Viper Lair
Sponsor
Menu
Latest Stuff


MSI K8N Neo Platinum
Albatron FX5700P Turbo 128MB
Creative MuVo 2 CF Mod
X-Arcade Dual MAME Controller
Flexiglow Illuminated KB
Cooler Master Hyper 6 Heatsink
Danger Den RBX AMD Kit
Interview with Intel
Apacer Wireless Steno MB112
CoolerGiant 485W SFMA PSU
Latest Stuff
Search for lowest prices:


for 


Price Search:    for    

MSI i865 NEO2-FIS2R: Supporting Intel Hyper-Threading Technology, 800FSB and Dual Channel support, the 865PE Neo2-FIS2R is sure to to be a speedy board. Is speed enough though?

Date: October 1, 2003
Manufacturer:
Written By:
Price:
 



Benchmark Setup

CPU:

Intel Pentium IV 2.4C - 200MHz FSB

Intel Pentium IV 1.8A @ 2.4GHz -b 133MHz FSB

Motherboard:

MSI 845PE Max2

MSI 865PE Neo2

Memory: 1GB Corsair XMS TwinX PC4000 Kit (2*512MB)
Hard Drives :

40GB Seagate ST340016A, 40GB Maxtor 34098H4, 2*80GB Maxtor 6Y080M0 SATA Hard Drives

Video Card: Matrox Parhelia 128MB (200MHz/250MHz)
Operating System: Windows 2000 Pro SP3 Direct X 9
Drivers: Parhelia 1.04.00.78
Other Cards: Turtle Beach Santa Cruz, ATi TV Wonder
Cooler: Retail Pentium IV Heatsink
Case: CoolerGuys Windtunnel IV
Power Supply: RaidMax 400Watt Power Supply
Direct X Benchmarks: Unreal Tournament 2003 (HardOCP software 2.1 - CPU Test)
OpenGL Benchmarks: Jedi Knight II Jedi Outcast (time demo)
Other Benchmarks VirtualDub 1.4.10 DivX 5.03
  Truespace 4.2 DUMeter 3.03 Build 110
  TMPGEnc Plus 2.59.47.155 H2benchw
  Pi Fast 4.2  


    Both CPU's were used for testing, with the 2.4b being the only one that worked on both motherboards.  We have also taken some comparison results for Pi Fast from Hubert's Abit IS7 Max II Advance review, as well as his latest MSI Mega 651 PC review.  The memory was run at its SPD settings, and even so was running at 2.5-4-4-8 even at 133MHz on the 865PE motherboard.  The 845PE had no trouble with the RAM and happily ran it at 2-2-2-5 at 133MHz.  

   All tests were run at least three times, with the highest of the two scores being used.  If the scores were not within a 1% range the test was redone until that occurred.  The Unreal Tournament test used the HardOCP test software's CPU test and was run at 640*480.  The Jedi Knight tests were run at 1024*768 with high quality settings, as test was not GPU limited at this resolution.  Pi Fast was rung using 10000000 digits of Pi, Chudnovsky method, 1024 K FFT, and no disk memory, just as we at VL have done previously.

    For DUMeter, the host machine was the 865PE, with the CSA Gigabit network card running at a maximum of 1Gbps, using CAT5E crossover cable to connect to the other computers.  The files used for both the Network transfer tests and the real life hard drive tests totaled just over 9GB and had within the 15000 files various file sizes, from 1KB to over 1GB in size.  Our real life hard drive test consisted of copying the files from a SATA RAID0 array to the other hard drives.  The H2benchw software was run on each hard drive using the following command line from DOS, C:\h2benchw.exe -english -! (Hard drive number) -a -w filename, which provided a full read and write test of the hard drives.

    Both our DivX 5.03 and our TMPGEnc tests used the same video file, a HuffYUV encoding of 8555 frames of "The Hitchhiker's Guide to the Galaxy" (NTSC).  With the TMPGEnc tests the following settings were used, and the video was encoded with the audio as well.  With our DivX 5.03 tests the same video was encoded minus the audio at 1500kbps with the only other change being made to the default settings was that every 30 frames there was a keyframe, not 300.  Lastly Truespace was run using the same process mentioned in my previous article using the following settings, the image was then rendered to a 1600*1200 tga file.

Synthetic / 3D Rendering

    Synthetic benchmarks don't necessarily mean much to most people, but can provide a direct look at the specific areas of the system.  So we will take a quick look at Pi Fast, which, as the name suggests calculates Pi to as many digits as you want.  In our case we will calculate Pi to 10000000 places.  Lets see how the 865PE compares to the MSI 845PE, Abit IC7, and the Abit IS7.

    As we can see, the moving straight to the 865PE offers an 8% improvement with all other settings equal.  When moving to dual channel mode, we see a 14% improvement over the older 845PE chipset.  Compared to the Abit motherboards, the MSI 865PE does well with only a 1% difference with a 2.4b CPU.  Once we get to the 2.4C there is more of a difference, 3% compared to the IS7, and 4.5% difference compared to the IC7.  We can see in this first test that the 865PE does indeed have a fair amount of improvement compared to the older 845PE, which makes it faster in this test at least, how about other tests?

    Truespace is considered the beginners 3D animation program, but that doesn't mean that it doesn't create high quality images.  The following image was rendered at 1600*1200.  How did the 865PE perform in this test.

    As we can see with the results, there is really no real difference between the two different chipsets.  Even dual channel memory doesn't show much of a difference, only less than 0.5%.  When we increase the FSB to 200MHz we see a actual 6% increase in performance in single channel and a 9.5% increase with dual channel memory.  It is only with the 200MHz FSB that there is a improvement with dual channel memory, which could be explained as the memory at 133FSB is running at 178MHz.  Lets see how some games perform, specifically UT2003 and Jedi Knight II.


Shop for MSI Products.
Copyright © 2001-2004 Viper Lair. All Rights Reserved.