|
|
|||||||||
|
|||||||||
|
|||||||||
| |
||
| |||||||||
![]() |
|
|
«
Previous Thread
|
Next Thread
»
|
Thread Tools | Search this Thread | Display Modes |
|
|||
|
A different paradigm
There are new materials and methods coming out, transparent silicon and other semiconductors are fast becoming reality. Diamond-growing methods have advanced to the point where that material can be utilized in circuit-design. Organic LEDs are hitting the shelf as we speak, and persistant pixels are in development.
So here's my thought: What if you could see the contents of your Video RAM directly. The 32/64/128 bit contents of each memory location represented by a single pixel of a specific hue, saturation, luminance. No need to worry about refresh rates or scanning frequencies, the pixels reflect changes in the memory as they are made by the video subprocessor. The video card circuitry can fit comfortably on the back-side of the panel, and cooling is not a problem because the entire panel can be heat-sinked. There's more room in the computer case since there would be an interface card plugging into the video slot and a high-speed networking cable connects it to the display. Now the questions: Is it worth the time/money/resources to try and effect this dream? What disadvantages can be seen? The one that comes to mind immediately is that ugrading memory involves replacing the entire panel. Expense would be a very limiting factor initially, as well. Thoughts?
__________________
Yes, it is very possible, in fact, highly probable, that you are wrong. You would be a better person if you would just admit that. And I would be less annoyed. ![]() "I like doggies....and apparently dinousaurs, mostly because they eat cats!" --Nilpo |
|
|||
|
The Ultimate for Gaming/video snobs? imagine buying your XXXXX vid system, and deciding you want 2gig onboard ram. Remove a plate inside that plugs into a slot, and then drop in a new plate, complete with ramsinks and an upgraded heatpipe cooling system. Cost? High. But, as with everything else, it drops until Everyman can afford them.
The future excites me.
__________________
Farsi for Teh W1n |
|
|||
|
An interesting concept. The display panel also being the framebuffer, with any further buffer being offscreen somewhere. As hey said though, drop in a new plate, increased res = increased onboard ram.
The upgradable graphics subsystems draw closer, hell, laptops already have it...not all of course. Graphics motherboard, with optional single, double or even quad gpu sockets, several so-dimm slots for memory upgrades, all bolted to the back of your OLED screen. Sounds like a dream. Snapping back to reality a moment though, graphics hardware requires massive bandwidth. Much more so than just pushing pixels ala dvi. Think about it, take a screen, lets say 1600x1200, a good res, give it 24bit colour depth (lets face it TRUE >=32bit screens cost a ton, and will probably be quite a way off for us mortals), and lets assume this screen is an lcd, and give it a 60Hz vsync. The bw required is simply 1600*1200*24*60 = 2.7Gbps (w/ G as 10^9 rather than 2^30). Mind you, this is raw pushing all pixels all the time. If you just updated the pixels that change, it should reduce somewhat. Keeping that figure in mind, graphics cards are now using PCIe x16. 16 lanes of PCI Express. 16 * 2.5Gbps. 40Gbps. Over this massive pipe, all the geometry, texture and lighting information is pushed though. Not a compiled image, but how to compile the image. The source is always larger than the finished version ("Hello, World!" et al not included). Improvements in wireless tech, with UWB starting to finally come onto the market in various guises, a decent wireless display should be able to become a reality soon ![]() |
![]() |
| Viewing: Dev Hardware Forums > HARDWARE > Hardware Development > A different paradigm |
| Thread Tools | Search this Thread |
| Display Modes | Rate This Thread |
|
|
|
|
|