From clusters, to cloud, to desktop workstations, to GPU. Now there are many options for getting your computing done quickly.
Recently I had the pleasure of sitting down with Desktop engineering and discussing the opportunities in improving computing speed for engineers.
You can catch the full interview here
https://www.digitalengineering247.com/summit23/computing
Here is a summary in case you want to get the gist of what I talked about.
The most important factor in computing speed is defining the right problem to solve.
We typically work with our new customer on understanding their simulation needs and defining the right problem for them to tackle.
Defining the right problem allows for the use of different methodologies, software and computing techniques. This makes an enormous impact on the speed and efficiency of engineering computing.
GPU computing is a true game changer in engineering computing.
Problems that used to take days or require supercomputers can now be solved on a desktop workstation.
Here is an example
The power available on a desktop workstation today is the same as that of a cluster from past years back. This allows for an amazing increase in productivity. Many tools now supports native GPU computing.
I'm certain more and more will have this capability in the near future.
The big limitation to GPU computing is the amount of memory available. If problems can't be solved on a workstation the logical is to try to solve on the cloud.
Cloud has become easy and much more accessible.
There are a variety of options for utilizing the cloud. All signs points to improved ease of use and setup, lower cost and more access to software.
I go into more detail in the presentation, but hopefully this was an informative summary.