Ian Harry

Oh wait, it's called a second time here (??): https://github.com/gwastro/pycbc/blob/master/pycbc/filter/matchedfilter.py#L1810 Except here the delta_f is different (no idea why). I think this is a possible cause of this, as this...

After chatting with Tito, the issue is here: https://github.com/gwastro/pycbc/blob/master/pycbc/filter/matchedfilter.py#L1855 and arises because the code asks for additional look back time for [reasons I haven't understood]. However, as we are only...

So looking more closely at the code, there's a line here https://github.com/gwastro/pycbc/blob/master/pycbc/strain/strain.py#L1991 which *doubles* max_length. So when we ask for 256, we get 512. This would then mean that rather...

Apologies, I missed this! How would you know what the *original* distances etc. are if doing this? I'm also not sure it would let you *independently* jitter distance/phase/etc. in H1...

Many of these are still using more memory than they should/could ... Just not enough that it's impossible to run them

It should be possible to call `time_delay_from_earth_center` *once* and have it compute N_IFO * N_skypoints quantities. I think it could take vectorized input (might need some poking). If that doesn't...

@ahnitz A few points to answer (I won't bother trying to fix the tests until we're at least happy with the motivation): * The `--dax-file-directory` option should *never* be set...

@WuShichao Does this https://github.com/gwastro/pycbc/pull/4225 fix the issue?

The lines in the PSD tending to 0 definitely cause an issue when trying to whiten.

Something to discuss at the next call!