Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

By default nVidia settings sets the GPU as the physics processor. In cases where the CPU is powerful enough, is it better to set CPU as the physx processor?

In my case my i7 3770k is mostly unused (at least not all cores) during gaming, but my GTX 660 is not, as it tries its maximum to increase framerate.

In general, is it better to set CPU as physx processor when the CPU is powerful enough?

nvidia control panel

share|improve this question
considering the fact that you play games which require heavy graphics... the GPU is almost always better equipped to handle the load. The CPU works is ways other than the GPU, which won't allow you to see massive performance even with your powerful i7. so let the default settings be the GPU – Spandan Chatterjee May 5 '13 at 20:13
up vote 2 down vote accepted

Emulation physical processor PhysX on intel CPU's just worse.) Use nVidia PhysX physical processor.

High level PhysX Architecture

High level overview of AGEIA's PhysX Architecture

Are distinguished by different mathematics. For example a straight can be represented as the sum of multiple sine waves. But a few uncomfortable, and the costs are large. This was essentially different from the CPU GPU. It is better to use the right tool for its intended purpose and all. GPU geometric solves the problem. CPU solves the problem of integer and floating point. In the first approximation.

share|improve this answer

For games that use that Physx heavily, you want to use the GPU. Light physics work the CPU would do fine, but under a geavy physics (Or specifically Physx) the GPU is massively parallel which means significant improvement for this type of work (The same reason a decent GPU also smokes a top of the line CPU in things like @Home).

A Cpu is much more serial and does not handle these types of tasks nearly as well.

share|improve this answer

Take a look at this paper A Survey of CPU-GPU Heterogeneous Computing Techniques which discusses unique strength of both CPU and GPU. Published in ACM Computing Surveys 2015, it reviews nearly 200 papers and classifies them on several characteristics, e.g. based on their application domain (e.g. physics, image processing). This paper provides a good discussion on 'CPU vs GPU' or 'CPU-GPU collaborative computing'.

share|improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .