HyperX Solver is the latest evolution of our Sizing and Analysis engine, built with performance and customization in mind. As the aerospace industry shifts to increasingly large and complex models, analysis technologies have to keep pace. The technical challenges presented by these trends provide new opportunities to expand HyperX's capabilities and capacity.
Note
Built with High Performance Computing (HPC) and Customizability in mind, HyperX Solver is the latest evolution of our analysis engine.
Our biggest goal with HyperX Solver is to match/improve analysis accuracy while reducing the time it takes to analyze.
With that in mind, we expect HyperX Solver results to generally match the Standard analysis results for the same inputs (with some exceptions). To ensure that we do not sacrifice quality while continuing to develop Solver, there are some analysis methods that are not currently available via Solver. For up-to-date information on Solver's current capabilities and an in-depth comparison of Solver and Standard analysis methods, please request the complete Solver documentation.
Important
While we continue to roll out Solver methods please note that there may be some bugs or instabilities and results should be verified before they are used for production. We greatly appreciate feedback so please let us know if you experience any issues while using HyperX Solver.
From the foundation of HyperSizer, HyperX built up new capabilities, increased usability, and improved computational performance. Similarly, HyperX Solver builds on HyperX to bring you significantly increased performance that is more customizable and easier to use than ever before.
Note
HyperX Solver is not just a "drop-in" replacement of HyperX's computational sizing engine; it's a new framework built to solve new problems - even problems we haven't discovered yet.
HyperX Solver is built with performance and modularity in mind. Problems in aerospace are only getting bigger and more complicated. Solver is built to handle the biggest problems of today with the flexibility adapt to the problems of tomorrow.
Some of these capabilities already exist in HyperX but can be used in new ways in Solver. For instance, HyperX provides various methods to analyze the producibility of parts but those must be considered after Sizing. With Solver, you can consider producibility and Sizing simultaneously to arrive at better compromise solutions in less time. New Solver features, such as multi-threading, can solve problems involving symmetry and variable linking more easily and quickly by factoring in related Zones and/or variability rules while Sizing.
Note
Processes that were previously serial can now be run in parallel. You can also modify or add to them at more places than ever before.
Aerospace engineering continues to trend towards larger, more complex, and higher fidelity models. Models with 10k+ zones and 100k+ load cases are becoming commonplace. All this extra data must be processed, and it must be processed faster and more efficiently than before. HyperX Solver is designed to help you do just that.
The first thing to consider is the data that you're looking for. HyperX Solver affords much more output data granularity than our Standard methods. For instance, when Sizing, you may want the best candidate for every Zone...and the margins of safety...for all the criteria...but not just the top load case...let's say the top 100...and each element in the Zone, just in case...with Analysis Details of course. With HyperX Solver, we've introduced Output Databases that can store whatever data you specify.
Note
Output Databases are scalable to huge datasets and are designed to be optionally standalone. That means they can archive extensive results without needing to archive the even larger HyperX Database that generated them.
In addition to providing more control over what data you want, HyperX Solver lets you choose how you want to process it. HyperX Solver launches with Linux support and can be run remotely with on-premises clusters or commercial HPC solutions (i.e. AWS, Azure, etc.).
Note
HyperX Solver can be run wherever you want it to: from individual PCs to in-house Linux Servers to offsite HPC environments and everything in between.
To truly be flexible to your needs, HyperX Solver has to be easy to use. We are making it as painless as possible and providing more tools to manage processes inside of HyperX. For instance, we've added new forms to HyperX to set up, execute, and monitor remote jobs. These features help you focus more attention on the data instead of the means of acquiring it.
Combining these features with multi-threading and multiprocessing in cluster environments allows HyperX Solver to take advantage of as much hardware you are willing to throw at it. Empowering HyperX Solver with this level of HPC support means huge gains in performance on its own, but we didn't stop there. We've also leveraged our 30+ years of industry experience to improve our algorithms with domain-specific optimizations.
Decades of assisting customers and internally using HyperSizer and HyperX for aerospace stress analysis and design have provided countless lessons and tricks of the trade. HyperX Solver implements this hard-earned knowledge to utilize computing resources as efficiently as possible. Some examples include domain-specific handling of criteria and load cases and only calculating as-needed based on the outputs you request and the inputs you specify.
Note
HPC support and multi-threading/processing lets HyperX Solver work harder and 30+ years of aerospace stress analysis and design experience baked into our algorithms lets Solver work smarter.
In addition to performance increases, we are designing HyperX Solver to be as modular as possible. This will make it easier to customize Solver for your purposes today while providing a more adaptable framework for future problems. To accomplish this, we first asked: what aspects of our Sizing/Analysis process would someone want to customize?
With HyperX Solver, the processing steps above are all customizable modules. Some of these were already customizable, such as custom analysis criteria being Analysis Plugins, but we've extended that level of customization to all aspects of Sizing and Analysis. Now you can focus your customization more granularly instead of having to shoehorn all of your customization into one or two places.
Sometimes, added customization means added work but that's not the case here. We've made it easy to mix and match native and custom modules with Packages and the Package Manager. Packages can represent one or many of the customizable modules and the Package Manager is a one-stop-shop for discovering, managing, and distributing Packages.
Discoverability
We often wonder how to make sure users know about what extensions and customizations are available to them. We've heard similar questions from customers too. The Package Manager is designed to make available Packages easy to discover.
When you add Package Sources, all the Packages they contain are visible and installable from the Package Manager. When updates are available, they're displayed in the Package Manager and can be installed with a few clicks. You can also configured the Package Manager to automatically update Packages if desired.
Management
The Package Manager also offers comprehensive version control. You can see the version of Packages and the Modules they contain while also being able to switch between available versions easily.
Changelogs, readmes, and in-depth methods documentation are included with every Package. These can all be accessed directly in the Package Manager. Since these are associated with each version of the Package, you'll know the documentation you're reading matches the methods you're using.
Distribution
The Package Manager can be linked with Sources, folders containing Packages, making all the contained Packages visible and installable. From a program management standpoint, this makes it easy to deploy and standardize Packages amongst your team.
You can provide a Package Source that’s a company-internal server, and your methods group can publish new Packages to that server whenever they have them. End users will automatically see them as available since they configured HyperX to use your internal server source. This also makes deploying bugfixes and updates easy. Automatic upgrade of Packages is possible, and the software will show you when an update is available for something you’re using if you want to manually manage updates.
As we continue to develop HyperX Solver, we're tracking the performance on a number of case studies (including the examples below). This allows us to track our progress as we improve performance and correct any accidental slowdowns.
Consider the following case studies as examples of the performance increases provided by HyperX Solver.
Sizing Composite Wing Cover
This case represents sizing a wing cover of a generic commercial aircraft.
Setup Details:
-
216 Panel Zones
-
Smeared sizing with 41,310 candidates per Zone
-
Peak load processing (~35 design-to cases)
-
Panel Buckling, Local Buckling, Composite Strength, Composite Crippling
-
20 threads on a desktop PC
Time with Standard Methods: 43 minutes.
Time with HyperX Solver: 31 seconds. An 81x speedup
Sizing Composite UAM Wing with HyperXpert
This case represents sizing a UAM vehicle wing with HyperXpert.
Setup Details:
-
54 Panel Zones
-
Smeared sizing with 2856 candidates per Zone
-
Returning all viable candidates (HyperXpert mode)
-
Panel Buckling, Local Buckling, Composite Strength, Composite Crippling
-
20 threads on a desktop PC
Time with Standard Methods: 36 minutes.
Time with HyperX Solver: 1 minute 6 seconds. A 33x speedup
The Solver API can be used to develop custom panel and beam analysis criteria. This is similar to custom Analysis Plugins but with additional Solver-specific features. You can find a full Solver API Reference here:
Version 2025.1
HyperX Solver is available as an initial release in HyperX version 2025.1.4 as we continue to fill out its features. To use Solver, you must enable "Experimental Features" in Viewport Settings which can be accessed from the View tab of the Ribbon.
Enabling these features will add a new Solver section to the Run Tab of the Ribbon as well as enable new options for HyperFEA, Batch Runner, and HyperXpert.
Version 2026.1
Solver is no longer in the Experimental Features options. Instead, it has been brought to the forefront of the UI as an option on the Project Run Settings form, found on the Run tab of the Ribbon, HyperFEA, Batch Runner, HyperXpert, and Project Runner forms.
There are several locations throughout the software where Solver is mentioned. These features can only be run with an additional Solver license. See the note below to request a license.
Important
In order to run Solver in v2026.1 or later, an additional HyperX license type is required. Contact us to make a request.
Tip
The new HyperX Solver has not replaced the existing Standard Analysis Engine in any version. Rather, it must be enabled separately, and the user can return to running the Standard analysis code at any time.
While in this initial release window, feedback on Solver is especially welcome! We'd love to hear what works well for you and what doesn't. We also want to hear what you'd like to see Solver handle next so that we can prioritize Solver development to better suit your needs. Please reach out with any and all feedback/requests!