Will this trend always stand?

firewolfsm

Golden Member
Oct 16, 2005
1,848
29
91
It seems that for many generations now, for enthusiest systems, graphics cards have had about 1/4 as much onboard memory as was available to the processor. Now, a high end system will have 1GB of VRAM and 4GB of system ram. 4GB of RAM only recently became cheap and popular, earlier this year it was 2GB, that was also when only 512-768MB of VRAM was on the highest end cards.

The trend does go back to the 9800PRO era at least, if not further.

Any thoughts?
 

Rebel44

Senior member
Jun 19, 2006
742
1
76
I dont think this trend will continue. More then 1GB VRAM per GPU is now useless and you cant have too much system RAM :)

I have HD4870 512MB VRAM and 8GB RAM (I often use 5+GB).
 

QuixoticOne

Golden Member
Nov 4, 2005
1,855
0
0
I think that the architecture of GPUs will change pretty radically over the next 4 years or so.

Some companies are working on combined GPU + CPU chips in which case one would expect there to be less of a distinction between system RAM and GPU RAM.

Certainly today even a 4 core Intel CPU is starved for memory bandwidth and can't execute a full 128 bit wide load + store from/to RAM per every core per every clock cycle per every core. That is only going to get worse with CPUs of 8 cores or 4 cores + hyper threading or whatever.

Thus RAM bandwidth needs to expand past the current approximate 4GBy/second out to at least 16GBy/second and beyond just to keep up with moderate multi-core CPU based processing.

High performance GPUs usually need around 50GBy/second to 150GBy/second of VRAM bandwidth for today's top generation of single GPU cards.

CPUs with dual channel memory controllers as is now typical have 2 x 64 bit wide paths to the system RAM for 128 bit wide parallel I/O.

Mid-range GPUs also have 128 bit wide data paths to VRAM, with higher end GPUs having 256, 384, or even 512 bit wide paths to VRAM.

The 'Nehalem' generation of PC CPUs / chipsets will have triple channel and quad channel DDR3 models handling 192 bit wide and 256 bit wide paths to system RAM.
http://en.wikipedia.org/wiki/N...em_(microarchitecture)

This suggests a possible convergence of system RAM and VRAM for mid-range GPUs integrated with mid to high end CPUs over the next 2-3 years concordantly with a possible integration of CPU and GPU either on a single chip or on a coprocessor slot / socket basis (quickpath?).

The other factor is that memory prices fall and memory densities increase. It isn't altogether clear that the average desktop (or even gaming) PC will need more than 4GB-8GB of memory for at least a few years until software catches up with the hardware capabilities and 64 bit applications begin to dominate and common applications that can actually use more than 2GB usefully emerge.
Thus the amount of PC ram may stall while the amount of VRAM still is increasing to the 1GB-2GB / GPU level and beyond. Once you have 2GB-4GB of fast GPU RAM you almost might as well use it as system RAM as well if you have the architecture to support that.

Another factor is that graphic resolutions are increasing with 30" sized LCDs becoming common, and bigger ones following rapidly. This will shift the need for VRAM out to 1GB minimum and 2GB common within the next 12 months. This is simply to permit today's "cartoon quality" graphics. Within another 2-4 years GPU power will keep increasing to the point that whole new levels of image quality become feasible with the ultimate goal being basically photo-realistic 3D in real time. That would need something like 8-16GBy of VRAM at least, and probably with new generations of dynamic compression / 3D interpolation et. al.


Another factor is the limited lifetime of 2D displays; eventually we'll start to have more common true 3D display technology in which case we'll see even more need for 16GBy+ VRAM quantities and GPUs that fully outpace today's CPUs in every way. At such a point it will be irrelevant to talk about VRAM vs system RAM since the GPU will be quite enough to be your CPU and VRAM will be far bigger than you likely need for system RAM.

 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
some companies? AMD is calling it fusion, and intel is calling it larabee (later larabee will involve a "CPU" with small FLOPS intensive chips and large, single thread optimized chips, aka CPU + GPU on one chip).

Speaking of ram requirements for realistic graphics... movie level special effects today require computers with 32GB, 64GB or even 128GB of ram to render (to be able to load all that detailed data).
 

QuixoticOne

Golden Member
Nov 4, 2005
1,855
0
0
Indeed. And IIRC NVIDIA is trying to buy out a CPU company as well presumably to produce its own CPU + GPU chip(set).

As for memory requirements, you're certainly correct that those figures would be more representative of what is needed, I was just assuming we'd pass through an evolutionary stage or two between today's 2GB and the ideal 128GB or whatever.

X^3 increases quickly. An 8 Megapixel camera is basically photo realistic for a monitor sized print out, and that's with 2800 x 2800 resolution at 3 bytes (Red = 8 bits, Green = 8 bits, Blue = 8 bits) per pixel.
Add another dimension and that's 2800 x 2800 x 2800 = 21,952,000,000 or 22 Giga pixels * 3 = 66
Gigabytes just to store one static 3D "scene" excluding transparency data which at a minimum would add one more byte per pixel to give 88 GBy/scene at that resolution.

Of course using NURBS or 3D wavelets instead of pixels would help to compress that considerably...

Maybe in a few years I'll finally get the personal planetarium (or likely better) that I've always dreamed of.

If anything maybe they'll start building the GPU into the display device just so they won't need 1TBy/second data links between the GPU and the display to push all that data across a cable.

It'd also be a good opportunity for "smart RAM" where processing technology is integrated right into the fabric of the 'RAM' more like the way a human retina / visual cortex works.