Google Acquires Rights to 20 Year Usenet Archive

[Hammer]

The Processor Core: The Amiga would have stayed with its
parallel processor core design. We always knew that one of the things
that made this such a versatile and powerful machine was the fact
that, even though main processor speed was relatively slow, much of
the burdensome work of processing graphics and sound was done by not
one, but two other chips, which when combined, made a very powerful
factory for multimedia apps. So, the Amiga today would sport not one,
but, actually three equally powerful, equally advanced, state of the
art Motorola chips, each designed specifically for jobs in sound, and
graphics. Also, each of these powerful processors should have direct
control and access to the system resources they require, including
proprietary and system-wide RAM and transport buses. In other words,
no cards!

NVIDIA’s Geforce 256 GPU murders any current Motorola CPU when it comes to 3D acceleration.

Note that, modern DX9 VPUs are massively paralleled VLIW(1)/SIMD(2)/MIMD(2) pipelined cores. They are noted to be one of the fastest hidden** DSPs in the market place (when loaded with BionicFX Apps).

Examples
1. NVIDIA Geforce FX family.
2. NVIDIA Geforce 6x00 family and ATI Radeon Xx00 family.

ATI Radeon X800 has theoretical of +200GFLOPS i.e. equalling ~1 Sony Cell. "Rage Max" is an ATI reserve technology for multi-VPU cores.

3DLabs's Wildcat Realizm 800 is quite dangerous since this sucker has theoretical of ~700GFLOPS (shame about the drivers).  

[Hammer]

nVidia has removed their much-applauded audio engine from their nForce chipset, and they have publicly stated that hardware-accelerated audio is dead.

Note, NVIDIA is currently revaluating SoundstormII i.e. to compete with the joint Intel/Dolby Lab’s Dolby Live initiative.  

I saw a demo where an ATI GPU was performing realistic fluid simulation as a smoke demo -- in the GPU itself. A few years ago, only powerful, full-featured CPUs with floating point support could do stuff like that.

Particle physics and blood simulation on NV GPU was nice.

GPU’s large cache such as ATI’s hyper-memory and NV’s turbo-cache are just the beginning for large cache equipped GPU cores.  

The next evolution for DX class VPUs is the Unified Shader Model 4.0 i.e. probably under DirectX9d moniker.