Duane's Homepage : System Benchmarks

Main Menu

M-II / 6x86MX VS Pentium MMX
Final Reality Benchmarks

General Overview

The objective of synthetic benchmarking is to perform a series of test on a system and subsequently rate its performance based on the test results. Although these tests cannot be 100% accurate in representing real world performance, more often that not, they are useful as a rough indicator.

	Test Bed Disclosure Operating system: Windows 98 Build 4.10.1998 & DirectX 6.0 Benchmark programme: VNU Final Reality version 1.01 ATI Xpression+ driver version: 5.21-9CN-WEB Advanced Benchmark option used and sound disabled by default For accuracy, all benchmark tests were run 5 times I/O voltage raised from 3.38V to 3.52V for stability in 75MHz FSB tests
	Socket 7 Configuration Processor : Processor : Motherboard : Memory : 2D/3D card :	IBM 6x86MX HF Series (Part Number IBM26x86MX-DVAPR333HF) Intel Pentium with MMX Technology P55C SY059 Abit SM-5 Triton VX 512PLB Revision 1.01 64Mb 10ns Hyundai SDRAM in 2x32MB DIMM configuration ATI Xpression+ 4MB (3D RageII+DVD) P/N 113-38806-101
	2D Benchmark Results

3D Benchmark Results

Final Reality is the intellectual property of VNU Business Publications Limited.

Discussion and Conclusion

The results of the Final Reality benchmarks indicate that the floating point performance of the M-II is slower than that of a Pentium MMX processor running at the same clock speed. This conclusion can be inferred from the lower scores obtained by the M-II in the 3D benchmarks which are known to be floating point intensive.

The 3D benchmarks also show a linear increase in performance when moving to 75MHz from the 66MHz Front Side Bus (FSB) frequency. Hence, it can conclude that the limiting factor affecting the performance of the system in the 3D benchmarks is the avaliable mathematical processing power.

However, in the 2D benchmarks, the non-linear gain in system performance is highly unexpected. The M-II at 225MHz outperforms that of an M-II running at 233MHz by a 1.7% margin. This quite clearly illustrates that the bottleneck hindering overall system performance of 2D applications is not the floating point performance, but instead the Front Side Bus frequency.

[Previous]	Glossary Of Terms
[Next]	The Dream System
[Up]	Duane's Hardware Homepage
[Home]	Home Page
[Mail]	Send EMail to Duane's Homepage
[Search]	Search Duane's Homepage
[Contents]	Duane's Homepage Contents

This site is optimized for viewing at 800 x 600 x 24 bit color with Netscape Communicator 4.06 or later.

All content on this web page and its related pages is copyright © 1996-98 by Duane Bong.

Last modified on September 03, 1998