search

radical-cybertools / radical.repex.at

This is the github location for RepEx developed by the RADICAL team in conjunction with the York Lab.
Other
4 stars 3 forks source link

problem with generalized branch #76

Closed haoyuanchen closed 8 years ago

haoyuanchen commented 8 years ago

I tried to run a 2D UU simulation with the feature/perfopt_gen branch, but I got this error:

2016-03-25 17:43:54,291: radical.saga.pty : MainProcess : MainThread : ERROR : DoesNotExist: file copy failed: mput "/home/haoyuan/radical.repex/examples/amber/tuu_remd_inputs/ace_ala_nme_coors_8x8/salt_conc_pre_exec.py" "/work/02938/chen1990/radical.pilot.sandbox/rp.session.prot-55-247.rutgers.edu.haoyuan.016885.0000-pilot.0000/staging_area/ace_ala_nme.inpcrd.0.0"

It's trying to copy a script for salt concentration exchange, which is not needed here. Also, it seems that it's trying to copy that file to a coordinate file...

antonst commented 8 years ago

Can you please provide your simulation input file?

haoyuanchen commented 8 years ago

{ "remd.input": { "re_pattern": "S", "number_of_cycles": "4", "input_folder": "tuu_remd_inputs", "input_file_basename": "ace_ala_nme_remd", "amber_input": "ace_ala_nme.mdin", "us_template": "ace_ala_nme_us.RST", "amber_parameters": "ace_ala_nme.parm7", "amber_coordinates_folder": "ace_ala_nme_coors_8x8", "same_coordinates": "False", "group_exec": "False", "init_temp": "300.0", "replica_mpi": "True", "replica_cores": "2", "steps_per_cycle": "5000", "download_mdinfo": "False", "download_mdout" : "False" }, "dim.input": { "d1": { "type" : "umbrella", "number_of_replicas": "4", "min_us_param": "45.0", "max_us_param": "360.0" }, "d2": { "type" : "umbrella", "number_of_replicas": "4", "min_us_param": "45.0", "max_us_param": "360.0" } } }

antonst commented 8 years ago

Thanks, can you please try to re-run now?

haoyuanchen commented 8 years ago

It ran for several cycles and then crashed. However, I looked at each replica, none of them actually finished their MD cycles--Amber wasn't even executed in the MD cycle. But it still went through several cycles and no exchange occurred.

antonst commented 8 years ago

This is fixed now in feature/perfopt_gen branch. Also I recommend to set number of cores = node size.