3D, Gaming and Video Playback
First let’s look at Mad Onion’s 3D Mark 2001. It’s a fairly up to date benchmark that looks mostly at 3D gaming, but also 3D performance by itself. Not only does it give an overall score, it also gives individual scores for different types of 3D gaming at two different resolutions. This can be quite helpful in looking at differences in not only CPU speed but also by differences in the CPU’s architecture.
3D Mark 2001 performance gain with iP4 system
Since the percentage difference in each test does vary significantly, it appears the not only are the scores higher with the iP4 due to it’s higher CPU speed, but the difference in the architecture of the two CPU’s is quite different. Otherwise we would not see the range go from a low of 17% to a high of 97% for the different tests run.
It is interesting to see the overall score come in at 43% in favor of the iP4 since it is over 100% faster in pure CPU MHz. Just remember the CPU is only one part of the complete system. But just as interesting is to see that this test did not come close to the difference reported by the Winstone test. My guess would be that more than anything it’s due to 3D Mark 2001 results being limited by other hardware, such as the video card, 3D Mark 2001 not using some of the newer features of the iP4, or because the iP4 really shines is other types of multimedia applications (such as the more business based multimedia applications used in Content Creation Winstone).
Video and Gaming performance gains with iP4 system
One application Intel states the P4 series of CPU’s really excels is in video playback and manipulation – here that is confirmed with the Video 200 score of 90% faster then the P3 system. That is a very significant difference and with more and more video type applications being used it’s also a very important factor in deciding what CPU to buy. I’d also guess it’s a good indication of what the high bandwidth of DDR memory and a large L2 cache can do for overall system performance in some applications.
I find the two game benchmarks quite interesting – even though both are reporting FPS (Frames Per Second) they each show quite different gains. Quake is an older benchmark and quite a good indicator of pure CPU and memory speeds, Unreal is a newer program that does not seem to be quite as sensitive in showing those factors. So Quake gives the type of score I’d suspect, almost twice as fast. While the 42% gain shown for Unreal is quite significant, it just goes to show that not all applications react the same to changes in a system… or is it due to a limitation of other hardware not allowing the CPU to be used to it’s full potential?
Be the first to discuss this article!