BBS | Home | Mainboards | CPU | Networking | Memory | Video | Tech Advice
What's new | Cartoons | email me | About the author | Links | Advertising Info

ASUS K7V
The Next Generation
(Review by MS)

Performance

Overall the performance of the K7V is more than satisfactory. In business applications using 128 MB system RAM it still doesn’t reach up to the high scores set by the MSI K7Pro. However, in a 128 MB system memory configuration, the HDD I/O times are the crucial performance limitation parameters, at least in ContentCreation Winstone2000. Therefore it is not surprising that regardless of whether HSDRAM or VCM was used in either 3:3 or 4:3 setting, the results were statistically indistinguishable from each other.

This changes quite a bit as soon as the system RAM is increased to 256 MB. In addition, overclocking the bus speed shows the impact of the memory configuration on the performance in a multitasking environment.

ContentCreation Winstone 2000 scores depending on clock speed and memory configuration. The lower 4 columns show the scores at 800 MHz using 128 MB of either VCM (V) or HSDRAM (H) at 3:3 and 4:3 bus frequency. All other results shown were obtained using 256 MB system memory to reduce HDD performance issues. The numbers indicate the clockspeed at which the data were obtained (mean of 3 runs).

What appears to be counterintuitive is the fact that increasing the latency in form of the DRAM read latch time increases the performance. However, this is not as much a conundrum as it appears at first glance. In most high performance systems, memory performance is no longer limited by the data transfer capability of the Bus itself but rather by the ratio of page hit to page miss. The latter increases with the bus / memory frequency and the amount of memory / number of applications. Page thrashing can be held responsible, particularly in 3D applications but also the overall speed of the memory bus plays a role. In other words, by increasing the latency of the DRAM read latch, the system gets more time to correctly place the data out (DIMM) into the data valid window (chipset), thus reducing the read errors. In business applications the performance gain is negligible since most processing is handled from the cache anyway but what about 3D performance?

There are some rumors ....

=> Quake3 Arena =>

Top page | specs | contents / manual | layout | connectors | BIOS | set up / stability | ContentCreation | 3D performance | conclusion |
BBS | Home | Mainboards | CPU | Networking | Memory | Video | Tech Advice
General disclaimer: This page only reflects the author's personal opinion and assumes no responsibility whatsoever regarding any of the contents or any damages that may occur explicitly or implicitly from reading the contents of this site. All names and trademarks mentioned in this review are the exclusive property of the respective parent companies.

All contents of this site are protected by international copyright laws. Reproduction of the contents even in parts is not allowed except after written permission by the author and referral to this site.
Copyright 2000 LostCircuits

Click Here!