search

lacker / seticore

A high-performance implementation of some core SETI algorithms that can be included in other programs.
MIT License
3 stars 6 forks source link

Removing a DC spike should be optional (main.cu) #1

Closed texadactyl closed 2 years ago

texadactyl commented 2 years ago

Some telescope arrays, have no DC spike. According to David McMahon, the spike comes about because of a shortcut taken in the FPGA design of the coarse channelizer (aka "F engine").

So, this should be a flag-driven feature.

lacker commented 2 years ago

That makes sense. Maybe it should even just make the correct decision based on the h5 file - one of the headers has the information about which telescope it is, right? Then we wouldn't have to ask people to get a bunch of flags right all the time.

texadactyl commented 2 years ago

DC spikes might be just Parkes and GBT. LOFAR? Need verification on all of them.

lacker commented 2 years ago

Wael says the ATA does not have a DC spike

texadactyl commented 2 years ago

Makes sense. Danny said it was mostly a characteristic of single-antenna telescopes. Base it on telescope ID?

lacker commented 2 years ago

I'll make it part of the telescope-guessing logic, since I still see plenty of data with telescope id -1...

lacker commented 2 years ago

Meerkat data also has no DC spike. Mentioning here for later reference

lacker commented 2 years ago

DC spike is now part of the telescope-guessing logic:

https://github.com/lacker/seticore/commit/987cc9c0f452200811a36aa5d1ec1a558a266aa1