| fuwuyodksuwt | Date: Saturday, 03 Aug 2013, 03:50:03 | Message # 1 |
|
Private
Group: Users
Messages: 3
Status: Offline
| Still, graphics card will be more powerful then your CPU will by a huge margin, least for the next five years or so. Don't know how AMD's fuzion is going to work out yet.
However, most gaming logics require short, branched operations that such a processor doesn't handle well.
Anyone have anymore insight on that?
'Current' processors have an integrated FPU; however, it is much less powerful than a graphic card's dedicated GPU, making software rendering of a scene much slower on those. It doesn't mean processors can't render a scene, it just means they need MUCH more time.
Note: softwares like 3D studio used to be CPUonly. Even now a 3D card is used for prerendering, not for final rendering. This is due to the various optimizations used in the card and in the driver that make several computations approximative and a bit unreliable.
Maybe we'll get a hybrid CPU/GPU in the future (AMD FUSION??? or INTEL's next moveforgot the name!!! ?) Right now, what you pay for in a CPU is mostly SRAM not DRAM (one SRAM byte uses 8 transistors while a DRAM byte would only need one) which is one reason why a GPU is cheaper (!!!!?????) than a CPU.
The Cell processor is a step towards creating a hybrid CPU/GPU; it is very good at computing loads of floating point operations very fast <a href=http://www.asicsrunningshoes.eu>Asics Run Shoes</a> and does away with a graphics card.
Yeah GPUs have a specific instructionset and architecture designed for 3D rendering. CPUs simply have a broader range of things to worry about. This is called RISC, I think, and <a href=http://www.asicsrunningshoes.eu>Asics Run Shoes</a> you see it in a lot of places such as encryption cards in servers, and sound cards.
If you say yes, let's take a look on the fastest GPU and on the fastest CPU. With 700$ you got http://www.asicsrunningshoes.eu 8800 GTX 768MB DDR3 RAM, while Quadcore X6700 cost 900$ and you don't get RAM yet. It runs at a lot more Hz than your desktop CPU. that doesn't make it more advanced. It just makes it different. A GPU is a streaming processor. It took it a while to be able to handle branching (conditional loops) and when it did, we ended up with a huge latency on the G80. CPUs generally sacifice absolute performance with backward compatibility and ease of programming (ever tried to calculate a GPU?, even with Clike code!!!).
As such, the very ancient idea of a core CPU targeted at computing integers with a FP coprocessor is coming back today as graphic cards with programmable shaders (which aim at computing large amounts of FP data as fast as possible).
Is it possible a CPU for 3D Rendering
I believe the architecture itself also limits the speed. GPUs will always be slower clockwise because of the job they do.
I asked this question myself <a href=http://www.asicsrunningshoes.eu>asicsrunningshoes</a> not to long ago, and 3DMark06 answered my question. By 2.4 gigahertz Core2 with 2 megs of Cache couldn't render at a stead 1 frame per second for it's test. My lowend PCIExpress graphics card could have rendered the same test at hundreds of frames per second. Intel's 80 core processor running at 78 gighertzs would equal about what my graphics card does for graphics. That's about what it seems, anyway. Could be way off on that last point.
|
| |
|
|
