Galaxy GeForce 7600 GS GDDR3 video card review

For several generations now, NVIDIA have impressed with the performance on offer from their mid-range offerings.  Even during the lean GeForce FX era, NVIDIA followed up the poor GeForce FX 5600 with the 5700, which turned out to be arguably the most impressive member of the product family at that time.  This was then followed up by the introduction of the GeForce 6600 during the GeForce 6 series era, which quite simply blew away the competition in its price range.

Until recently, with a late challenge courtesy of ATI's Radeon X1800 GTO part, this story has repeated itself courtesy of the GeForce 7600, which takes everything that was good about the GeForce 6600 and gives it a GeForce 7 series coating.  Of course, NVIDIA's flagship GeForce 7600 part bears the GT moniker, but for those looking for a slightly cheaper mid-range part, the GeForce 7600 GS is the way to go, being as it is a fully featured version of the 7600 GT, albeit with lower clock speeds.

One NVIDIA AIB partner, Galaxy, is currently producing two boards based around the GeForce 7600 GS - One is clocked at stock speeds and utilising GDDR2 memory modules, while the other uses GDDR3 as well as featuring some other intriguing goodies.  It is the latter of these two parts we'll be looking at today, so strap yourself in for our analysis of a GeForce 7600 GS board that really is rather special...

GeForce 7600 architecture

As we mentioned in our introduction, the GeForce 7600 series of parts take what made the GeForce 6600 so impressive, and gives it a glossy new coat.  At its heart, the GeForce 7600 utilises NVIDIA's 90 nanometre G73 core - This features twelve pixel pipelines (or three fragment/pixel quads), which also means that twelve texturing units are present on the die.  This is matched up with just eight ROP units, meaning that although twelve pixels can be worked on by the architecture at any one time, only eight can be output.  This is a similar set-up to its predecessor, NV43, which utilised eight pixel pipelines paired up with just four ROPs.  If the concept is a little alien or baffling to you, let me reiterate and update slightly my explanation of this configuration from my GeForce 6600 GT review at the time.

The reason for this pipeline/ROP configuration is that ROPs are often left idle while work in the shader pipelines are taking more than one cycle, thus rarely would all twelve output units required at any one time, especially with the kind of work being asked of a mid-range board. If you're looking for a real-world comparison, think of queuing to enter a football stadium that has twelve turnstiles, but only eight entrances to the actual seating beyond that. It takes time to process peoples tickets as they enter the turnstiles, with some taking longer than others while they fiddle with change and the like, meaning that there is rarely twelve people making their way through the turnstiles at any one time, and thus hardly ever a crush of customers trying to get through the eight entrances to the seating area.

Finally on the architectural side of things, G73 contains a total of five vertex shading units.  Aside from this, the GeForce 7600 retains all of the features of its bigger brothers in the GeForce 7 family, including its additional processing abilities in the ALUs to put it some way beyond the GeForce 6600 series with regard to its shader power.  Clock speed wise, the GeForce 7600 GT sees itself granted reference clock speeds of 560MHz on the core and 700MHz DDR for memory, with the 7600 GS clocked lower at 400MHz core and just 350MHz memory.  For reference boards, we also normally see the GT part utilising GDDR 3 memory, while the GS boards feature GDDR 2.  We've already established the part we are looking at today from Galaxy features GDDR3 RAM modules, but that isn't where the differences from a reference board end by any means - More on that in a moment.

You can see the entire feature set of the GeForce 7600 GS below.

• CineFX 4.0 Architecture
  • Full DirectX9 Support
  • DirectX9 Shader Model 3.0 Support
    • Vertex Shader 3.0
    • Pixel Shader 3.0
    • Internal 128-bit Floating Point (FP32) Precisions
  • Unlimited Shader Lengths
  • Up to 16 textures per pass
  • Support for FP16 Texture Formats with Filtering, FP32 without
  • Non-Power of two texture support
  • Multiple Render Targets
• NVIDIA High Precision Dynamic Range Technology
  • Full FP16 Floating Point Support throughout the entire pipeline
  • FP16 Floating Point Frame Buffer Support
• Intellisample 4.0
  • Up to 4X, Gamma Adjusted, Native Multi-sampling FSAA with rotated grid sampling
  • Transparent Multi-Sampling and Super-Sampling
  • Lossless colour, texture, z-data compression
  • Fast Z Clear
  • Up to 16x Anisotropic Filtering
• UltraShadow Technology
• NVIDIA SLI Support
• NVIDIA Pure Video Technology
  • Adaptable Programmable video processor
  • High Definition MPEG2 and WMV9 acceleration
  • Spatial Temporal de-interlacing
  • Inverse 2:2 and 3:2 pull-down (Inverse Telecine)
  • 4-tap horizontal, 5-tap vertical scaling
  • Overlay color temperature correction
  • Microsoft® Video Mixing Renderer (VMR) supports multiple video windows with full video quality and features in each window
  • Integrated HDTV output
• Advanced Display options
  • Dedicated on-chip video processor
  • nView Multi Display technology
  • Single and Dual-Link TMDS Transmitter
  • Digital Vibrance Control 3.




Prev - Next >>