ARTEMIS: a versatile backend for transient searches

ARTEMIS: a versatile backend for transient searches

Aris Karastergiou
LOFAR Transients Key Project Meeting, Meudon, December 2011

Ab44292d7d6f032baf342a98230a6654?s=128

transientskp

June 23, 2012
Tweet

Transcript

  1. The Sun Pulsars Radio Transients Interstellar Medium Nanosecond Astronomy Ionosphere

    Cosmic Rays The Unknown LOFAR station Beamformed Data Thursday, 15 December 2011
  2. LOFAR international stations • Large collecting area (96 coherently added

    dipoles), high sensitivity. 9 international stations = superterp • Very fast sampling rate (5 ns), reduced to 5.12 µs for the station beams. • Large fields of view, 10 to 100s of square degrees. • Not used for all LOFAR KSPs. Thursday, 15 December 2011
  3. Dispersion searches DM 0 N2 or N logN process per

    time sample (N > 1000 -- 10000+ samples/sec HPC computing Thursday, 15 December 2011
  4. Dispersion searches DM 0 N2 or N logN process per

    time sample (N > 1000 -- 10000+ samples/sec HPC computing Thursday, 15 December 2011
  5. Dispersion searches DM 0 N2 or N logN process per

    time sample (N > 1000 -- 10000+ samples/sec HPC computing Thursday, 15 December 2011
  6. Dispersion searches DM 0 N2 or N logN process per

    time sample (N > 1000 -- 10000+ samples/sec HPC computing Thursday, 15 December 2011
  7. 3.2Gbps data reduced data 800Mbps data 800Mbps data 800Mbps data

    800Mbps data Thursday, 15 December 2011
  8. Dell Poweredge C6100 Dell Poweredge C410x Stef Salvini Wes Armour

    Dalal Ait Allal Mike Giles Chris Williams Steve Roberts Fred Dulwich Ben Mort Alessio Magro Aris Karastergiou hardware Thursday, 15 December 2011
  9. time dispersion measure First light -> but too many detections!

    Thursday, 15 December 2011
  10. Thursday, 15 December 2011

  11. Thursday, 15 December 2011

  12. Radio interference is localised Thursday, 15 December 2011

  13. Thursday, 15 December 2011

  14. Terrestrial RFI pulsar dispersed transient? Thursday, 15 December 2011

  15. frequency time Bright, short duration, broadband and frequency swept signals

    from... Thursday, 15 December 2011
  16. Thursday, 15 December 2011

  17. Frequency (MHz) Time (ms) 8000 8200 8400 8600 8800 9000

    143 144 145 146 147 148 149 20 30 40 50 60 200 400 600 800 0 2000 4000 6000 8000 Time bin 40 60 100 200 300 400 500 600 700 800 900 dB Frequency channels Pulse Thursday, 15 December 2011
  18. Frequency (MHz) Time (ms) 8000 8200 8400 8600 8800 9000

    143 144 145 146 147 148 149 16 18 20 22 24 26 28 30 200 400 600 800 0 200 400 Time bin 22 24 26 100 200 300 400 500 600 700 800 900 dB Frequency channels Thursday, 15 December 2011
  19. Dispersion measure pc/cm3 Time (s) Crab H AM SAT HAM

    SAT • Working prototype for non-image processing • Real-time searches for Individual Radio Pulses using CPU/GPU • Dedispersion over 4000 DMs per beam in real time with GPUs • Pilot survey 1: 6-8 beams tracking circumpolar targets • Pilot survey 2: 6 beams fixed on the meridian from 8o to 28o dec Pipeline diagnostic plot from Pilot survey – daily summary from one of six beams Receive udp stream Polyphase filter Raw to Stokes RFI removal Dedispersion output S/N DM Time (s) Aris Karastergiou (aris@astro.ox.ac.uk) Crab Thursday, 15 December 2011
  20. S/N DM ?? ? Detecting known sources of dispersed pulses

    including pulsars and frequency swept radio emitters (HAMSAT). Anticoincidence experiments with Effelsberg + Nancay will reveal nature of enigmatic detections • Performed on 800 mbps streams • Each node has: 12 Xeon CPU cores + 1 NVIDIA M2050 (or GTX) • DM processing is ~10x better than real time • RFI rejection performing well in very contaminated environment Dispersion measure pc/cm3 Time (s) Time (s) Aris Karastergiou (aris@astro.ox.ac.uk) Thursday, 15 December 2011
  21. Day 1 Thursday, 15 December 2011

  22. Day 2 Thursday, 15 December 2011

  23. Day 3 Thursday, 15 December 2011

  24. frequency intensity frequency intensity intensity frequency 1. Model the bandpass

    2. Match model to data 3. Subtract model and filter 4. Store spectrum stats in ring buffer 5. Update model with running averages Thursday, 15 December 2011
  25. dynamic noise modelling for effective RFI removal and transient detection

    DM time (s) GPU dedispersion: 4000 dispersion measures searched, only pulses from B0329+54 detected, no spurious signals filterbank noise pre-filtering filterbank noise post-filtering Thursday, 15 December 2011
  26. Thursday, 15 December 2011

  27. CPUs, GPUs... Wes Armour GPU de-dispersion : http://www.oerc.ox.ac.uk/research/wes ARTEMIS: http://www.oerc.ox.ac.uk/research/artemis

    Thursday, 15 December 2011
  28. Artemis 2.0 3.2Gbps port mirror 10gbe to 4x1gbe •9U including

    storage •parallel operations with ASTRON •cheaper and upgradable Thursday, 15 December 2011