3D Tests II
X2 at 800x600 picks up differences in L2 and main memory performance. Increasing your memory module latency and running X2 has the visible desired effect of less performance, so it's nice to use.
What's been interesting to see throughout is the performance of the single channel P4 setups, compared to the dual channel setups. It's not been devastatingly slow, vindicating the performance of VIA's PT800 recently, along with the dual channel PT880.
More of the same, apologies for the unexciting graph commentary.
Who'd have known the Comanche 4 graph would have looked like that? I think we've established the performance order don't you? Comanche 4 reacts nicely to changes in memory subsystem, PAT is the biggest culprit in the performance difference between 848P and 865PE single channel. Let's wrap it up.