-- Singapore Hardware Zone -- Reviews -- Motherboards -- Microstar MS-6167 Ver.1

Introduction

The MS-6167 is one of the first motherboard based on the AMD Irongate chipset for their latest 7th generation processor K7, which is codenamed Athlon. When Hardware Zone first got this board, we were very excited to put it to the test to see how well it fared against AMD's rival (need I spell out the name?). The first benchmarks we performed was on Microstar's engineering sample which was loaned to us. Needless to say, the build looked the same and its features were also similar. Nothing more and nothing less.

I checked the motherboard layout with that of the AMD's reference board and it did look extremely similar. If I am not wrong, the board was designed according to the AMD's reference design. Of course, I did not have the reference board from AMD, but from large close-ups of AMD's reference design (that you can find almost anywhere in the web), there is little doubt. Even the placement of the DIMM slots, ATX power connector are the same. No wonder Microstar could roll out the board in such a short time.

Opening the attractively packaged motherboard, you are at once greeted with the the manual, the guarantee card, a CD-ROM containing the drivers, a packet containing FDD/HDD cables, an 80-conductor UDMA/66 cable, heatsink retention clips, a Microstar sticker (for those sticker freaks) and the motherboard itself. I would have hyper-ventilated if I were to repeat that without stopping as I am no Ace Ventura. Besides, I don't speak with my buttocks (just kidding folks!).

Anyway, there's also a piece of notice in the box which contains a list of recommended power supplies for the motherboard. I guess the board is pretty fussy with poorly made power supply units, and a good regulated unit is required and definitely a must, if you want your system to function well, and problem free. In the list, it also mentioned some of the tested and approved AGP cards, but as we all know by now, the board works fine with almost all of the popular graphic cards available in the market. Let's face it, no sane nut will use an outdated Cirrus Logic 4MB AGP card on their state of the art Athlon system. Anyway, here's a list of power supplies recommended by the folks at Microstar.

Turbo-Cool 300ATX
American Media CWT-300ATX
Emacs AP2-5300F-RV2
Astec SA302-3515-980
Enlight HPC-250G2, A0-01

Sparkle FSP250-61GN
Enhance ATX-1125B
FSF Group Inc FSP 250-61GN
DELTA ELECT INC 200PB-103A
POWERMAN FSP300-60GT

Anyway, I suppose most of you should not be alarmed if you don't have the right power supply. A good thing to note is that the power supply I used in the test was not in the list at all. In fact, it was one of those power supplies which I ripped out from my old 1.5 year old ATX casing. I guess you should make sure the power supply comes with some of those certifications like TUV and CE.

The Good

I guess the most attractive part about this motherboard lies in its simplicity. There are very little jumpers or BIOS settings that you need to change on the motherboard to get your Athlon up and running. There are no voltage settings, multiplier settings, and no FSB settings too. Just plug and play. Of course, most tweakers would not like the lack of control. However, if you do not already know, overclocking an Athlon processor isn't that simple, considering the multiplier settings are locked on the CPU board. You will need to open the CPU case (warranty will be void) and do some resoldering to change to the desired multiplier. Such modifications are rather tricky and risky, if you ask me. I doubt many will take the risk. But, I'm sure someone will come out with an ingenious way of overclocking the Athlon.

The Southbridge chipset, which is codenamed Viper, can support UDMA/66 IDE devices. This is really a plus for those who already have UDMA/66 hard disk drives, or are planning to purchase one. Although there are not much improvement in the hard disk performance as compared to UDMA/33, it is still comforting to know that your motherboard is prepared for fast UDMA/66 transfers, when the right hard drives come along.

Some users have wondered what kind of RAM should a 200MHz FSB system should use. Well, the 200MHz EV6 bus does not apply to the RAM. The chipset still communicates with the SDRAM at 100MHz. This allows you to re-use your old PC-100 SDRAM. Good for those unwilling to spend the extra buck on RAMs since the price has skyrocketed lately.

The board also comes with 4 USB ports, something that most motherboards do no have, as of now. There's a 10-pin header located near the front panel for 2 additional USB ports. This is really convenient for people using portable USB devices. I think locating your USB ports on the front (or side) panel of your system is a good idea. How many of you actually dread reaching the back of your PC to plug your USB connector? As my system is located under the table, reaching below to plug in any USB device can sometimes be a back-aching experience. However, we still need casings that has built-in USB ports on the front panel, something that we only see in other well-branded PCs.

Besides that, the board generously comes with 3 fan headers. Since the Athlon is known to be very hot, I think that it is handy to have 3 fan connectors. You may never know when you'll need all of them. In addition, it is neat to plug directly on the motherboard as you do not need extra cabling to convert your power supply's power connectors to fit your fan. Additional cablings make your system look messy, and can sometimes be a potential air flow barrier.

Running the system through all kinds of test, I must say that this is one of the most stable motherboards I have ever encountered. I did not run into any GPF errors or any blue screen of death. There were no errors, and I am very very impressed indeed. I expected some problems in Windows NT, and again, none whatsoever. I churned RC5 blocks with the system for a few days, non-stop. Again, it was perfect. Don't you just love it?

In the tests, I compared a K7-500 with the P!!!-500, with both running on Microstar motherboards. The P!!! ran on the MS-6163Pro, while the K7 was benchmarked on the MS-6167. The same RAM, hard drive and video card was used, unless otherwise stated. Of course, I couldn't make use of the same drivers and motherboards, otherwise, it wouldn't make the K7 that special after all. So, don't go around accusing me of being biased. We should look at the benchmarks based on the motherboards and processors (system as a whole), and not just one component alone. Also, to make things a little fair, I used the Promise Ultra66 IDE controller to enable UDMA/66 on the hard disk as it really matters in benchmarks like Winstone99 and Disk Benchmarks.

Benchmark	MS-6163Pro			MS-6167
Benchmark	P!!!-500	P!!!-600	P!!!-600B	K7-500	K7-650
CPU mark 99	38.4	44.3	46.4	47.7	59.7
FPU Winmark	2550	3040	3070	2710	3530
Business Winstone 99	24.5	26.6	26.8	24.5	27.5

It is very evident that the K7 is far more superior than the Pentium !!!. No wonder Intel is panicking. The K7-500 CPUmark99 scores were surprisingly faster than the P!!!-600 (currently the fastest Intel CPU around). Even the 133MHz FSB version (as denoted with a B) was slower than the K7-500 in the CPUmark99 tests. Unfortunately, the FPU Winmark was not faster than the P!!!-600, but if you compare MHz to MHz, the K7 beats the P!!! flat. I've placed the K7-650 results there as a reference and it should not be compared to the rest (don't call me bias again.. hahaha). I guess most of you must be wondering how fast the 650MHz version is, so here it is. Also note that the 600B CPU was benchmarked on the 6163Pro which was not designed for 133MHz operation. Results may vary when it is benchmarked on the soon to be released i820 (a.k.a. Camino) motherboard. Comparing the Winstone 99 results, you can see very little difference between the K7 and P!!! even though the K7 was faster. This may be due to the poor disk performance, as discussed in the next benchmark.

Disk Performance Benchmarks (Windows 98)

Benchmark	MS-6163Pro		MS-6167
Benchmark	P!!!-500	P!!!-600	K7-500	K7-650
ZD Business Disk Winmark	4650	3840	4330	4030
ZD High-End Disk Winmark	14400	13200	10400	12700
Threadmark 2.0 (Data Transfer Rate MB/s)	8.48	8.45	8.21	-
Threadmark 2.0 (Average CPU Utilisation)	32.51%	27.76%	31.69%	-

If you compare the P!!!-500 and K7-500, it shows that the MS6167 disk performance is weaker than the Promise Ultra66 IDE controller running on the MS6163Pro. This basically explains why the Winstone99 results were on par. The disk performance pulled the score lower, making the K7 results in Winstone on par with the P!!!. This shows the weakness in AMD's Southbridge (Viper) chipset. Hopefully it will be better with motherboards incorporating VIA's southbridge chipset. Comparing the disk performance of faster CPUs, it is awkward to see poorer performance with the P!!!-600 and K7-650. I have absolutely no idea why it behaved that way, but I suppose we should just take the results with a tiny weeny pinch of salt. I did not run any disk performance tests with the MS6163Pro on board IDE controller since it is only UDMA/33. It wouldn't be fair, isn't it?

Wintune 98 1.0.39 Benchmarks (Windows 98)

Benchmark	MS-6163Pro		MS-6167
Benchmark	P!!!-500	P!!!-600	K7-500	K7-650
CPU Integer	1464.2	1758.9	1514.4	1970.0
CPU Floating Point	575.0	690.7	618.2	800.2
Video 2D	99.6	140.1	157.8	170.7
Direct 3D	94.3	169.1	168.4	168.6
Open GL	76.8	80.71	80.4	84.4
Memory	840.2	991.0	1272.7	1601.3

The Wintune 98 benchmark seems to paint a different picture. Although the K7 is still significantly faster in all the tests, the K7-500 did not beat the P!!!-600 in the CPU Integer scores. However, all the video tests ran faster on the K7. The memory benchmark was stunningly fast on the K7. Overall, the K7 is still much more superior than the P!!!. Two different benchmarks don't lie. So, how about Windows NT?

Ziff-Davis Winbench 99 and Winstone 99 v1.1 Results (Windows NT)

Benchmark	MS-6163Pro		MS-6167
	P!!!-500*	P!!!-600	K7-500*	K7-650
CPU mark 99	38.7	44.5	48.3	61.1
FPU Winmark	2560	3050	2740	3560
Business Winstone 99	31.9	35.5	32.6	37.6

Disk Performance Benchmarks (Windows NT)

Benchmark	MS-6163Pro		MS-6167
Benchmark	P!!!-500*	P!!!-600	K7-500*	K7-650
ZD Business Disk Winmark	3460	4280	4610	4030
ZD High-End Disk Winmark	9770	9910	10300	12700
Threadmark 2.0 (Data Transfer Rate MB/s)	9.6	9.53	9.62	-
Threadmark 2.0 (Average CPU Utilisation)	13.46%	8.02%	29.05%	-

Wintune 98 1.0.39 Benchmarks (Windows NT)

Benchmark	MS-6163Pro		MS-6167
Benchmark	P!!!-500*	P!!!-600	K7-500*	K7-650
CPU Integer	1467.9	1763.0	1522.1	1981.6
CPU Floating Point	581.9	698.6	624.9	812.8
Video 2D	113.7	163.5	65.8	129.2
Direct 3D	77.2	86.3	37.3	43.4
Open GL	72.9	51.4	75.6	52.6
Memory	872.6	1067.684	1339.5	1708.9

* MS-8808 (TNT2-M64) was used instead of V3-2000.

The results in Windows NT showed similar trends. However, the disk performance was better in Windows NT as compared to Windows 98. Well, at least the Winstone 99 results made a lot more sense in the Windows NT benchmarks.

Shall we look at the graphics performance?

Quake III Test 1.08
q3demo1	q3demo2

The performance in Quake III Arena is pretty much the same. This is probably due to the graphic card used in the benchmarks. It has probably reached the stage where CPU performance have no affect on its performance. However, at 640x480, the K7-500 is slightly faster than the P!!!-500, and it is on par with the P!!!-600.

Resolutions	3D Mark overall scores	Fill Rate performance (MTexels/s)
640x480
800x600
1024x768

In the 3D Mark scores, the K7-500 scored higher than the P!!!-500. In certain tests, the K7-500 was also faster than the P!!!-600. This shows the K7's superiority over the Pentium !!!. It is a waste that the chipset only supported AGP 2x transfers, otherwise it would have been great to run some benchmarks using some of the latest graphic cards like GeForce256, or even the Matrox G400Max. Anyway, I did not get the chance to run graphic benchmarks using other video cards to get a better picture of the motherboard. However, isolated tests revealed certain jerkiness during the demo and Game 2 benchmark. It looked as though frames were dropped. I suppose the AGP implementation is still a little buggy. I'm not sure if this jerkiness has to do with the motherboard or chipset, but I tend to believe that it is the latter's fault. Still, I hope the next generation of chipset for the Athlon would be better, and would include AGP 4x transfers.

Motherboard Specifications
Processor	AMD K7 Athlon Processors (500Mhz - 650MHz or higher)
Chipset	AMD-751 Irongate Chipset (Northbridge) 200MHz EV6 System Interface speed 100MHz SDRAM 1x/2x AGP AMD-756 Viper Chipset (Southbridge)
Cache memory	CPU Built-in 128KB L1 & 512KB L2 cache
System Memory	3 x 168-pin 3.3v PC100 SDRAM DIMM support Supports 8/16/32/64/128/256 MB DIMM Module Supports 168-pin unbuffered PC100/PC133 SDRAM (Supports ECC, 1-bit Error Code Correct function) Supports up to 768MB of memory size
PCI IDE	2 X PCI Bus Master UDMA/66 IDE ports (up to 4 ATAPI Devices) Support for PIO Mode 0-4, UDMA/33, UDMA/66 IDE & ATAPI CD-ROM
I/O Interface	1x floppy port (360KB-2.88MB) 2x serial ports 1x parallel port (SPP/EPP/ECP) PS/2 Keyboard PS/2 Mouse 4x USB 1 IrDA connector for SIR
Expansion slot	5 x PCI 32-bit slots, PCI 2.2 compliant (one PCI/ISA shared) Supports 3.3V/5V PCI bus interface 2 x ISA 16-bit slots 1x AGP (1x & 2x Mode, 66/133MHz) slot (3.3v device support)
Power Management	Power On by Soft-Power Switch Power Off by Windows^®95/98 Shut down & Soft-Power Switch ACPI (Advanced Configuration and Power Interface) feature
Form Factor	ATX Form Factor : 30.5cm(L) x 20.9cm(W) Fits into regular ATX Case (8 mounting holes) ATX Connector on Board
BIOS	2 Mbit (256KB) FLASH memory Award PCI BIOS with Green, PnP, DMI and Anti-Virus Functions LS120, ZIP, ATAPI CD-ROM, LAN, IDE #1, #2, #3, #4 bootable

Test Configuration
Processor(s):	AMD K7-500 AMD K7-650 Intel Pentium !!!-500 Intel Pentium !!!-600 (100MHz FSB)
RAM:	1 x 128MB Mitsubishi PC100 SDRAM DIMM
Hard Drive(s):	IBM Deskstar 22GXP DJNA-371350
Video Card(s):	3dfx Voodoo3 2000 16MB Microstar NVidia TNT2-M64 32MB (MS8808)
Bus Master Drivers:	AMD Bus Master Drivers 1.12 Promise Technology Ultra66 1.43
Video Drivers:	3dfx Voodoo3 drivers 1.02.13 (Win98) 3dfx Voodoo3 drivers 1.02.11 (WinNT) NVidia Reference Detonator Drivers 2.08 (Win98/NT)
DirectX version:	v.6.1a (Win98), v.3 (NT 4.0)
Operating System(s):	Windows 98 Second Edition (build 4.10.2222A) Windows NT 4.0 Workstation (Service Pack 5)

Overall Rating (Out of a maximum of 5 Star)
Installation	*****
Performance	*****
Price	***½
Overclockability	N.A.
Material Quality	********
Stability	*****
Overall Rating	********

Introduction

The Good

The Test

Conclusion