Page History
...
The default setting will analyze run 124 on psanaq with 24 CPUs. Number of events to process set to -1 analyzes all the events.
For this demo, let's analyze only 50 events with 12 CPUs.
- Run(s): 124
- Queue: psanaq
- CPUs: 12
- Number of events to process: 50
Click "Launch peak finder”. To check your job progress in the queue, type the following in the terminal with your own username:
| Code Block | ||
|---|---|---|
| ||
$ squeue -u <username> |
At the time of writing this documentation, psanaq is quite busy processing 7466 jobs. 6612 jobs are pending. 854 jobs are currently running. For more information on which queue you are allowed to use, see see Batch Nodes And Queues#BatchNodes
Click "Launch peak finder”. You can check the status of launched jobs by typing "bjobs" in your Terminal. This will display the job ID and the status amongst other information.A logfile of the peak finding is also saved under the same directory, . <jobID>.log. Note the "." in front of the logfile name.
If the status stays in PEND PD or SUSP mode for awhile, then you may want to kill the jobs. To kill a batch job, type "bkill scancel <jobID>". For more information, please refer to: Submitting Batch Jobs
You can check the status of your peak finding job here: /reg/d/psdm/cxi/cxic00318/scratch/<username>/psocake/r0124/status_peaks.txt
| Code Block | ||
|---|---|---|
| ||
$ cat /reg/d/psdm/cxi/ |
cxic00318/scratch/<username>/psocake/ |
r0124/status_ |
peaks.txt
{"numHits": 19, "hitRate(%)": 38.0, "fracDone(%)": 100.0} |
We found 19 hits from 50 events. The hit rate is 38%. Fraction of events processed is 100%.
Also, click "Refresh" in Small Data panel to see the progress. You should see a red plot showing event number vs number of peaks found. When peak finding finishes, the plot will turn blue.
...
Two virtual powder patterns will also be generated in the run directory:
1) cxitut13cxic00318_00100124_maxHits.npy: Maximum pixel values for hits found
2) cxitut13cxic00318_00100124_maxMisses.npy: Maximum pixel Maximum pixel values for misses
If the hit finding parameters are good, cxitut13cxic00318_00100124_maxMisses.npy should not contain many Bragg spots. You can look at these images by using the "Load image" button in the Image Control panel.At CXI, the detector geometry may not be centred exactly. You can fix the centre by clicking "Deploy automatically centred geometry" in the Diffraction geometry panel. This uses cxitut13_0010_maxHits.npy and cxitut13_0010_maxMisses.npy.
Jumping to interesting images based on the number of peaks
Once you have submitted the peak finder job, let's plot the number of peaks found for each event by clicking the "Refresh" button. The refresh button will sometimes not work if your .cxi file is busy writing.
In the small data panel, you should see the CXIDB filename:
- filename: /reg/d/psdm/cxi/cxitut13cxic00318/scratch/<username>/psocake/r0010r0124/cxitut13cxic00318_00100124.cxi
- dataset: /entry_1/result_1/nPeaksAll
...
Overview
Content Tools