Slashdot Mirror
Pushing The 512MB Barrier On Video Cards
Hack Jandy writes "Remeber your ancient TNT graphics card that had 16MB of memory? ATI is pushing the texture barrier by incorporating 512MB in their newest X850 video card lineup. The catch? Even ATI acknowledges there will probably be no performance benefits to bumping the memory support from 256MB to 512MB as the cards are 'intended to demonstrate the next-generation capability to gamers." An anonymous reader points out that Gainward (which sells NVidia-based graphics cards), will shortly introduce its own 512MB card, according to Hexus.net.
"Remeber your ancient TNT graphics card that had 16MB of memory?
Man you were lucky. I had to deal with a 1MB video card in my job workstation.
Honestly, its not all that impressive to see these high numbers for video card ram. Different needs pushes the limit nowadays. It used to be pushed to deal with higher color palettes at higher resolutions. Now its all about texture mapping.
Even ATI acknowledges there will probably be no performance benefits to bumping the memory support from 256MB to 512MB as the cards are 'intended to demonstrate the next-generation capability to gamers."
Translation: Even though it's not practical, we'll sell it since gamers will buy it.
"If my email tells me anything, size DOES matter."
Seeing as how most of the 'realism' of a 3D game comes from detailed textures, yes, size of texture ram does matter.
"Derp de derp."
So many cards today are bloated with RAM that they don't utilize. Case in point: 256MB FX5200. But consumers think they are getting more, and getting people to buy your product is all that matters.
I find most of the realism comes from the physics engine. The texture just makes it look a bit prettier, but by no means makes the game any better.
Ever wonder why GPUs are such a big deal and sound cards are such an after thought? It's all about numbers. ATI and nVidia can increase clock speed and double memory and make it look really impressive. Sound cards can't really do that.
If I were Creative I'd start including massive amounts of RAM on my cards. Plus, I'd throw a CPU in there too, if there isn't one already, and start hyping the clock speed. I'd even have a program to overclock both.
That way all the ignorant fanboys would start buying them simply for bragging rights.
If someone says he and his monkey have nothing to hide, they almost certainly do.
Kinda sad when your video card has more ram than first/second/third computer had disk space (combined). Your processor has more cache than your first comp had ram. And I'm waiting for the day that a processor has more cache than my first comp had disk space.
Fly me to the moon Let me sing among those stars Let me see what spring is like On jupiter and mars
If you use uncompressed textures on Ultra setting, it tries to use 512 MB of video memory. I really doubt that the difference quality is really noticeable during actual gameplay. However, hardware review sites love to take screenshots and magnify them 1000x to analyze the lines on ceiling tiles, so I guess it will make them happy.
It will lower prices on the 256MB versions and all the cards beneath, giving me better deals when I pay $200 for a 7 month-old card that still plays the latest and greatest just fine.
Are we seeing a generation of "boy gamers" equivalent to the "boy racers" that add big tail-pipes, chrome and LEDs to their cars. 512MB sounds good, but basically you're buying features - not performance.
Engineering is the art of compromise.
Current games were developed targeting 64 or 128MB graphics cards as the majority and 256MB as the high end, so there are very few scenes with more than 128MB of content in them. Adding more RAM will not increase performance or quality in this case, since it's just left empty. Games won't begin to really take advantage of 512MB of VRAM until a significant percentage of the market owns it.
3DLabs pixel shader / vertex shader implementation is broken and incomplete. Forget about DOOM3, Half-Life 2, FarCry, or other modern games rendering properly.
All in all this is a pretty fair assessment, but I think it leans a bit to the pessemistic.
First, I doubt a window manager would actually use a 32 bit z-buffer. 8 bit would be overkill here (enought to specify a unique depth for 256 windows). Even a 3D window manager would get by on 16 bit depth no problem, I believe it's the most commonly used depth for most true 3D apps now.
Also, I doubt that in many cases more than a small number of true 3D windows would be needed. Someone who is working with 30+ windows open most likely has mostly terminals, web pages and text editors open, with maybe a few 3D apps.
The ultimate plays for Madden 2006
Eh.. this guy still looks unnatural to me. The chains around him STILL look like the flat textures they are, as do his teeth, and the joints on the gun. I'm sure the chains will stretch unnaturally as the creature moves, and the barrel of the gun is still a hexagon.
There's still a long way to go and, in fact, I don't think we'll ever reach the point where a single processor will be capable of creating an image on the fly that matches the quality of a prerendered.
https://www.eff.org/https-everywhere
Someone who is working with 30+ windows open most likely has mostly terminals, web pages and text editors open, with maybe a few 3D apps.
Doesn't matter. The window server (Quartz in my case) treats all of them as texture-mapped polygons, where the "texture" is their actual content.
How to solve most of our problems: 1.Lots of nuclear plants. 2.Cure aging.
Yes, but the parent was stating a difference between apps which require 3D capabilities themselves vs. those that only require 3D like capabilities from the window manager.
2D apps treated like textured polys do not need their own depth buffer, frame buffer, etc. They just need a texture buffer, and the window manager treats them like texture polys in a single, comprehensive 3D app. Conventional 3D apps require their own depth buffer, frame buffer, textures, etc. in addition to that used by the window manager. I was pointing out that while the requirements for 30+ 2D apps are fairly large, they are considereably smaller than the same number of true 3D apps, because they have needs beyond the WM overhead.
The ultimate plays for Madden 2006