Experimental Physics and
| |||||||||||||||||
|
[...] Iâd like to do some array calculations (FFT, rectangular to polar, filtering etc). Iâd like the option to run at high rates (KHz) which would suggest using asyn. Is there any standard for this? I could easily put the calculations in the asyn layer but I canât see an obvious way to build a processing chain with only one link. Any suggestions? [...]My 2 cent: my "standard approach" to implement complex algorithms on IOCs is, to use many records connecting to the same device support. You can use one waveform record to feed the raw data to the device support and another to contain the FFT results. You may need to feed additional parameters of your FFT into the device support by output records or read out diagnostics into input records. I personally use SNL only for state machines, for algorithms I prefer a simple device support. This approach does not produce any overhead, since you are basically passing pointers. I'm always reluctant to pass MB of data with kHz over channel access (but maybe I'm old fashioned here ;-) Also is there a standard for inverse subarray? Iâd like to combine 100s of scalar PVs into one waveform. [...]I agree that an "inverse subarray" and a "waveout" record would be nice in EPICS base. Of course one can do both with a device support, but I would find a record implementation cool. Has somebody developed any of both records? If yes: how about adding it to base? Andreas
| ||||||||||||||||
ANJ, 02 Sep 2010 |
·
Home
·
News
·
About
·
Base
·
Modules
·
Extensions
·
Distributions
·
Download
·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing · |