Home

News

Forums

Hardware

CPUs

Mainboards

Video

Guides

CPU Prices

Memory Prices

Shop



Sharky Extreme :


Latest News


- SanDisk's SSD As Rapid As It Is Reliable
- OCZ Launches Limited-Edition SSD
- AMD Offers Low-Profile DirectX 11 Graphics Card
- Microsoft Gaming Keyboard Has Awesome Anti-Ghosting
- Matrox Lets Multi-Monitor Fans Double Up
News Archives

Features

- SharkyExtreme.com: Interview with Microsoft's Dan Odell
- SharkyExtreme.com: Interview with ATI's Terry Makedon
- SharkyExtreme.com: Interview with Seagate's Joni Clark
- Half-Life 2 Review
- DOOM 3 Review

Buyer's Guides

- February High-end Gaming PC Buyer's Guide
- November Value Gaming PC Buyer's Guide
- September Extreme Gaming PC Buyer's Guide

HARDWARE

  • CPUs


  • Motherboards


  • Video Cards






  • A quick glance at the specification sheet reminded us of something NVIDIA told us right before the launch of the GeForce3 - mainly, "processing graphics takes raw transistors - there is no way around it." Apparently, STMicro did not share NVIDIA's sentiment, for the KYRO II houses only 3 million more transistors than its predecessor. Moreover, the 175/175MHz core/memory frequency did not exactly bowl us over. As we said though, the board and specifications are deceptively simple.

    The KYRO's memory controller is not DDR RAM-capable, so the board ships with 64MB of 175MHz SDRAM. If we calculate the 3D Prophet 4500's memory bandwidth, we come up with the following:

    175MHz * (128 bit bus / 8 = 16 bytes) = 2,800MB/s

    Compared with the 7.3GB/s that the GeForce3 possesses, the KYRO II looks to be at an extreme disadvantage. On the contrary though, Hercules estimates that the KYRO II is up to three times more efficient with its usage of the available memory bandwidth, making it much more competitive with speedy DDR solutions. Granted the level of efficiency is highly dependant on the depth complexity of a given application - that is, if a given 3D application involves lots of rendered objects on top of each other, KYRO II becomes super-efficient since those objects that are not visible never get rendered, while the traditional accelerator must render all of the objects and throw out the polygons that are not viewed.

    If we once again assume that the KYRO II is three times more efficient than typical rendering pipelines, we could take the theoretical fillrate of 350Mpixels/s and multiply that number by three, giving us 1.05Gpixels/s effective fillrate - an impressive figure indeed.

    Each of the two pixel pipelines on the KYRO II is equipped with a single texture unit - a seemingly minimalist approach when compared to the quad-pipeline/dual-texture unit design of the GeForce 2 and 3. However, another feature that carried over from the original KYRO design is the ability to apply up to 8 textures per pass without having to resend geometric data back across the memory bus during instances of multi-texturing.

    Like the KYRO before it, the KYRO II offers support for Environmental Bump Mapping, DOT3 bump mapping, texture compression and anti-aliasing - all standard features for a modern accelerator. Unfortunately, the KYRO II offers no T&L support or any semblance of support for the features introduced by DirectX 8. We have found hardwired T&L to be a more or less unremarkable feature, but the lack of DirectX support (which can be found in ATI's affordable 32MB RADEON) raises our concerns about the shelf life of such a product.





    Copyright © 2002 INT Media Group, Incorporated. All Rights Reserved. About INT Media Group | Press Releases | Privacy Policy | Career Opportunities