NVidia GeForce2 MX w/ Overclocking

By Dave "Davo" Altavilla
7/7/00

There is something strange about 3D Graphics Cards that brings out the "crazies" in people.  What amuses me the most is when people get into very emotional and heated discussions in our HotHardware Conference Room Forum.  It is as if people hold their video cards so close to their hearts, the damn things are like spouses or children to them.  While we personally don't understand this phenomenon here at HotHardware, one thing is for sure people get down right belligerent and opinionated about their 3D Graphics cards.  With that said, there is one thing neither NVidia fans or 3dfx fans can debate, NVidia has been pumping out new products with a frequency and regularity that no other player in this market can even come close to.  Frankly, we at HotHardware get a little dizzy with all the NVidia product that is available for evaluation.  This is good thing mind you.

NVidia first released the NV10, known as the GeForce and then followed suit not much more than 6 months after with the NV15 or GeForce 2 GTS.  Right around the time of the NV15 introduction, there were a few reports that NVidia had a product called the NV11 in the works.  At that point, most of the Online Media thought this product was targeted at the Mobile Market.  Indeed the NV11 or GeForce 2MX, is all that and more, with a firm stake pitched into the Business and OEM market segments as well.  Here is the run down on its features.

Specifications / Features Of The NV11 Jack of all trades.....

• 0.18 micron manufacturing process

• approximately 4 watts active power consumption

• 175MHz core clock and 166MHz memory clock

• 2 pixels per clock cycle, 4 texels per clock cycle

• 350 Mpixels/s fill rate, 700 Mtexels/s fill rate

• 20 million triangles/sec

• 8-64MB frame buffer

• 128-bit Single Data Rate (SDR) or 64-bit Double Data Rate (DDR) memory

• "TwinView"  Dual Simultaneous and Independent Display Output

• "Digital Vibrance Control" for enhance 2D/3D Color Quality

• Integrated Dual-Link TMDS transmitters

• NVIDIA Shading Rasterizer (NSR)

• High-Definition Video Processor (HDVP)

• AGP 4X with Fast Writes

• 32-bit color

• 32-bit Z/stencil buffer

• Cube environment mapping

• DirectX and S3 texture compression

• 350MHz RAMDAC

Memory bandwidth will be the limiting factor with the GeForce2 MX.  The GeForce 2 GTS and GeForce1  have an interface to 128 bit DDR or SDR memory and the MX will handle only 64 bit DDR or 128 bit SDR SGRAM or SDRAM.  We are sure it was a "real estate"/pin count issue.  Get the pin count down and drive the cost down with it.  Either way you cut it, the MX will be bandwidth limited from a memory standpoint.  It would be interesting to see if DDR would help even though it would be running on half  the bus width.  We'll have to wait and see on this one.

Also of note is that the GeForce2 MX only cranks out 4 Texels per clock cycle versus the GeForce2 GTS at 8.  Overall fillrate is still much higher than the GeForce256 however.

So, these features look a lot like a jacked up but low cost GeForce 1, right?  True, the core clock speed is up a notch to 175 from the original GeForce and the card we received was fitted with Hyundai 6ns. SDR memory but take a look closer at the "Twin View" and "Digital Vibrance" bullets. The drivers we have received to date for the NV11 (5.30 version and 5.32) don't allow you to utilize either of these to features yet.  However, we are told we'll be in possession of a version that does soon.  For now, we will just cover these features briefly and get back to them with an update when we have had the time to test the features out.

Digital Vibrance is a method of allowing the user to manipulate the actual bit stream coming from the GeForce chipset to the "Phy" layer.  Whoa... there's a new buzz for word for you huh?  In Communications/Networking Technologies, the "Phy" stands for the "physical layer" in the network hierarchy model.  This is just a fancy schmancy term for the connection to the outside world or outside of the box.  So, in short Digital Vibrance let's you play with the output of the card before it gets to your monitor and on the screen.  Specifically, it lets you adjust color saturation and vibrance to improve overall image quality.  This is not just a gamma tweak.  We saw the effect live and in action when we met with NVidia and it is much more prominent and "prettier" (for lack of a better word) than gamma adjustments.  Here is a shot to illustrate the effect.

(click for full view)

This will be an adjustable feature on the control panel and will have three settings of low, medium and high.  It will also be available on each of the independent outputs of the "Twin View" feature.  That brings us to a logical break here for that discussion.

Twin View is a feature that competes directly with Matrox's "Dual Head" technology.  Integrated Dual TMDS (transmission minimized differential signaling) transmitters (wow that's a mouthful) are on board and capable of driving two independent displays in either digital or analog mode.  You can have two digital or analog flat panels running, one CRT and one Flat Panel, two CRTs (as long as the OEM has populated the board with a second RAMDAC), you get the idea.  Twin View is a natural for the Business market and the flat panel loving folks on Wall Street are going to eat this up, just like the G400.

Another market segment that Twin View fits well into is the mobile market.  The GeForce2 MX runs on about 4 watts of power and with its ability to drive flat panels and CRTs and even simultaneously, one would think NVidia could score big with this chip at some of the major OEMs.  Of course, the official mobile version of the chip is not out yet.  But we were told it is coming.

Setup, Installation and Overclocking