Low-latency and offline

Low-latency h(t) reconstruction server (Aug - Oct/31: DT,SH,TY,NK,...)

• We already have three computing nodes to test the offline h(t) reconstruction via frame file but we need to install another for the low-latency h(t) reconstruction via shared memory inside the tunnel

• (Aug/20-) Connection and software test between a prototype PC and a test bench
  • Prototype PC is just a standard desktop PC and will be prepared by SH
• (Aug-Sep) Procurement of server (2U height, at least two 10G network card, standard CPU,memory and disk)
  • According to Darkhan’s test with LIGO data and discussion with LIGO people, standard spec should be OK.
  • DT is working to figure out the spec and model.
  • Academia Sinica will make the procurement
• (Mid.Sep) Connection test between the server and the test bench
  • We also need the collaboration with DMG people for the data transfer
• (Sep-Oct) Connection test between the server and the KAGRA online system
  • We also need the collaboration with DMG people for the data transfer

Current tasks (May 26 - July 15):

• Ask to LIGO CAL team for the detailed information of gstlal-calibration (DT) [status: ongoing | expected finish: N/A]

• Make gstlal-calibration running on a machine at KAGRA (DT, SH) [status: ongoing | expected finish: Jul 15]

  • Find names and versions of all prerequisite libraries
  • Install prerequisites and gstlal-cal package on a machine at AS
  • Test gstlal-cal pipeline on an AS machine
  • Install prerequisites and gstlal-cal package on a machine at KAGRA
  • Test the pipeline on a KAGRA machine

Progress so far (June 21):

• Ask LIGO CAL team for the detailed information of gstlal-calibration

  • (ongoing) gstlal-calibration maintainers in LIGO are helping to resolve some installation and testing issues.

• Find names and versions of all prerequisite libraries,

  • (done, entry from June 7) SH made progress on this, see gstlal installation

• Install required packages and the gstlal-cal package on a machine at AS

  • DT has written scripts for installing lalsuite and gstlal software on a clean CentOS7 distro based on SH's instructions. Scripts were uploaded to KAGRA-AS group's Gitlab repositories:

    • containerized "cold" version of lalsuite

      containerized "cold" version of gstlal-calibration

  • CI pipelines that build docker-lalsuite-cold and docker-gstlal-cold containers were set up to run on AS IoP machine.

• Test gstlal-cal pipeline on an AS machine

  • As a prerequisite to this task, DT reproduced a portion of LIGO h(t) from raw data with gstlal-calibration-v1.1.8 (w/ a "docker-gstlal-cold" image). This test was done on a private server (2-cores on Intel Xeon 2.4GHz, 8GB RAM).

  • Downloaded a few hours of bKAGRA Phase-1 data on AS machine.

  • (doing) Will try to reproduce C00 with gstlal-calibration. Details in

    • gstlal-calibration testing on AS machine (cal-iop)

  • (to do) Verify the output of gstlal-calibration.

• Install prerequisites and gstlal-cal package on a machine at KAGRA

  • (done, entry from June 7) SH made progress on this, see gstlal installation.

• Test the pipeline on a KAGRA machine gstlal tests

  • to do

• Documentation

  • (doing) Started documenting the work in KAGRA-AS group's Gitlab repositories:

    • containerized "cold" version of lalsuite

      containerized "cold" version of gstlal-calibration

      gstlal-calibration testing on AS machine (cal-iop)

Progress so far (June 7):

• Ask LIGO CAL team for the detailed information of gstlal-calibration

  • Communication with people working on LIGO gstlal has been initiated.

• Find names and versions of all prerequisite libraries,

  • SH made progress on this, see gstlal installation

• Install prerequisites and gstlal-cal package on a machine at AS

  • As a prerequisite to this task, SH worked on installation process for CentOS, DT for Debian.

  • The goal is to have a better understanding of what is happening under the hood when the gstlal-calibration package is compiled.

• Test gstlal-cal pipeline on an AS machine

  • As a prerequisite to this task, DT is currently trying to reproduce LIGO h(t) from raw data.

  • A sub-goal of this task is to come up with a gstlal code testing procedure.

  • DT addressed some questions to LIGO gstlal people.

• Install prerequisites and gstlal-cal package on a machine at KAGRA

  • SH made progress on this, see gstlal installation

• Test the pipeline on a KAGRA machine gstlal tests

  • to do

New tasks:

• Put more effort on documentation.