[MINC-users] Installation issues - make check fails

Claude LEPAGE claude at bic.mni.mcgill.ca
Wed Sep 16 10:26:45 EDT 2009


Hi,

> I am having trouble with one of the binaries in the N3 software. make check fails on test 8 and passes successfully on the other 7 tests.
>
> I have installed:
> N3-1.11.0
> EBTKS-1.6.2
> MINC-2.0.12
> NetCDF-4.0.1
> HDF5-1.8.3
>
> Here is the relevant portion of test.log :
>
> [resample_labels] [root at cortex:/usr/local/N3-1.11.0/testing] [2009-09-16 09:53:24] /usr/local/mni/bin/extracttag /usr/tmp/resample_labels_10573/icbm_avg_152_t1_tal_nlin_symmetric_VI_mask.mnc_rsl.mnc -volume /usr/tmp/nu_estimate_np_and_em_10558/icbm_avg_152_t1_tal_nlin_symmetric_VI_mask.mnc.temp -threshold 0.5 1.1 -label 1 -maxtags 0
> Updating history
> (from miopen): Can't write compressed file
> ncredef: ncid -1: NetCDF: Not a valid ID
> ncattput: ncid -1: NetCDF: Not a valid ID
> ncclose: ncid -1: NetCDF: Not a valid ID
> resample_labels: crashed while running minc_modify_header (termination status=768)
> nu_estimate_np_and_em: crashed while running resample_labels (termination status=768)

I don't know if this will solve the problem, but netcdf-4 has not been
tested with minc. The API changed from version 3 to 4, so there might be
some incompatibilities. I suggest you link with netcdf 3.6.1 (which I 
know works). Similarly, we use hdf5 1.6.6. I haven't tested 1.8.X. 

And while you are at it, why not use minc-2.0.18? 2.0.12 is old and many
bugs have been fixed since then. Moreover, 2.0.18 offers much better/faster
internal compression. Make sure to activate minc2 compression using the
shell environment variables MINC_FORCE_V2=1 and MINC_COMPRESS=4.

Let me know if this works.

Claude


More information about the MINC-users mailing list