NVIDIA's GeForce FX Ultra
A Pre-Production Showcase

By - Dave Atlavilla
January 27, 2003

The wait for NVIDIA's new GeForce has been almost an eternity it seems.  Since their initial hints at their next generation technology, back in June, there has been much speculation and rumor milling with regards to the specifics of the retail board level product and its forthcoming release to the public.  Also since that time, NVIDIA's arch rival ATi Technologies, has released several  variants of retail product, based on their R300 VPU.  This release of ATi product not only arrived much sooner than NVIDIA's next gen technology, but also delivered a stinging blow to NVIDIA's historic dominance in 3D Graphics.  The move to .13 micron technology proved to be too much of a stretch, for the then flawlessly executing 3D Graphics ASIC design machine out of Silicon Valley.  They bet the farm on TSMC and their bleeding edge manufacturing process but their horse just didn't come through in time.  Since it's initial debut in November at Comdex,   NVIDIA's Marketing Teams were hard at work positioning the product and readying strong branding campaigns, targeted at maximizing the impact of the product in the market place upon its release.  The dates came and went.  Targets were set for a pre-Christmas delivery but again the GeForce FX was held up, all the while ATi has been enjoying the spoils.

However, the day is finally drawing near, when NVIDIA will have their GeForce FX product out into retail channels in volume.  With an extremely large Heat Sink and Blower attached to the front, it seems as though the Design Team at NVIDIA has finally gotten the core stable and ready for prime time.  However, the product will still be in very short supply for the next few weeks and we are sorry to report to you, that we still haven't had the opportunity to test the product in our labs at HotHardware.  Instead, what we've prepared for you today, is a closer look at some of the hardware specifics of the GeForce FX as well as the software, drivers and functionality of the product.
 

Specifications & Features of the NVIDIA GeForce FX
Retail specs confirmed...


CLICK ANY IMAGE FOR AN ENLARGED VIEW

GPU:
  • GeForce FX 5800 Ultra -- 500 MHz core
  • GeForce FX 5800 -- 400MHz core
  • .13u Manufacturing Process
  • 125 Million Transistors ( 2X GeForce 4)
  • 256-Bit GPU
  • Flip-Chip BGA Package with copper interconnects
  • 8 Pixel Per Clock (8 Pixel Pipelines)
  • 1 TMU Per Pipe (16 Textures per unit)
  • 350 Million Triangles per Second - 3x The Geometry Performance of a GF4 Ti
  • AGP8X (2.1GB/s bandwidth)

Memory:

  • Memory with 1GHz Data Rate (500MHz DDR)
  • GeForce FX 5800 Ultra -- 1GHz Memory Clock
  • GeForce FX 5800 -- 800MHz Memory Clock
  • 128-Bit "DDRII" Type
  • 128MB & 256MB Memory Capacity
  • 3rd. Generation Lightspeed Memory Architecture

  • 48GB/s Effective bandwidth through the use of compression techniques  (16GB/s actual @ 500MHz)

Features:
  • Full DX9 Compliance
  • 64-Bit Floating-Point Color
  • 128-Bit Floating-Point Color
  • 2 x 400MHz Internal RAMDACs
  • Long Program length for Pixel and Vertex Shading
  • Conditional Execution for both Pixel and Vertex
  • True Data-Dependant at Vertex
  • Unified Vertex and Pixel Shading instruction set
  • Unified Driver Architecture
  • nView 2.0 - Multi-Display Technology
  • Digital Vibrance Control 3.0

 


We're not going to comb through all of the above specifications again for you here.  If you would like a good refresher on the technology, please see Marco's Debut Article, right here.  However, we now have full confirmation on retail product speed grades and models.  The GeForce FX 5800 Ultra (the card we expect to have in the lab here for testing very soon) runs at a 500MHz core clock speed, supported by a 1GHz DDR2 (500MHz Double Data Rate) 128 Bit DRAM subsystem.  The GeForce FX 5800 version has a lower core speed at 400MHz and 800MHz DDR2 Memory.  Both cards have 128MB of on board DDR2 memory and are fully AGP8X capable and compliant.  The GeForce FX 5800 Ultra variant, as you can see, has an impressive and large heat sink and blower unit, strapped on top of the GPU and it also makes contact with the small BGA DDR2 memory chips.  We have mixed emotions about this cooling setup that NVIDIA has designed for their new flagship card. 

On one hand, this cooling system is an impressively designed solution that most likely provides fantastic heat transfer and radiation performance.  It's probably one of the reasons why NVIDIA is able to clock their GPU so high at 500MHz, in the first place.  On the other hand, while it is impressive to look at, it seems to be a much less elegant and more costly solution, than the standard Heat Sink and Fan assemblies of more traditional cards on the market.  Also, this cooler consumes an extra PCI slot location in your chassis, adjacent to the AGP slot it plugs into.  Although this isn't much of a show stopper, since most folks don't use all PCI slots on a motherboard these days, it still seems like a waste of real-estate.  Regardless, if you know the way we think here around HotHardware.com, you'll know that if we can get even an extra 10MHz speed bump, we'll take it and good cooling is essential.

More Technology and Technology Demos