Software for LBA disk based experiments
- Contemporary messages sorted: [ by date ] [ by thread ] [ by subject ] [ by author ] [ by messages with attachments ]
From: <Chris.Phillips_at_email.protected>
Date: Fri, 12 Nov 2004 12:04:50 +1100 (EST)
Hi All,
I have made a copy of the software needed for the LBA disk experiments on:
http://www.atnf.csiro.au/vlbi/evlbi/download.html
You can either download individual files or a tar file of the whole lot.
Currently the date is meaningless.
The "Recording" software includes the vsib.o kernel module (you should not
have to re-install this!). And vsib_record. This is a name change to wr to
distinguish our (mostly Richard's) changes from the official Finnish
version.
The test software are a few programs I use for testing the raw data files.
Documentation for some of this software can be found on:
http://www.atnf.csiro.au/vlbi/evlbi/software.html
I *will* update this over the next couple of days.
A brief summary of some of the test software:
************
vsib_checker
************
This can be used to check the 1 PPS markers are where you expect them,
calculate the sampler statistics of the data file, to a thorough check of
data from the VSIC countdown test mode (0110/6) or search for a particular
bit pattern in the data file. It works on 8 bit (256 Mbps), 16 bit (512
Mbps) data (ie single and dual DAs recording) and 64 MHz data (single DAS)
in VLBA or AT bit encoding.
Typical usage:
- Check the 1 PPS markers are where we expect them. For 256 Mbps (8 bit)
recording, such as the imaging test (vt001h) typical usage would be
>> vsib_checker TEST002*
For the Huygens style recording (512 Mbps from 2 DAS)
>> vsib_checker -b 16 TEST0123
No output indicates no problems. This is safe to run while recording,
though I suggest you on run it on a single file. When running on lots of
files, use the -fail option to stop running as soon as it finds a
problem.
- Check sampler statistics:
>> vsib_checker -stat TEST0123
>> vsib_checker -stat -vlba TEST1234
>> vsib_checker -b 16 -n 100 TEST234
These will print out the sampler statistics for the 4 states of the
2 bit data. Use:
-n <size> to limit the check to n bytes *rounded up to the internal
buffer size (about 2 MB)
-vlba If the das is configured to use VLBA statistics. Note that
You still expect to see 17/33% statistics (not 50% like the
S2 reports) as this looks at the 4 sampling levels, while the
S2 looks at the two bit streams separately.
-64 The data is from the 64 MHz tests (vt001i)
- Check data from the VSIC count down test mode. This compares each sample
with the previous and next sample and checks the data stream is a series
of monotonically decreasing numbers.
>> vsib_checker -c -fail TEST*
>> vsib_checker -c -b 16 -fail TEST*
Use the first when running in "8 bit" mode (vsib_record -m 3, which is
the default). Use the second version when running in "16 bit" mode
(vsib_record -m).
******
fauto
******
This autocorrelates the data and plots the passband. It works on all forms
of data (8/16 bit and 64 MHz).
Typical usage would be
>> fauto TEST000
Plot the bandpass of channel 1
>> fauto -c 2 TEST000
Plot the bandpass of channel 2
>> fauto -c 1 -c 3 TEST000
Plot the bandpass of channel 1 and 3
>> fauto -if1 1 -if2 4 TEST000
Plot the bandpass of channels 1 to 4
>> fauto -b 16 -if1 1 -if2 8 TEST000
Plot channels 1 to 8 for the Huygens data
Other useful options:
-sp1 First spectral point to plot
-sp2 Last spectral point to plot
-v/-vlba Calculate statistics for VLBA data
-w/-wide/-64 Assume 64 MHz bandwidth data
-sp1 <n> First spectral point to plot
-sp2 <n> Last spectral point to plot
-n/-nlags <n> # lags to correlate
-h/-help List brief help
*NOTE* Currently the only pgplot option is "/xs" (no hard copy). This
will be added soon.
You are forced to auto-correlate the entire recorded file.
I will changes this also soon.
***********
band_invert
***********
This program will invert the bandpass for one or all data channels in a
file (it creates a new file). If probably only works on normal 256 Mbps
recording modes. No guarentee it still works at all. It is a very simple
program so can be quickly modifed to support cases on a as needed basis.
Cheers
Chris
Received on 2004-11-12 12:05:18
Date: Fri, 12 Nov 2004 12:04:50 +1100 (EST)
Hi All,
I have made a copy of the software needed for the LBA disk experiments on:
http://www.atnf.csiro.au/vlbi/evlbi/download.html
You can either download individual files or a tar file of the whole lot.
Currently the date is meaningless.
The "Recording" software includes the vsib.o kernel module (you should not
have to re-install this!). And vsib_record. This is a name change to wr to
distinguish our (mostly Richard's) changes from the official Finnish
version.
The test software are a few programs I use for testing the raw data files.
Documentation for some of this software can be found on:
http://www.atnf.csiro.au/vlbi/evlbi/software.html
I *will* update this over the next couple of days.
A brief summary of some of the test software:
************
vsib_checker
************
This can be used to check the 1 PPS markers are where you expect them,
calculate the sampler statistics of the data file, to a thorough check of
data from the VSIC countdown test mode (0110/6) or search for a particular
bit pattern in the data file. It works on 8 bit (256 Mbps), 16 bit (512
Mbps) data (ie single and dual DAs recording) and 64 MHz data (single DAS)
in VLBA or AT bit encoding.
Typical usage:
- Check the 1 PPS markers are where we expect them. For 256 Mbps (8 bit)
recording, such as the imaging test (vt001h) typical usage would be
>> vsib_checker TEST002*
For the Huygens style recording (512 Mbps from 2 DAS)
>> vsib_checker -b 16 TEST0123
No output indicates no problems. This is safe to run while recording,
though I suggest you on run it on a single file. When running on lots of
files, use the -fail option to stop running as soon as it finds a
problem.
- Check sampler statistics:
>> vsib_checker -stat TEST0123
>> vsib_checker -stat -vlba TEST1234
>> vsib_checker -b 16 -n 100 TEST234
These will print out the sampler statistics for the 4 states of the
2 bit data. Use:
-n <size> to limit the check to n bytes *rounded up to the internal
buffer size (about 2 MB)
-vlba If the das is configured to use VLBA statistics. Note that
You still expect to see 17/33% statistics (not 50% like the
S2 reports) as this looks at the 4 sampling levels, while the
S2 looks at the two bit streams separately.
-64 The data is from the 64 MHz tests (vt001i)
- Check data from the VSIC count down test mode. This compares each sample
with the previous and next sample and checks the data stream is a series
of monotonically decreasing numbers.
>> vsib_checker -c -fail TEST*
>> vsib_checker -c -b 16 -fail TEST*
Use the first when running in "8 bit" mode (vsib_record -m 3, which is
the default). Use the second version when running in "16 bit" mode
(vsib_record -m).
******
fauto
******
This autocorrelates the data and plots the passband. It works on all forms
of data (8/16 bit and 64 MHz).
Typical usage would be
>> fauto TEST000
Plot the bandpass of channel 1
>> fauto -c 2 TEST000
Plot the bandpass of channel 2
>> fauto -c 1 -c 3 TEST000
Plot the bandpass of channel 1 and 3
>> fauto -if1 1 -if2 4 TEST000
Plot the bandpass of channels 1 to 4
>> fauto -b 16 -if1 1 -if2 8 TEST000
Plot channels 1 to 8 for the Huygens data
Other useful options:
-sp1 First spectral point to plot
-sp2 Last spectral point to plot
-v/-vlba Calculate statistics for VLBA data
-w/-wide/-64 Assume 64 MHz bandwidth data
-sp1 <n> First spectral point to plot
-sp2 <n> Last spectral point to plot
-n/-nlags <n> # lags to correlate
-h/-help List brief help
*NOTE* Currently the only pgplot option is "/xs" (no hard copy). This
will be added soon.
You are forced to auto-correlate the entire recorded file.
I will changes this also soon.
***********
band_invert
***********
This program will invert the bandpass for one or all data channels in a
file (it creates a new file). If probably only works on normal 256 Mbps
recording modes. No guarentee it still works at all. It is a very simple
program so can be quickly modifed to support cases on a as needed basis.
Cheers
Chris
Received on 2004-11-12 12:05:18