nVidia GeForce4 MX 440 in the test: Inno3D, Gainward, Leadtek and Prolink in comparison

nVidia GeForce4 MX 440 in the test: Inno3D, Gainward, Leadtek and Prolink in comparison

The chip in detail

We also have the GeForce4MX a little more detailed in our Geforce4 preview illuminated. For a more detailed explanation of the new features of the chip and infrastructure, please stop by, here are the most important things in brief:

Features of the GeForce4 MX series

  • nView Display Technology
  • Lightspeed Memory Architecture (LMA) II
  • Accuview Antialiasing
  • DirectX and S3TC TextureCompression
  • High-Performance 2D Rendering Engine
  • Integrated Hardware T&L Engine
  • Cube Environment Mapping
  • Video Processing Engine (VPE)
  • NVIDIA Shading Rasterizer (NSR)
  • DirectX 8.1 support (without pixel and vertex shaders!)
  • OpenGL 1.3 support

At first glance, it looks quite a complete list of features, but on closer inspection (and this is also appropriate when buying a graphics card!) It is noticeable that neither a vertex nor a pixel shader, as was already to be found on the GeForce3, is available is. Strictly speaking, despite its name, the GeForce4 MX is made up of elements from the GeForce2 that have been expanded a little here and there. These extensions mainly include the improved dual monitor technology called nView, thanks to which it is finally possible to control two output devices separately from each other (Matrox and ATi have been offering this for a long time), as well as the integrated LMA-II, which protects the im literally serves precious memory bandwidth.

With the elimination of the rendering unit of the bigger brother and namesake of the series, fell Unfortunately, the loop-back option is also gone, with which the chip can control its two texture units twice in succession for a maximum of four texture layers without having to write the half-finished image to memory in between. On the one hand, this worsens the quality below 16-bit color depth and, on the other hand, a large part of the bandwidth saved by LMA-II is lost again through these unnecessary accesses. It is also possible that LMA-II was only used in the GeForce4MX to compensate for this disadvantage of the lack of loop back.

GeForce4 MX in comparison
chip GeForce4 MX440 GeForce2Ti GeForce3Ti200 GeForce2MX400 architecture NV17 NV15 NV20 NV11 core clock speed 270MHz 250MHz 175MHz 200MHz memory clock 200MHz 200MHz 200MHz 166MHz DDR connection 128-128bit -DDR 128Bit DDR 128Bit SDR memory bandwidth 6.4 GB/s 6.4 GB/s 6.4 GB/s 2.7 GB/s T&L performance 34.37 MT/s 31.25 MT/s 43, 75 MT/s 25 MT/s Rendering Pipelines 2 4 4 2 Max. Pixel fill rate 0.54 GPix/s 1.0 GPix/s 0.7 GPix/s 0.4 GPix/s Max. Texel fill rate 1.08 GTex/s 2.0 GTex/s 1.4 GTex/s 0.8 GTex/s FSAA technology Multisampling Supersampling Multisampling Supersampling MultiMonitor Support nView - - TwinView TnL unit NSR NSR Pixel-/Vertexshader NSR Other VPE, LMA-II - LMA -

As you can see, the GeForce4MX seems to be a complete DirectX-7 T&L chip. The disadvantage of only two rendering pipelines is tried to counteract simply by using a higher chip clock. Otherwise, with nView, the VPE video processing engine and the high memory bandwidth, it offers sufficient theoretical plus points to be drawn into the shortlist.

Only for gamers who want to keep their graphics card a little longer but the lack of support for DirectX 8 pixel and vertex shaders would be a knockout argument against the GeForce4MX from the start.

On the next page: Test system