If his argument was that they were fine at running 3D, then that would be a fair reply. But since he's talking about them running non-3D tasks fine, which they were.v1o wrote:And Computers were not running just fine without 3D non-accelerated graphics, in those days people used highly expensive SGI Workstations for heavy graphics work because PC's were just not up to the task of complex 3D graphics.EvilDragon wrote:
Wrong counter-argument. GPUs are there because gaming industry "made them necessary". Computers were running just fine with non-3D accelerated graphics cards long before nVidia, ATi, Matrox, 3dfx, etc.
And for the record, the SGI kit of fifteen years ago might have been better at realtime 3D, but by 2000 it was being surpassed by generic PCs; at that point in time I was getting faster renders out of £1K PCs than I was £18K SGI Octanes, and only a year or two later, a programmer colleague was able to get 3 linked generic PCs to drive the same kind of realtime immersive 3D as we'd been using a £500K SGI Onyx2 for.
Most 'special effects' work doesnt run on the GPU, its primarily rendered using the native CPU. And whilst there are some GPU-enabled renderers today, an awful lot of what peope consider 'special effects work' consists of compositing, which rarely uses GPUs.That includes special effects work for the movie industry, scientific work and the games industry.
It wasn't until modern GPU's came out that performance levelled out, to the point that PC's were good enough to use and there was no need for exotic hardware such as the likes made by SGI. Ironically it was mainly former SGI engineers designing the silicon at Nvidida and ATI.
OTOH, these do generally get done on CPUs, when it comes to rendering.Additionally you could run early 3D games like Quake and Unreal with software-only rendering on the host CPU, but the performance and feature set was always severely reduced compared to running on the GPU. That's because GPUs are massively parallel (an order of magnitude more than CPUs) and have dedicated silicon dedicated specific to the task. They're not just designed for general purpose code. The gulf in performance can be proven by running benchmarks or even writing your own code and trying to do intensive tasks like global illumination, tessellation and self shadowing on a host CPU.
If you've ever watched a renderfarm grind to a crawl on frames which exceed the memory of the system they're running on, and thus start swapping out, its a reminder of why raw GPU power doesnt necessarily help when you need 16Gb to calculate your frame.
For some rendering tasks, for some renderers, with specific circumstances, and particularly for realtime, with the associated compromises, GPUs make sense. But for the actual stuff you see in the movies, game cutscenes, etc, that's mainly CPU work. There's a reason VRay was native first, and the GPU based stuff is still in progress.
