NAMD with GPU error : "Number of VDW types exceeds CUDA LJ_TABLE_SIZE"

From: David McGiven (davidmcgivenn_at_gmail.com)
Date: Fri May 27 2011 - 06:12:42 CDT

Dear NAMD Users,

When trying to run NAMD, I get the following error :

FATAL ERROR: Number of VDW types exceeds CUDA LJ_TABLE_SIZE
FATAL ERROR: Number of VDW types exceeds CUDA LJ_TABLE_SIZE
[2] Stack Traceback:
   [2:0] CmiAbort+0x7b [0xadcae5]
   [2:1] _Z8NAMD_diePKc+0x62 [0x5360a2]
   [2:2] _Z22build_cuda_force_tablev+0x28 [0x6c8a84]
   [2:3] _ZN20ComputeNonbondedUtil6selectEv+0x30c7 [0x5debbb]
   [2:4] _ZN10ComputeMgr14createComputesEP10ComputeMap+0x4d [0x5c4b9d]
   [2:5] _ZN4Node7startupEv+0x2d0 [0x8ef29e]
   [2:6] _ZN12CkIndex_Node18_call_startup_voidEPvP4Node+0x12 [0x8eefca]
[...] I've reduced the amount of output. Basically it repeats the
FATAL ERROR and the Stack tracebacks.

The input has the following description :

# Continuing a job from restarting files
# Min of GPCR, raise temperature to 300K and equil without constraints
# embedded in POPC membrane, ions and water.
# PME, Constant Pressure.

I can attach the full log and the full input if needed.

Does anybody know what could be the problem ? I haven't found anyone
having this problem before.

Thanks.

Best Regards,
D.

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 05:24:00 CST