Member Avatar for Catweazle
Catweazle 140 Grandad

Has the next major upheaval in PC hardware reared it’s head?

Every once in a long while, a new type of component arrives which dramatically changes how we consider our computers. The development of components such as video capture cards and TV tuners have added functionality, but that functionality is very much task oriented. Such add-ins don’t really change the way we work with out PCs – they simply add an extra function for those who need it.

The most recent really major change in PC hardware came with the introduction of 3D acceleration. Manufacturers of the first 3D acceleration cards had a difficult task in front of them to convince people that their product was needed, and for quite some time afterwards 3D cards were almost solely used for gaming. But over time other uses have been accepted, and nowadays for any desktop computer other than those used for the most basic of tasks it would be unthinkable not to have 3D display capability included. We are now confronting the likelihood of everyday computing where everything on the system becomes a 3D object.


Several months back now we started to hear more and more mention of the development of what just may prove to be the next ‘biggie’ in hardware development. A company called AGEIA has been readying what is the world’s first dedicated physics processing card. Very soon now, they will be available in retail.

Dedicated physics processing cards take the task of collision detection and the calculation of such things as fragmentation and motion of fluids away from the system’s central processor, and provide a previously unparalleled capability to the desktop machine. They do so by conceptually constructing solid and liquid objects from rigid particles known a ‘bones’. Software which utilizes the PhysX Software Development Kit can detect collisions between those particles and calculate their effects. As currently under development, the Physics processing card can handle 40,000 ‘bones’ onscreen at any one time, or even more when representing liquids. Even a modern dual-core processor can only handle a few hundred!


Initially, the cards will have impact in the gaming world, just as many other hardware innovations do. It isn’t hard to conceive, after all, what impact it will have on the capacity to create a realistic game world environment. When you consider those words “realistic environment however, you get a hint at what further capabilities are also introduced. Initially there will be plug-ins for the major CAD/CAM software packages. Later, no doubt, there will be software uses which only the imagination can place limits on.

I’m sitting here dreaming about waving my PC mouse about a bit to shift furniture in my Living Room and see where to best place my home theatre speakers for best effect. I’m dreaming about setting a bit of growth in motion so I can work out how best to prune the nearby plants to show that new shrub off to best effect as it grows. I’m wondering about how good it would be if I dragged that file over to the filing room door in the corner of the screen, dropped it in the filing cabinet I wanted it in and had the thing automatically titled and catalogued for me in accordance with the index system I had set for that drawer!


What does it make YOU wonder about? Welcome to the future, if you want to be part of it :D

  • 1 Contributor
  • 0 Replies
  • 210 Views

Dani AI

Generated

had the right hunch: physics acceleration promised a real shift. The actual history, however, shows that the standalone “physics card” idea morphed into something else — the company behind those cards was absorbed and the add-in hardware never became a mainstream consumer platform. The technology survived, but as libraries and GPU compute rather than as a mass-market peripheral. Nvidia completes absorption of Ageia. (theregister.com)

What followed was consolidation and evolution. The PhysX engine moved from dedicated PPUs to GPU-accelerated and CPU paths under Nvidia’s stewardship, and the SDK later became open-source so developers could extend and inspect it. More recently, Nvidia has published the previously withheld GPU simulation kernels to the public repositories, so the GPU-accelerated pieces are now available to researchers and studios. See reports on PhysX going open and the official GitHub discussion. Tom's Hardware on PhysX open-sourcing (2018) · PhysX GPU code on GitHub. (tomshw.it)

Today the practical route for “physics everywhere” is engine middleware and GPU compute rather than a separate board. Unity’s built-in 3D physics integrates Nvidia PhysX (with DOTS/Havok options for high scale), and Unreal Engine 5 ships Chaos as its native solver — both provide production workflows and performance guidance. For real projects: profile early, favor primitive colliders for dynamics, pre-cook or simplify mesh colliders, use sleeping and LOD/approximation for nonessential objects, keep physics on a fixed timestep, and push large particle or fluid sims to GPU/compute where practical. See Unity’s physics docs and Epic’s Chaos performance notes for concrete optimization techniques. Unity Physics docs · Chaos in UE5. (docs.unity3d.com)

The UX dreams you sketched — dragging furniture to test layout or physically moving files — are closer now thanks to spatial/AR platforms and real-time engines, but they live mainly in AR/VR and specialized apps (not as the default desktop). If prototyping those interactions, build in adjustable fidelity: approximate physics for UI affordance, higher-fidelity simulation only where it matters, and test on target hardware (and for networked scenarios, design determinism/authoritative servers from the start). See recent spatial-computing platform docs for example workflows. Apple visionOS / Vision Pro examples. (apple.com)

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.