Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Crash after running over an entire brick #1

Open
ngbusca opened this issue Jun 16, 2016 · 5 comments
Open

Crash after running over an entire brick #1

ngbusca opened this issue Jun 16, 2016 · 5 comments

Comments

@ngbusca
Copy link

ngbusca commented Jun 16, 2016

After running without multiprocessing over brick brick-b-4200p000.fits brick-r-4200p000.fits of the BOSS quasar training sample, I get a crash at the very end, after seemingly successfully fitting over all spectra.

I can't reproduce the bug if I run only over a few spectra in this brick.

Traceback (most recent call last):
File "/Users/ngbusca/Boulot/DESI/github/desispec/bin/desi_zfind", line 8, in
zfind.main(args)
File "/Users/ngbusca/Boulot/DESI/github/desispec/py/desispec/scripts/zfind.py", line 199, in main
zff = RedMonsterZfind(**arg)
File "/Users/ngbusca/Boulot/DESI/github/desispec/py/desispec/zfind/redmonster.py", line 113, in init
self.zpicker = ZPicker(specobj, self.zfinders, self.zfitters, flags)
File "/Volumes/LyA4ToB/ngbusca/DESI/redmonster/python/redmonster/physics/zpicker2.py", line 65, in init
self.classify_obj(zfindobjs, zfitobjs, flags)
File "/Volumes/LyA4ToB/ngbusca/DESI/redmonster/python/redmonster/physics/zpicker2.py", line 140, in classify_obj
n.max(zfitobjs[itemp].chi2vals[ifiber]) /
File "/Users/ngbusca/Boulot/homebrew/lib/python2.7/site-packages/numpy/core/fromnumeric.py", line 2258, in amax
out=out, keepdims=keepdims)
File "/Users/ngbusca/Boulot/homebrew/lib/python2.7/site-packages/numpy/core/_methods.py", line 26, in _amax
return umr_maximum(a, axis, None, out, keepdims)
ValueError: zero-size array to reduction operation maximum which has no identity

@julienguy
Copy link
Owner

ok, checking that ...

@julienguy
Copy link
Owner

I have the same error ... I am going to rerun with temporary much shorter template arrays to go faster

@ngbusca
Copy link
Author

ngbusca commented Jun 16, 2016

@julienguy Can we predict which spectra cause trouble and skip them in the wrapper?

@julienguy
Copy link
Owner

This is now fixed in the head version.
There was a bug when no solution was found for one template class (for instance, no valid solution for stars but solutions for QSO and galaxies). I checked that the best solution is found at the end in those cases.

@julienguy
Copy link
Owner

I processed brick 4200p000 in 38 minutes on 20 cores with the original redmonster templates.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants