH --export=NONE #SBATCH --output=/astro/mwavcs/vcs/1320499816/batch/RTS_1320499816.out #SBATCH --account=mwavcs #SBATCH --clusters=garrawarla #SBATCH --partition=gpuq # #SBATCH --cpus-per-task=1 #SBATCH --mem-per-cpu=10240MB #SBATCH --nice=0 #SBATCH --chdir=/astro/mwavcs/vcs/1320499816/cal/1320499816/rts #SBATCH --time=0:45:00 #SBATCH --nodes=13 #SBATCH --cpus-per-gpu=1 #SBATCH --gres=gpu:1 ncpus=1 export OMP_NUM_THREADS=1 module use /pawsey/mwa/software/python3/modulefiles/ module load vcstools/master module load RTS/master cd /astro/mwavcs/vcs/1320499816/cal/1320499816/rts export UCX_MEMTYPE_CACHE=n srun --export=all -N 25 -n 25 rts_gpu /astro/mwavcs/vcs/1320499816/cal/1320499816/rts/rts_1320499816.in

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1320499816/cal/1320499816/rts/RTS_1320499816.batch
collapse	true

language	bash
theme	Midnight
title	Plot Hyperdrive solutions

$ mwa_plot_calibration -m 1320499816_legacy.metafits -c 1320412440.fits -C hyperdrive_solutions.bin -R NONE -U 0,0 -O -X -z nan > solution.txt
$ plot_calibration.py --phases_png phases.png --amps_png amps.png solution.txt

Image RemovedImage Removed

Weirdly, even though I flagged HexE2 in Hyperdrive (corresponding to Antenna #57), it still shows up in these solution plots. This might be a bug in Hyperdrive. Fornow, I'll just also flag it when I try to beamform. Fingers crossed this solution is still suitable for my observation!

RTS (failed!)

Create source list

Code Block

language	bash
theme	Midnight
title	Create source list

$ cd /astro/mwavcs/vcs/1320499816/cal/1320499816
$ module load srclists/v1.0.0
$ wget -O 1320499816_metafits_ppds.fits http://ws.mwatelescope.org/metadata/fits?obs_id=1320499816&include_ppds=1
$ srclist_by_beam.py -m 1320499816_metafits_ppds.fits -n 1000 -s ${SRCLISTS_DIR}/srclist_pumav3_EoR0aegean_EoR1pietro+ForA.txt

Set up initial RTS job

First, we need a calibration metafits file (mode = 'HW_LFILES'), which for the moment I edited by hand. It doesn't matter whether the file is one with PPDs or without, but apparently some script somewhere is expected the file to be called \[obsid\]_metafits_ppds.fits. So I manually edited the (MWAX VCS) metafits file and just changed the mode, and then made a symbolic link to is called /astro/mwavcs/vcs/1320499816/cal/1320499816/rts/1320499816_metafits_ppds.fits. Then I ran the script from VCSTools:

Code Block

language	bash
theme	Midnight
title	Run the RTS

cd /astro/mwavcs/vcs/1320499816/cal/1320499816/rts
$ calibrate_vcs.py -O 1320499816 -o 1320499816 -m 1320499816_metafits_ppds.fits -s ../srclist_pumav3_EoR0aegean_EoR1pietro+ForA_1320499816_patch1000.txt --offline --nosubmit

There's no good reason I couldn't have left off the --nosubmit option and had it run, but I just wanted to check it worked before trying it.

Run the RTS

Code Block

language	bash
theme	Midnight
title	Run the RTS

$ cd /astro/mwavcs/vcs/1320499816/cal/1320499816/rts
$ sbatch /astro/mwavcs/vcs/1320499816/batch/RTS_1320499816.batch

offline_correlate.sbatch
collapse	true

#!/bin/bash -l

#SBATCH --nodes=1
#SBATCH --mem=370gb
#SBATCH --partition=gpuq
#SBATCH --gres=gpu:1
#SBATCH --time=01:00:00
#SBATCH --account=mwavcs
#SBATCH --export=NONE

module use /pawsey/mwa/software/python3/modulefiles
module load offline_correlator

for t in {8..599}
do

    gpssecond=$(echo "$t + 1320499816" | bc)
    gpubox=$(echo "FREQ - 108" | bc)

    INPUT_DATA_FILE=/astro/mwavcs/vcs/1320499816/combined/1320499816_${gpssecond}_chFREQ.dat
    START_SECOND=${gpssecond}
    DUMPS_PER_SECOND=1 # This sets the output time resolution
                    # (e.g. 20 --> 1/20 = 0.05s = 50 ms)
                    # Minimum allowed resolution is 2 ms
    CHANS_TO_AVERAGE=1 # This sets the output frequency resolution
                   # (e.g. 1 --> 1x10 kHz = 10 kHz)
    GPUBOX_CHANNEL_NUMBER=${gpubox} # This should be chosen to "match" the input channel
                         # This is not easy! (mwalib handles this, but at the
                         # moment, offline_correlator is not using mwalib)
    OUTPUT_PREFIX=1320499816 # Output files begin with this

    srun -N 1 -n 1 offline_correlator \
        -d ${INPUT_DATA_FILE} \
        -s ${START_SECOND} \
        -r ${DUMPS_PER_SECOND} \
        -n ${CHANS_TO_AVERAGE} \
        -c ${GPUBOX_CHANNEL_NUMBER} \
        -o ${OUTPUT_PREFIX}

done

RTS (failed!)

Create source list

/rts/rts_1320499816.in

Code Block

language	bash
theme	Midnight
title	Create source list

$ cd /astro/mwavcs/vcs/1320499816/cal/1320499816

collapse

true

ReadAllFromSingleFile=
BaseFilename=/astro/mwavcs/vcs/1318345216/cal/1318345216/vis/*_GPUBOX
ReadGpuboxDirect=0
UseCorrelatorInput=1

ReadMetafitsFile=1
MetafitsFilename=/astro/mwavcs/vcs/1318345216/1318345216

DoCalibration=
doMWArxCorrections=1
doRawDataCorrections=1
doRFIflagging=0
useFastPrimaryBeamModels=0
generateDIjones=1
applyDIcalibration=1
UsePacketInput=0
UseThreadedVI=1

MaxFrequency=170.24
ObservationFrequencyBase=138.895
ObservationTimeBase=2455846.1252314816
ObservationPointCentreHA=-0.35952189154909175
ObservationPointCentreDec=-5.98757073636307
ChannelBandwidth=0.04
NumberOfChannels=32

CorrDumpsPerCadence=128
CorrDumpTime=1.0
NumberOfIntegrationBins=6
NumberOfIterations=1

StartProcessingAt=0

ArrayPositionLat=-26.7033194
ArrayPositionLong=116.6708152
ArrayNumberOfStations=128

ArrayFile=

SourceCatalogueFile=
$ module load srclists/v1.0.0
$ wget -O 1320499816_metafits_ppds.fits http://ws.mwatelescope.org/metadata/fits?obs_id=1320499816&include_ppds=1
$ srclist_by_beam.py -m 1320499816_metafits_ppds.fits -n 1000 -s ${SRCLISTS_DIR}/srclist_pumav3_EoR0aegean_EoR1pietro+ForA.txt

Set up initial RTS job

First, we need a calibration metafits file (mode = 'HW_LFILES'), which for the moment I edited by hand. It doesn't matter whether the file is one with PPDs or without, but apparently some script somewhere is expected the file to be called \[obsid\]_metafits_ppds.fits. So I manually edited the (MWAX VCS) metafits file and just changed the mode, and then made a symbolic link to is called /astro/mwavcs/vcs/

...

1320499816/cal/1320499816/

...

rts/

...

1320499816_

...

For an as yet unknown reason, this RTS job timed out, even after allowing 4 hours of wall time.

Beamforming

Beamforming directly from MWAX (which does the fine PFB on GPU memory) requires >= 3.2.1.

...

metafits_ppds.fits. Then I ran the script from VCSTools:

Code Block

language	bash
theme	Midnight
title	Run the RTS

cd /astro/mwavcs/vcs/1320499816/cal/1320499816/rts
$ calibrate_vcs.py -O 1320499816 -o 1320499816 -m 1320499816_metafits_ppds.fits -s ../srclist_pumav3_EoR0aegean_EoR1pietro+ForA_1320499816_patch1000.txt --offline --nosubmit

There's no good reason I couldn't have left off the --nosubmit option and had it run, but I just wanted to check it worked before trying it.

Run the RTS

Code Block

language	bash
theme	Midnight
title	Run beamforming jobthe RTS

$ cd /astro/mwavcs/vcs/1320499816/vcsbeamcal/1320499816/cohrts
$ wget -O 1320412440_metafits_ppds.fits http://ws.mwatelescope.org/metadata/fits?obs_id=1320412440&include_ppds=1
$ sbatch vcsbeam.sbatchsbatch /astro/mwavcs/vcs/1320499816/batch/RTS_1320499816.batch

vcsbeamcoh/flaggedtilenames.txt

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1320499816/vcsbeamcal/1320499816/cohrts/pointingsRTS_1320499816.txtbatch
collapse	true

00:34:08.8703 -07:21:53.409

Code Block

language

bash

theme

Midnight

title

#!/bin/bash -l

#SBATCH --export=NONE
#SBATCH --output=/astro/mwavcs/vcs/1320499816/

batch/

RTS_

collapse

true

HexE2

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1320499816/vcsbeam/coh/vcsbeam.sbatch
collapse	true

#!/bin/bash -l

1320499816.out
#SBATCH --account=mwavcs
#SBATCH --nodesclusters=24garrawarla
#SBATCH --ntasks-per-node=1partition=gpuq
#
#SBATCH --cpus-per-task=1
#SBATCH --mem-per-cpu=370gb10240MB
#SBATCH --partitionnice=gpuq0
#SBATCH --gres=gpu:1chdir=/astro/mwavcs/vcs/1320499816/cal/1320499816/rts
#SBATCH --time=010:0045:00
#SBATCH --accountnodes=mwavcs13
#SBATCH --cpus-niceper-gpu=01
#SBATCH --gres=gpu:1

ncpus=1
export=NONE OMP_NUM_THREADS=1


module use /pawsey/mwa/software/python3/modulefiles/
module load vcsbeamvcstools/master


srun -N 24 -n 24 make_mwa_tied_array_beam \
    -m /astro/mwavcs/asvo/252057/1320499816.fits \
    -b 1320499824 \
    -T 592 \
    -f 109 \
    -d /astro/mwavcs/asvo/252057 \
    -P pointings.txt \
    -F flagged_tilenames.txt \
    -c 1320412440_metafits_ppds.fits \
    -C /astro/mwavcs/vcs/1320499816/cal/1320412440/hyperdrive/hyperdrive_solutions.bin \
    -p -R NONE -U 0,0 -O -X

Code Block

language	bash
theme	Midnight
title	Run beamforming job

$ module load singularity
$ /pawsey/mwa/singularity/presto/presto.sif prepfold -psr J0034-0721 -nosearch -nodmsearch C001_1320499816_00:34:08.87_-07:21:53.41_ch109-132_000?.fits

Image Removed

***(Needs to be updated)***

Single Channel

Changing all the number of nodes "24" to "1" in the sbatch script above produces the following:

A good single pulse for testing purposes

Code Block

language	bash
theme	Midnight
title	Run beamforming job

$module load RTS/master


cd /astro/mwavcs/vcs/1320499816/vcsbeam/coh
$ sbatch vcsbeam_single_pulse.sbatchcal/1320499816/rts
export UCX_MEMTYPE_CACHE=n
srun --export=all -N 25 -n 25 rts_gpu /astro/mwavcs/vcs/1320499816/cal/1320499816/rts/rts_1320499816.in

#!/bin/bash -l #SBATCH --nodes=24 #SBATCH --ntasks-per-node=1 #SBATCH --cpus-per-task=1 #SBATCH --mem=370gb #SBATCH --partition=gpuq #SBATCH --gres=gpu:1 #SBATCH --time=01:00:00 #SBATCH --account=mwavcs #SBATCH --nice=0 #SBATCH --export=NONE module use /pawsey/mwa/software/python3/modulefiles module load vcsbeam srun -N 24 -n 24 make_mwa_tied_array_beam \ -m /astro/mwavcs/vcs/1320499816/vcsbeam/1320499816_legacy.metafits \ -b 1320500016 \ -T 1 \ -f 109 \ -d /astro/mwavcs/vcs/1320499816/combined \ -P pointings.txt \ -F flagged_tilenames.txt \ -c 1320412440_metafits_ppds.fits

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1320499816/vcsbeamcal/1320499816/cohrts/vcsbeamrts_single_pulse1320499816.sbatchin
collapse	true

ReadAllFromSingleFile=
BaseFilename=/astro/mwavcs/vcs/1318345216/cal/1318345216/vis/*_GPUBOX
ReadGpuboxDirect=0
UseCorrelatorInput=1

ReadMetafitsFile=1
MetafitsFilename=/astro/mwavcs/vcs/1318345216/1318345216

DoCalibration=
doMWArxCorrections=1
doRawDataCorrections=1
doRFIflagging=0
useFastPrimaryBeamModels=0
generateDIjones=1
applyDIcalibration=1
UsePacketInput=0
UseThreadedVI=1

MaxFrequency=170.24
ObservationFrequencyBase=138.895
ObservationTimeBase=2455846.1252314816
ObservationPointCentreHA=-0.35952189154909175
ObservationPointCentreDec=-5.98757073636307
ChannelBandwidth=0.04
NumberOfChannels=32

CorrDumpsPerCadence=128
CorrDumpTime=1.0
NumberOfIntegrationBins=6
NumberOfIterations=1

StartProcessingAt=0

ArrayPositionLat=-26.7033194
ArrayPositionLong=116.6708152
ArrayNumberOfStations=128

ArrayFile=

SourceCatalogueFile=/astro/mwavcs/vcs/1318345216/cal/1318345216/srclist_pumav3_EoR0aegean_EoR1pietro+ForA_1318345216_patch1000.txt
NumberOfCalibrators=1
NumberOfSourcesToPeel=0
calBaselineMin=20.0
calShortBaselineTaper=40.0
FieldOfViewDegrees=1

For an as yet unknown reason, this RTS job timed out, even after allowing 4 hours of wall time.

Incoherent Beam

(Currently not tested since v2.18. Results not guaranteed. As of this writing, 21 Dec 2021, only supported for legacy data.)

Code Block

language	bash
theme	Midnight

$ cd /astro/mwavcs/vcs/1320499816/vcsbeam/incoh
$ sbatch vcsbeam_incoh.sbatch

Code Block

language	bash
theme	Midnight
title	/astro/mwavcs/vcs/1318345216/vcsbeam/incoh/vcsbeam_incoh.sbatch
collapse	true

#!/bin/bash -l

#SBATCH --nodes=24
#SBATCH --ntasks-per-node=1
#SBATCH --cpus-per-task=1
#SBATCH --mem=370gb
#SBATCH --partition=gpuq
#SBATCH --gres=gpu:1
#SBATCH --time=00:30:00
#SBATCH --account=mwavcs
#SBATCH --nice=0
#SBATCH --export=NONE
 
module use /pawsey/mwa/software/python3/modulefiles
module load vcsbeam
 
srun -N 24 -n 24 make_mwa_incoh_beam \
    -Cm /astro/mwavcs/vcs/1320499816/cal/1320412440/hyperdrive/hyperdrive_solutions.binvcsbeam/1320499816_legacy.metafits \
    -p -R NONE -U 0,0 -O -X

Code Block

language	bash
theme	Midnight

$ module load singularity
$ /pawsey/mwa/singularity/dspsr/dspsr.sif dspsr -c 1 -D 0 -b 1024 -turns 1 -U 600 -A C001_1320499816_00:34:08.87_-07:21:53.41_ch109-132_0001.fits
$ singularity run -B ~/.Xauthorityb 1320499816 \
    -T 600 \
    -f 109 \
    -d /astro/mwavcs/vcs/1320499816/combined \
    -o incoh

After beamforming finishes, fold on B0031-07:

Code Block

language	bash
theme	Midnight

$ module load singularity
$ /pawsey/mwa/singularity/psrchive_tempo2/psrchive_tempo2presto/presto.sif pavprepfold -Gpsr J0034-f0721 4-nosearch 2021-11-09-13\:29\:57.ar

...

nodmsearch -start 0.01 incoh*.fits

Image Added

Versions Compared

Old Version 21

New Version Current

Key

Beamforming on B0031-07

VCSBeam version

Calibration (Hyperdrive, 1320412440)

Incoherent Beam

Calibration

Offline Correlator

Beamforming

All 24 channels

Hyperdrive

Attempt to run Hyperdrive for in-beam calibration (failed!)

Single Channel

Image Added

A good single pulse for testing purposes

On MWAX_CORRELATOR observation 1320412440 (SUCCESS!)

Other processes

Offline FinePFB

Legacy Offline Correlator

RTS (failed!)

Create source list

Set up initial RTS job

Run the RTS

RTS (failed!)

Create source list

Set up initial RTS job

Beamforming

Run the RTS

Single Channel

A good single pulse for testing purposes

Incoherent Beam

Page Comparison

Versions Compared

Old Version 21

New Version Current

Key

Beamforming on B0031-07

VCSBeam version

Calibration (Hyperdrive, 1320412440)

Incoherent Beam

Calibration

Offline Correlator

Beamforming

All 24 channels

Hyperdrive

Attempt to run Hyperdrive for in-beam calibration (failed!)

Single Channel

Image Added

A good single pulse for testing purposes

On MWAX_CORRELATOR observation 1320412440 (SUCCESS!)

Other processes

Offline FinePFB

Legacy Offline Correlator

RTS (failed!)

Create source list

Set up initial RTS job

Run the RTS

RTS (failed!)

Create source list

Set up initial RTS job

Beamforming

Run the RTS

Single Channel

A good single pulse for testing purposes

Incoherent Beam