OSNews

total memory = 128 MB
avail memory = 115 MB
bus 80 MHz
cpu 3192 MHz

??? My best guess is they have to fix something (:if not anything, access to whole 256MB would be appreciated:) But, hail to the hackers working on this one.

Define "really good" bits. Linux gets full access to the Cell chip in the system, which is probably the most interesting part of the machine. The RSX isn't much different than any other NVIDIA chip.

The RSX isn't much different than any other NVIDIA chip.

Actually, RSX is legacy part of PS3. Sony can't expect coders will start working on games for Cell the same moment PS3 hits the streets so they included RSX, while real preferred method would be avoiding RSX as much as possible and offload those jobs on Cell. Mesa for example running on Cell would be far more attractive than any other DRI, GLX implementation.

In perfect world, the only thing needed from RSX is to be able to display 2D.

Edited 2007-01-09 09:27

"real preferred method would be avoiding RSX as much as possible and offload those jobs on Cell."
ROTFL

"total memory = 128 MB"
64Mb is dedicated to PS3OS + hypervisor
256-64 = 192 Mb should be available.

"real preferred method would be avoiding RSX as much as possible and offload those jobs on Cell."
ROTFL

And where do you think those Cell 3D demonstrations were running (1mach plane simulation, complete heart simulation)? On some graphics card? Dream on.

If you check current engines you can see that they don't support Cell yet.

256-64 = 192 Mb should be available.

I stand corrected

Edited 2007-01-09 17:25

This isn't even close to true. Cell is far from being ideal for rasterizing 3D graphics. Take something very simple like trilinear filtering. For every pixel you write out, you need to load 8 texels and blend them together. The SPE can do one memory load and one arithmetic operation per clock cycle. You're talking dozens of clock cycles just to get a single texel value, not to mention the final blending or multi-texturing or bump-mapping, or any of those things. Meanwhile, GPUs have multi-ported texture caches that can do the memory reads in just a couple of clock cycles, and lots of dedicated interpolation hardware in the texture units that can allow the chip to provide a texel to the pipeline every clock cycle.

There is a reason GPUs are advertised as performing trillions of operations per second. They have a ton of fixed-function hardware that is necessary for the general-purpose units (the shaders) to do their work efficiently.