[gmx-users] Scaling GROMACS

[Justin Lemkul]

On 6/23/17 10:04 AM, Sergio Manzetti wrote:
> Hi, I have the following system:
> 
> Running on 1 node with total 24 cores, 24 logical cores
> Hardware detected:
> CPU info:
> Vendor: AuthenticAMD
> Brand: AMD Opteron(tm) Processor 6344
> SIMD instructions most likely to fit this hardware: AVX_128_FMA
> SIMD instructions selected at GROMACS compile time: SSE2
> 
> Compiled SIMD instructions: SSE2, GROMACS could use AVX_128_FMA on this machine, which is better
> 

Make sure to re-compile properly to get good performance.

> Reading file topol.tpr, VERSION 5.1.2 (single precision)
> Using 12 MPI threads
> Using 1 OpenMP thread per tMPI thread
> 
> 
> When running 4 separate jobs, does gcom -1 or other options in mdrun help in subdividing the tasks?
> 

You can pin a job to specific cores to avoid resources fighting one another with 
-pin and -pinoffset.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================