biochem-fan
commented
3 years ago Refine3D needs two half sets for FSC calculation. Thus you MUST use MPI.
Closed yuanhaowang1213 closed 3 years ago
biochem-fan
commented
3 years ago Refine3D needs two half sets for FSC calculation. Thus you MUST use MPI.
After doing ctf correction and particle extraction, I tried with relion_refine (not relion_refine_mpi:
relion_refine --o Refine3D/job005/run --auto_refine --split_random_halves --i particles_subtomo.star --ref riboReScale200.mrc --firstiter_cc --ini_high 60 --dont_combine_weights_via_disc --pool 3 --ctf --particle_diameter 200 --flatten_solvent --zero_mask --oversampling 1 --healpix_order 2 --auto_local_healpix_order 4 --offset_range 5 --offset_step 2 --sym C1 --low_resol_join_halves 40 --norm --scale --j 64
, but got this error:
in: /home/****/Downloads/software/relion/src/ml_optimiser.cpp, line 2417 Backtrace
relion_refine(_ZN11RelionErrorC2ERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEES7_l+0x77) [0x562c0c446757] relion_refine(_ZN11MlOptimiser7iterateEv+0x6cb) [0x562c0c4a318b] relion_refine(main+0x51) [0x562c0c434de1] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7) [0x150e0ddffbf7] relion_refine(_start+0x2a) [0x562c0c43740a]
ERROR: ERROR: Cannot split data into random halves without using MPI!
So my question is that whether I could remove --split_random_halves, will this cause artifects or? I only have a workstation with two xeon processor, so probably I cannot use mpi here or the mpi board number is 1 for me.
Environment:
Dataset: EMPIAR 10045