| Author |
Message |
johnnymamba
| | Posted on Sunday, November 18, 2001 - 3:08 pm: | |
Hi I have a Nvidia MX 200 card and would like some settings to help me get my 3D2001 Benchmark score up..!! I am running 1GHZ Athlon 256mb pc133 ram 20 gb HDD (5400)19"p991 dell monitor. So Far I have only managed to record 1059 at 1028X768 @32 bit...... Most reviews seem to be a lot higher than this. Any thoughts..?? |
Kevin
| | Posted on Friday, November 23, 2001 - 4:16 am: | |
Sorry to be the barer of bad news but....
You can try overclocking it with an overclocking utility or coolbits found at http://www.kickassgear.com/coolbits.htm
If overclocking try it in small stages. Usually overclocking this card is quite safe but if you see any funny things such as dots then go back a bit.
Basically the MX 200 isn't a good gaming card especially compared to other Nvidia cards in a similar price range such as the MX 400 which is about twice as fast.
http://www.amd3d.com/review/mx400/q3hi.gif
http://www.hardwareoc.com/mx_files/image004.gif
Article from http://www.hardwareoc.com/mx_4.php
640*480 results
Even at the lowest resolution the MX-200 starts to fall behind. This really shows the weakness of this card , due to the constricted memory bandwidth. It also seems that increasing the RAM density has nearly no effect on the performance at this resolution. Also note that the MX400s performance is nearly that of the GTS2 32meg at this resolution , this shows the fact that the low resolution tests rely on the CPU quite a bit more that the other benchmarks
1024*768 results
At this resolution the video card is of course forced to work much harder , and again the bandwidth bottleneck rears its ugly head. The MX 200 with its anemic 1.3 gb/s of bandwidth shows roughly 50% lower perfomance of the MX 400 with its 2.7 GB/s of bandwidth. Double the bandwidth and double the performance in this case. Keep in mind bandwidth is not only a function of memory speed but also of datapath width.
Again we see only a slight increase in the MX 400 64 meg over the 32 meg performance . Could it be that increasing memory density with this chipset is simply a ploy by NVIDIA's crafty marketing dept with at best a barely measurable performance increase?
We will see !
1280*1024 results
Must I even state that the MX-200 with it's inadaquate 64bit data path can hardly even run these tests at 1240*1024. The higher the resolution the more we can see that bandwidth is all important. Though I hate to be repetitive it seems that increasing the memory density once again has next to no effect on game performance.
* I didn't test at any higher resolution cards because all of the Mx's have already dropped out of being playable even at 1280*1024 resolution.
Conlusion
Well to be straighfoward the MX200 chipset just does not make the grade for any type of gaming , the crippled 64 bit datapath is a joke , Memory bandwidth is the issue here and by cutting the datapath in half Nvidia cut the performance in half. The lower cost can simply not justify the horrible perfomance this card eeks out .
The MX 400 chipset seems to be a fine budget gaming solution , While not the best on the market ,it performs adequately. There are several companies that offer this card with faster that the stock 6ns 166Mhz RAM, Faster ram does increase bandwidth and according to our theory as bandwidth increases , the performance increase the same percentage, at least within the limits of this technology, We will study this issue further and post our results
Now as far as increasing memory density, This, we think is a waste. Simply a marketing ploy by Nvidia. The MX400 chipset showed nary an increase in frame rate with increased memory density. Remember that when you buy an MX400 , With this particular chipset more video memory does not always equal better perfomance. A card with only 32Mb of a faster kind of RAM can outperform a 64Mb card if it is equipped with slower RAM.
Kevin |
|
