• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

NV20 Details

In fact, there is no official info on the upcoming monster from NVIDIA, NV20, yet. However, this chip is a No 1 topic for discussion today. There is the whole lot of rumors on the web about it and its possible specs. Here is a bit of some recent info we managed to collect from all over the Internet:

Core frequency: 300MHz.
Memory frequency: 500MHz. Graphics cards based on NV20 will be equipped with the memory chips from Samsung and Infinion, because these companies can boast the fastest DDR SDRAM chips.
Manufacturing technology: 0.15 micron. The chips will be manufactured by TSMC.
Amount of transistors: 50 million. It is twice as much as by GeForce2 GTS and 8 million transistors more than in Pentium 4.
4 rendering pipelines. NVIDIA seems to be increasing the speed of its upcoming product not by introducing more rendering pipelines but by optimizing memory requests, which is the today?s bottleneck of all classical graphics accelerators.
3 texturing units per pipeline. This is very similar to ATI RADEON. NVIDIA is probably expecting new games to come out, which will lay three textures over one pixel simultaneously.
DirectX 8.0 compatible geometric co-processor.
Curved surfaces and 3D-textures hardware support.
Hidden Surface Remove (HSR) support: this is a z-buffer algorithm, which allows to reduce the memory bus workload, because it eliminated the need to texture hidden surfaces.
Heat dissipation: around 15W. This is quite a lot for a graphics accelerator. That?s why new cooler construction may be required.
NV20 is expected to be announced in early spring 2001 and the first graphics cards on this chip should cost within $500.
 
My question is, why are new releases of cards priced higher and higher? I mean, when the TNT and Voodoos first came out, they were highend cards, but the prices were about $150.. then you had the GeForce, highend cards, but first batches were $200 or so. GF2, as soon as they hit the streets they're $300. Now with the NV20, they're expected to cost $500 at release?! What's going on here? Even taking into account inflation (which has been an all time low), these prices are insane. Can anybody explain this?

 
back then tnts did not have 50 million transistors or high speed ddr memory. more transistors = more space per sheet per core meaning lesser cores per platter. which causes costs to increase.
 
but die shrink lowers costs per chip.

and allows for lower power consumption, as well as faster clock speeds. <-- affected by core design of course, but holds if all else the same ;-)
 
and my super reliable undercover in nvidia told me it is almost tap-out. So, we are going to see it very soon.
 
Back
Top