Folding Forum

Does OpenMM entirely resolve the issue of having multi-threaded cores?

No, doesn't apply to threading really, though curious what issue you are referring to...

It's more related to creating a homogenous software interface to various and diverse hardwares. From the OpenMM page...

Description: OpenMM is a library which provides tools for modern molecular modeling simulation. As a library it can be hooked into any code, allowing that code to do molecular modeling with minimal extra coding.

Moreover, OpenMM has a strong emphasis on hardware acceleration, thus providing not just a consistent API, but much greater performance than what one could get from just about any other code available.

I am asking about the difference between a uniprocessor and smp library that OpenMM applies to. I'm wondering if OpenMM was designed to apply to both uniprocess/smp clients, or if any MD algorithms built for OpenMM are always going to be uniprocessor for now. not that I care if it supports SMP. there is always the ability to assign cpu affinity to multiple external processes instead of a multi-threaded appraoch.

Their page does indicate that one of the long term goals is...

- hardware support for multi-core, SSE, and/or graphics processing units (GPUs)

You'd have to ask them on the exact status of this goal, although their news page might be a good indicator...

Major Milestone for OpenMM: v1.0 Released January 21, 2010
Joy Ku
We are excited to announce the release of OpenMM 1.0. OpenMM is a freely available, extensible library that enables accelerated calculations for molecular dynamics on high-performance computer architectures, such as GPU cards. This release marks a major milestone. In addition to several new features, such as the ability to specify custom bond, angle, external, and Generalized Born forces, this release includes a stable API and reference and CUDA platforms that are suitable for production use. OpenMM and tutorial examples are available for download from http://simtk.org/home/openmm.