Table of information on head-on collision runs

Equal mass case (m1=m2=0.5, M_tot=1.0)

/th>
Init.Seperatio(C,P) Resol.(h) Crash time punc split (Y/N) Excision Gauge Movies Comments
C=2.3M, P~4.5M M/16,M/24 didn't Y . 1+log,HG(eta=3.1) . M/32 paper run didn't crash
C=2.3M, P~4.5M M/12 No crash Yes None 1+log,HG(eta=3.1) [Lapse: MPEG ] [Shiftz: MPEG ] [Shiftz_zoom: MPEG ] [conPhi: MPEG ] [HC: MPEG ] [HC_zoom: MPEG ] [RePsi4: MPEG ] [RePsi4_zoom: MPEG ] Waveform [PS]
C=2.3M, P~4.5M M/12 No crash No None 1+log,HG(eta=3.1) See above line See above line
C=2.3M, P~4.5M M/16 running N None 1+log,HG(eta=3.1) . waveform to come!
C=4M, P~6.8M M/16 56M Y mask too small 1+log,HG(eta=3.0) . See line below. Interesting? huh
C=4M, P~6.8M M/16 ran up to 131M Y None 1+log,HG(eta=3.0) . Waveforms[PS]
C=4M, P~6.8M M/12 didn't N None 1+log,HG(eta=3.0) Comparison bet. split/no-split [alpha][betaz] Waveform doesn't look good. Not terrible but should be improved...
C=4M, P~6.8M M/16 didn't N None 1+log,HG(eta=3.0) . Look reasonable
C=4M, P~6.8M M/24 37M N None 1+log,HG(eta=3.0) How I crashed [alpha:mpeg] [betaz:mpeg] .
C=4M, P~6.8M M/24 RUNNING NOW N None 1+log,HG(eta=3.0), DISSIPATION How I survived the crash with dissipation [alpha:mpeg] [betaz:mpeg] DISSIP helps
C=6M, P~9M M/12 63M N None 1+log,HG(eta=3.1) . Things go bad around origin, waveform TERRIBLE!
C=6M, P~9M M/12 RUNNING NOW N None 1+log,HG(eta=3.1) DISSIPATION . waveform terrible; it seems it needs more resolution will see
C=6M, P~9M M/16 31M Y mask too small 1+log,HG(eta=3.1) . MAD gives the same crash time
C=6M, P~9M M/16 66M Y mask too small 1+log,HG(eta=11.2) . Pablo P=10M
C=6M, P~9M M/16 didn't crash;ran to 81M! N None 1+log,HG(eta=3.1) . waveforms looks reasonable indeed!




  • General observation
    • Aware that there maybe *minimum* resolution requirement for a given separation for convergence and non-noisy waveforms. (e.g. M/16 for up to L_prop = 9M, M/12 may be ok for L_prop=4M or so)
    • In non-punc split strategy, High resolution problems around initial puncture locations can be dealt with dissipation.
    • (To be clarified further with more data and graphsics, but) Waveforms seem to indicate
      • that initial part of waveforms, often look noisy, are *really* initial data effect (so I am now less worried with sharp waveforms for this early part of waveforms..) and
      • therefore can be separated from the *collision* waveforms by farther separation. Given resources this can be pushed in full speed.
  • Next tsteps
    • Use at least h_finest=M/16, do parameter scanning covering the range of L_prop = 4.5M (QC0-equiv.) thru up to say L_prop=10--15M (going beyond Pablo).
    • clearly demonstarte separatoin of ID effect from the real collision waveforms...






  • Non-Equal mass case (m1=0.25,m2=0.75)

    Initial seperation (C,P) Resolution (h) Crash time punc split (Y/N) Excision Gauge Movies Comments
    C=2.3M M/12 30M Y meaningless 1+log,Gamma Drvier . .
    C=2.3M M/12 50M N meaningless 1+log,Gamma Drvier . .
    C=2.3M M/16 14M Y meaingless 1+log,Gamma Drvier . .
    C=2.3M M/16 28M Y with elongated mask turned on later 1+log,Gamma Drvier . .
    C=2.3M M/16 past 100M N meaningless 1+log,Gamma Drvier . waveform didn't look good. [PS]