NVIDIA's nForce 420/220: It's finally here

Memory Bandwidth Performance

There are a couple of things to take away from this initial performance measurement.  According to Sandra, the nForce 420-D only offers 26MB/s more useable memory bandwidth than the VIA KT266A.  This is in spite of the fact that the 420-D has twice the theoretical memory bandwidth of the KT266A.  The explanation is obvious and not too shocking; the Athlon is limited by its FSB in that the IGP can only get 2.1GB/s of data to it at any given time so having 4.2GB/s of memory bandwidth isn't all that useful.  This is also an indicator that the nForce 220 should perform very similar to the 420.

The next point to take away from this is that by enabling the integrated video, in normal 2D usage there is only a 10 – 30MB/s memory bandwidth penalty.  This is not significant enough to make a difference in any of the office/content creation benchmarks and sure enough it didn't. 

The standings were virtually duplicated in the FP-STREAM results so we have omitted them since they didn't really add much to the review.

Cachemem is a bit more on the theoretical side when it comes to performance and as you can see here, the peak memory bandwidth of both nForce solutions is just a tad less than what the KT266A can offer.  You can look at this one of two ways; either VIA designed an incredible memory controller with the KT266A or NVIDIA's solution is very well done for their first try.

Again we see that switching from an external AGP card to the nForce's integrated GPU results in a reduction of memory bandwidth which is to be expected.  Here you can see that on the theoretical level, the reduction in bandwidth is much greater than what we saw in the more "real world" Sandra test.  Keep in mind that neither of these benchmarks are very real world in that the performance differences here don't translate into real world performance figures; it's just useful to look at.