#########################
--GPU Comp Chem--
#########################
Required Hardware:
NVIDIA card with compute Compute Capability 2.0
Tested on: NVIDIA Tesla M2070
Required Software:
Linux Operating System
Tested on: SUSE Linux Enterprise Server 11 SP2
NVIDIA Developer Toolkit
Download from http://developer.nvidia.com/cuda-downloads based on your
OS and follow the instructions given to install.
Tested with: CUDA 4.2
If you are using the Alabama Supercomputer Center (ASC) configure based on the
instructions received when you set up your account.
When you have both of the above installed you should be ready to build.
Build Instructions:
>> git clone git://github.com/orlandoacevedo/MCGPU.git
>> cd MCGPU/
>> make
To Run a Simulation on a local machine:
>> cd /path/to/MCGPU/
>> cd bin/
>> ./linearSim demoConfiguration.txt
To Run a Simulation on the Alabama Supercomputer Center:
>> cd /path/to/MCGPU/
>> run_gpu demo_script.txt
Choose a batch job queue:
Queue CPU Mem # CPUs
-------------- ---------- ------ ------
small-serial 40:00:00 4gb 1
medium-serial 90:00:00 16gb 1
large-serial 240:00:00 120gb 1
class 2:00:00 64gb 1-64
daytime 4:00:00 16gb 1-4
express 01:00:00 500mb 1
Enter Queue Name (default <cr>: small-serial) <must be a serial queue>
Enter Time Limit (default <cr>: 40:00:00 HH:MM:SS) <enter time limit>
Enter memory limit (default <cr>: 500mb ) <enter required memory>
Enter GPU architecture [t10/fermi/any] (default <cr>: any) <fermi>
You standard out for your job will be written to
<jobname>.o<job number>
For more information see the Alabama Supercomputer Center manual.
Configuration File:
Configuration files are used to configure a simulation.
Line numbers are important
[1] #line 1 is a comment and will be ignored
[2] <x dimension>
[3] <y dimension>
[4] <z dimension>
[5] #line 5 is a comment and will be ignored
[6] <temperature in kelvin>
[7] #line seven is a comment and will be ignored
[8] <max translation for a molecule in angstroms>
[9] #line 9 is a comment and is ignored
[10] <number of steps for which to run the simulation>
[11] #line 11 is a commend and is ignored
[12] <number of molecules in the simulation>
[13] #line 13 is a comment and is ignored
[14] <path to the opls-aa.par file>
[15] #line 15 is a comment and is ignored
[16] <path to the z-matrix file to use>
[17] #line 17 is a comment and is ignored.
[18] <path to the state input file; overrides z-matrix setting is present>
[19] #line 19 is a comment and is ignored.
[20] <path to the stateoutput file>
[21] #line 21 is a comment and is ignored.
[22] <path to the pdb output file>
[23] #line 23 is a comment and is ignored.
[24] <nonbonded cutoff distance in angstroms>
[25] #line 25 is a comment and is ignored.
[26] <max rotation for a molecule in degrees>
[27] #line 27 is a comment and is ignored.
[28] <random number seed input as integer value>
[29] #line 29 is a comment and is ignored.
[30] <primary atom index to be used during cutoff as integer index of z-matrix atom in molecule>
Contributing Authors: Matt Campbell, Matthew Hardwick, Alexander Luchs, Riley Spahn,
Kalan Stowe, Ashley Tolbert, Seth Wooten, Xiao (David) Zhang, and Orlando Acevedo