Duh Voodoo Man's  PowerLeap PL-iP3/T Performance Review


To rigorously compare the PowerLeap rig's performance to the P3-850MHz and P3-1GHz processors, I selected a combination of synthetic CPU and video benchmarks, together with three popular 3D games commonly used for benchmarking purposes. These were:

For all tests, vsync was disabled and antialiasing was turned off.

As previously mentioned, the benchmarks were run on two different BX platforms--a Dell Dimension XPS-R and a homebuilt PC based upon an AOpen AX6BC motherboard. Additionally, the Dell machine was run in two different configurations. The initial benchmarking was done with a Voodoo5 5500 video adapter, an older card that is showing its limitations on newer 3D games and video benchmarks. Shortly after completing the initial review, I upgraded the machine to a Gainward Geforce3 Ti200, a much more powerful video card based upon what is generally considered to be the best currently available mainstream 3D rendering technology. I repeated the same series of benchmarks with this new configuration, and have updated the review to include the results. The two sets of Dell system benchmarks therefore give a good picture of the relative impact of a CPU upgrade in one PC with a very capable video subsystem versus a second with merely adequate graphic capabilities.

This second round of benchmarking on the Dell Dimension XPS-R platform also afforded another nice opportunity--to include an "original equipment" Pentium II 400MHz in the evaluation! (Many thanks to "Etherdude" at the DellTalk forums, for graciously agreeing to loan his old P2-400 to me for purposes of this testing.) Thus, on the tables and graphs that follow, you'll see a fourth entry for this CPU under the Dell/Geforce3 system results. Dimension XPS-R owners who are considering their first processor upgrade will doubtlessly find this "baseline" data of interest, because it gives an idea of the magnitude of performance improvement that they can expect to see when upgrading to one of the faster CPUs.

Basic system specs for the three machine configurations used are shown in the following table:

From the benchmark results, I also calculated an overall relative performance rating, using the P3-850 results as the reference value (i.e. relative performance = 1.00). For this calculation, I weighted the four synthetic benchmarks at 10% each and the three games at 20% each. I intentionally placed more weight on the game results because (1) they represent a "real world" use of the PC, and (2) the most common motivation for CPU upgraders is improvement of 3D gaming performance. And finally, using this relative performance ratio and the best available current price I could find for each of the three CPUs on Price Watch, I calculated a "bang-for-the-buck" ratio, again using the 850 as the base value of 1.00. So what were the results, you ask?

Previous page of article... Finally! The numbers...