The Visiometa Workflow - Part I: Performance

by Rene Romeike

We have been asked on social media about the utility of fast modulus computation, when conventional thermal analysis can also be done in a relatively short time today. So we decided to demonstrate the difference between "quite quickly" and "real fast" for you.

In the field of numerical simulation, there is a clear industry trend to leverage the high degree of parallelization that comes with modern GPU architectures. On occasion of the new GPU generation's recent release, we benchmarked our modulus analysis on different computer systems, ranging from older models to the latest and most powerful hardware.

The results speak for themselves: With current hardware it is easily possible to calculate the modulus field for the reference part shown below in just a few seconds, which is still at least two orders of magnitude faster than what regular thermal analysis can do today.

Comparison of Visiometa's computing performance on different machines. For the illustrated engine block, the solidification modulus was calculated in four discretizations between 4 million and 32 million nodes.

But why is this so important? Quite simply: because it is only with this kind of speed that it becomes possible to enable the "right first time" workflow that we are aiming for. With Visiometa, geometric part optimization, mold design and simulation are no longer separate work steps. Instead, they are partial aspects of an integrated planning approach that significantly reduces effort and avoids subsequent modification costs.

Want to know more? Then read also the second part of this feature series.

Go back