AOpen AX63-Pro Slot 1 Motherboard
Reviewed by Vijay Anand (4/8/99)

<Introduction><The Good><The Test><The Bad><Conclusion>
<The PC133 Test><CPU-Zilla's Conclusion><Rating>

Back to top

Introduction

We all know who Acer is, don't we? It is one of the most successful PC companies of Taiwan, which now is an International Corporation that sell PCs mostly made from their own components. They also manufacture lots of other peripherals and are the 2nd largest motherboard manufacturer in the world. Now some of you might ask what Acer has got to do with an AOpen motherboard. Since Acer is such a large company, they have a sister-company called AOpen, which stands for Acer-Open division (now you know why it's called AOpen?). This division produces far better products than Acer itself and also does its own research for new innovations (not that Acer isn't good but AOpen still does it better). Over time, AOpen has built itself a nice reputation of the most stable and well performing motherboards in the industry. Have a look at Jackie's trip to Taiwan, where he visited the AOpen factory (lots of interesting pictures await you!).

The board I'm reviewing here is an AX63-Pro, which is a new version of the AX63. The Pro version is now using VIA's latest chipset, Apollo PC133. What makes this chipset outstanding is that it officially support 133Mhz FSB, meaning the chipset is equipped with proper dividers to bring down AGP-bus to 66Mhz, PCI-bus to 33Mhz and any other minority buses to proper official speeds as compared to the Intel-BX at 133Mhz FSB. UDMA-66 is also supported with the 2 IDE connectors. The whole board is based upon the AOpen AX6BC-Pro (and a bit of the older AX63) as physically all the layouts is the same and so are the offered features with only the chipsets that have been changed.

The contents of the well known AOpen box include the following: 1 AX63-Pro motherboard, 1 packet of FDD & HDD cable, 1 UDMA-66 cable, 1 Drivers+Utility-CD, 1 Quick-Start manual.

Back to top

The Good

This board is equipped similarly like the AX6BC-Pro with 5 PCI, 2 ISA & 1 AGP-slot which, I suppose what most of us are looking for in a typical motherboard with a 6-PCI motherboard as a bonus (like the Asus P3B-F). Memory-wise, it has an adequate 3-DIMM slots to expand up to 768MB SDRAM (this Apollo PC133 supports up to 1.25GB SDRAM). The Apollo PC133 chipset supports many unusual memory standards like VCM/HSDRAM Support. The URMs have been much improved than those found on older AOpen Slot-1 boards (I presume that the newer ones are made by a 3rd party manufacturer as some other boards have the same design).

The board is tad bigger and spacious than many other BX motherboards of its category and the result is that most connectors are easily accessible without any hindrance. Examples are the HDD and Power connectors, which are located in front of the DIMM slots. In this board, there is an adequate gap between the DIMM slots and the above mentioned connectors but in many other boards those same connectors are nearly flushed against the DIMM slots, causing inconvenience to use the 1st DIMM slot. Other jumpers and connectors on board are the IrDA, WOL, WOM, FAN (2), CMOS-Clear jumper, KB/Mouse power-on jumpers, AGP-Ratio and our usual casing connectors. An SB-Link connector is also available for some PCI cards such as the Yamaha-192XG WaveForce that gives you legacy DOS sound support. Cards such as the Creative-Live do not use this connector. The ATX back-panel is colour-coded according to the PC99 specification for easy identification and plugging in of connectors. The VIA Apollo PC133's Northbridge has a gold-coloured heatsink with the words VIA PC133 and VIA's logo. I'm not sure if the heatsink is really gold-plated or just gold-coloured.

I find it a bit odd but this board has 3 different jumper settings for 3 different FSB ranges. You'll then have to enter the BIOS to select the exact FSB you require. The 3 ranges of FSB are 66 to 83Mhz, 90 to 124Mhz and 124 to 155Mhz. Alternatively, you still have the Auto selection jumper setting too. The following is a tiny part I took out from the PDF manual for reference:

Anyway here are all the FSB settings available: 66, 75, 78, 81, 83, 90, 95, 100, 105, 110, 112, 113.5, 115, 117, 118.5, 120, 124, 126, 133, 135, 137, 138.5, 140, 142, 144, 150, 155MHz, a total of 27 FSB settings! So the overclocking prospects are very good with this board coupled with the 0.1 or 0.2V increase of Vcore voltage in the BIOS. I've included here, the info from the manual:

Remember, the VIA Apollo PC133 has an Asynchronous memory bus speed, hence you can use an FSB of 100mhz and still use PC66 rams at 66Mhz or use an FSB of 66Mhz and use PC100 rams running at 100mhz. Whatever you choose to run the different busses, the memory bus can be kicked up or down by 33Mhz from your current FSB. This is set in the Bios under the Chipset-section. Here's another snip from the PDF manual for a clearer picture of how this feature works:

AOpen motherboards have a unique feature in which it allows you to save your Bios settings in EEPROM. This motherboard can also sustain the BIOS information in CMOS without the Lithium battery (but it's provided anyway) and use your power-supply instead. So long the power is connected to your motherboard, it will sustain the information. Now, since AOpen does provide the conventional Battery just in case, you don't have to worry about unplugging the power supply. If you saved your BIOS settings in EEPROM, and should you change any settings latter, or should the setting get erased, you can always load it back from the EEPROM. Very nifty and good for frequent fiddlers if you ask me :)

Suspend to Disk feature was first introduced by AOpen and it's still available in most of their current motherboards. There are 2 methods to implement it. The APM method is aimed at Win95 users and the ACPI method is only for Win98. Below are more details of each, taken from the manual:

APM Suspend To Hard Drive "Immediately" turns on system and goes back to the original screen before power down. You can resume your original work directly from hard disk without go through the Win95 booting process and run your application again. Suspend to Hard Drive saves your current work (system status, memory image) into hard disk. Note that you have to use VESA compatible PCI VGA, Sound Blaster compatible sound card with APM driver, for Suspend to Hard Drive to work properly.

ACPI Suspend to Hard Drive The conventional "Suspend to Hard Drive" function is a private design of AOpen motherboards, the requirement to run it is a BIOS revision and correct chipset that supports this function. However, nowadays "Suspend to Hard Drive" has been become widely known as the "S4 - Hibernation" state defined in the ACPI specification; in this case not only the BIOS itself but also the operation system (i.e. Windows98) and even the display card are involved in the execution of this function.

The correct way to go about implementing this feature is clearly described in PDF-format manual. The APM method was much harder to set than the ACPI method (this is similar to setting the STR function). Notice that in the ACPI paragraph above, Suspend to HDD is known as S4 and the Suspend To Ram is known as S3 (which I've explained in my i810 review and DFI PW65-D review). This obviously means S4 conserves even more than S3, hence it takes longer to recover to the last used state (also because HDD is slower than RAM).

The given CD comes with these software: Norton Anti-Virus, Norton CrashGuard, Hardware Monitoring Utility, Product Quality Assurance (PQA) tester, DocuCom Reader, Advanced Desktop Management (ADM), AOchip, Manual and FAQ. This is the AOpen Hardware Monitoring Utility:

The Test

Wintune 98 Results

ZD WinBench/WinStone-99, Norton SI & Sisoft Sandra-99

3DMark99-MAX

3DMark99-MAX Results
3DMark Results	CPU Geometry Speed	Rasterizer Score
The BX-based AX6BC-pro wins hands down on all the tests in 3D-Mark99 over the AX63-Pro using the Apollo PC-133...	This situation continues throughout all the tests.	No comments.
Game-1	Game-2	Fill-Rate
The results using a new Apollo PC133 chipset board is like downgrading your other system components!	Same comments.	This depends more on the video-card and cpu-speed.
Fill-Rate with Multi-texturing	16mb Texture Rendering Speed	Bump-Mapping Emboss, 3-pass
Simialr to previous comments.	The results are very similar for all other texture sizes (in terms of differnce between the both motherboards).	The results are very similar for Bump-Mapping Emboss (in terms of difference between the both motherboards).

On every test or benchmark, the results point out-right towards the BX chipset. As I expected, VIA's weakness in memory bandwidth and AGP implementation which brings down the scores quite a bit. What's the use of a PC133 chipset to utilise 133Mhz RAMs when the overall speed of the system is far below a system running an Intel BX chipset?? Well, if VIA's marketing is right, it can entice some OEM PC makers and some other PC vendors to implement a VIA Apollo PC133 chipset based motherboard just so that the vendor can add PC-133 compliant as his selling point!

Back to top

The Bad

The number-one bad point would be the speed of the system due to the use of this VIA Apollo PC133 board over the proven i440-BX chipset. In every category, it is slower and especially in games because the design of the AGP-bus or the accompanying drivers isn't as refined as Intel's. The results speak for themselves. When I started the Winbench and Winstone programs, there was a long pause, a pause I've never encountered. I thought the system hung but it was just a long delay. The same phenomenon happened in some other minor occasions but the board and system still worked flawlessly. Even maneuvering through the BIOS looks slow!

Given that we have 5 PCI slots and 2 ISA slots, only 4 PCI slots and 1 ISA slot can truly use full-length cards. The reason I use the word 'truly' is because there are some connectors located in front of the 5th PCI and 2nd ISA slots and depending whether you connect them or not, you'll have access to use full length cards on them. The 5th PCI slot could be blocked by a Casing-Fan connector and the 2nd ISA slot can be blocked by some casing-connectors.

The LED connectors & others are bunched into one corner which is not easy to pull out & put back if the board & other parts are all fixed inside the case. Although it's labeled, it's hard to see them within the case in one small corner.

The floppy connector is placed right at the back of the Slot-1, the traditional place where AOpen places the floppy and power-supply connectors but at least the big power connector is moved up front but the floppy connector still remains there. I'm really curious what made AOpen put the floppy connector in such an awkward position. I know that floppy connector is now closer to the I/O controller but still, this isn't a good excuse to position it behind the Slot-1 connector unlike many other boards. It's a possible heat accumulating area because of the floppy cable going over the CPU. You may try routing the floppy cable below the motherboard.

If this is a board aimed at overclockers, what happened to the 3rd fan-connector? I don't think this one should have been skipped. A thermal-header, thermistor and some proper s/w would also have been real useful like the MSI counterpart and this has made it a better value for money as both boards are priced similarly. A bigger complain other than the 3rd fan-connector would be the lack of a proper manual in booklet form! It's now provided in a PDF-format on the AOpen CD. Though setting up the board is easy with the use of the quick-start manual, it's very irritating to find out of the motherboard features, its offerings and many meanings of the BIOS menu from the CD. A normal motherboard booklet like the ones bundled with older AOpen motherboards would have been very useful. I do hope AOpen realise my constant complain about the over-concised manual as I believe it is quite important to portray a finished product for a company of its size and calibre, and considering how much effort they have put into designing the motherboard.

Furthermore, there is no AGP-4X support, STR functions. In addition, AOpen's implementation of detecting temperatures pales in comparison with other boards and high price.

Back to top

Conclusion

I've no idea who to recommend this board to buy because it's priced within $5 difference of the AX6BC-Pro!! Though this board works well, performance is slacking and minor pauses may occur time to time. Maybe a newer BIOS or drivers could iron some kinks out. But AOpen should price this board effectively lower than its higher classed AX6BC-Pro to at least win over people who want the newer/flexible features found in the Apollo-Pro chipset. AOpen can say that they've included the UDMA-66 cable but that doesn't warrant a price increase such that both VIA and Intel models are priced similarly. As for now PC-133 support and UDMA-66 are the only major points the AX63-Pro has over its AX6BC-Pro counterpart, but these 2 points are rather weak because of the AX63-Pro's usage of VIA's Apollo PC133 chipset.

Back to top

The PC-133 Test
Updated by CPU-Zilla (14/8/99)

Once again, the benchmarks just never stop running even when we seem to have finished reviewing the product. We managed to take the board to a PC-133 test using the Pentium II-350 SL37F which can be easily clocked up to a whopping 133MHz bus speed. Since my test system is different from that of Vijay's the results I'm about to present here are a little different in terms of the benchmarked scores, but you should be able to make a pretty good comparison.

Test Configuration
Processor(s):	Pentium II-350 SL37F Philippines
Motherboard(s):	AOpen AX63-Pro MS-6163 Pro
RAM:	1 - 64MB Micron PC133 SDRAM DIMM 1 - 128MB Micron PC133 SDRAM DIMM (48LC8M8A2 -75B)
Hard Drive(s):	Quantum Fireball EX 6.4GB
Video Card(s):	Microstar NVidia Vanta (16MB) M8807
Bus Master Drivers:	VIA Bus Mastering Drivers (provided on CD) Mircosoft Windows 98 (2nd Edition) generic VIA Bus Mastering drivers Latest VIA Bus Mastering Drivers from VIA's website (4-in-1 drivers v.4.11)
Video Drivers:	nVidia Detonator drivers build 2.08
Operation System(s):	Windows 98 (2nd Edition)

Firstly, let me say that overclocking the Pentium II-350 was very easy with this board. There's little to worry about the AGP and PCI clock speeds as they should be adjusted to the normal value of 66MHz and 33MHz respectively. Unlike other BX motherboards that can perform 133MHz, the AGP clock ratio used was still 2/3 of the bus speed which is actually running at about 88.66MHz. On the other hand the PCI clock would have been 44MHz instead! However, I do know the newer motherboards have proper PCI clock dividers of 1/4 (1/4 x 133MHz = 33MHz). Anyway, here are more benchmark results (drumroll)....

ZD Winstone 99 Results

Benchmark	133x3.5 (467MHz) @CAS2	142x3.5 (497MHz) @CAS3
AX63Pro (1)	23.2	23.7
AX63Pro (2)	23.2	23.7
AX63Pro (3)	23.5	24.0
MS-6163 Pro	23.5	-

ZD CPUmark 99 Results

Benchmark	133x3.5 (467MHz) @CAS2	142x3.5 (497MHz) @CAS3	144x3.5 (504MHz) @CAS3
AX63Pro (1)	36.5	38.6	39.2
AX63Pro (2)	36.5	38.6	39.3
AX63Pro (3)	36.5	38.6	39.3
MS-6163 Pro	36.9	-	-

ZD FPU Winmark Results

Benchmark	133x3.5 (467MHz) @CAS2	142x3.5 (497MHz) @CAS3	144x3.5 (504MHz) @CAS3
AX63Pro (1)	2400	2550	2590
AX63Pro (2)	2400	2550	2590
AX63Pro (3)	2400	2550	2590
MS-6163 Pro	2400	-	-

ZD Business Disk Winmark 99 Results

Benchmark	133x3.5 (467MHz) @CAS2	142x3.5 (497MHz) @CAS3	144x3.5 (504MHz) @CAS3
AX63Pro (1)	3520	3250	3510
AX63Pro (2)	2520	2750	2720
AX63Pro (3)	2720	2720	2730
MS-6163 Pro	2740	-	-

ZD High-End Disk Winmark 99 Results

Benchmark	133x3.5 (467MHz) @CAS2	142x3.5 (497MHz) @CAS3	144x3.5 (504MHz) @CAS3
AX63Pro (1)	10900	11900	12500
AX63Pro (2)	9850	10100	10000
AX63Pro (3)	9820	9970	9970
MS-6163 Pro	9990	-	-

Note : (1) using generic bus mastering drivers provided in Windows 98 Second Edition
(2) using VIA bus mastering drivers provided in the CD-ROM
(3) using latest VIA bus mastering drivers (dated 14 August 1999). VIA 4-in-1 v4.11

As we can see, the AOpen AX63 Pro performed pretty well at 133MHz as compared to an Intel BX chipset based motherboard (e.g. MS-6163 Pro). The CPU Mark99 was a little slower on the VIA chipset, but both MS-6163 Pro and AX63Pro had the same score for FPU and Winstone 99 benchmarks. However, the MS-6163 Pro seems to have a slightly better disk performance as compared to the AX63Pro. Note that the hard drive used in the benchmarks was only a UDMA33 hard disk. The Winstone and Disk Winmarks performance would have looked better on the AX63Pro if a UDMA66 hard disk was used.

The generic bus mastering drivers installed by Windows 98 Second Edition performed surprisingly better than drivers written by VIA as shown in the AX63Pro Disk Benchmarks. I didn't believe the scores, so I went to download the latest version of the 4-in-1 drivers and installed it. What we see was a very slight performance drop using the new drivers. This is indeed very puzzling to me. Somehow, VIA still have yet to properly tune their bus mastering drivers. Anyway, you should note that the generic drivers provided in Windows 98 Second Edition may not be the most stable driver around, although it worked fine during all my benchmarks. In addition, the strange behaviour of these drivers seems to give varying disk benchmark results as you can see from the inconsistencies tabulated above. The scores fluctuate quite a bit and I cannot really pinpoint the exact problem, so I used the best scores out of three different separate runs.

Using Micron's latest PC133 SDRAMs, I was surprised to find that these RAMs performed exceptionally well at 133MHz, even when the CAS Latency was set to 2. However, when I tried to use the same settings on the MS-6163 Pro, the system showed some instability, especially when running the Winstone benchmarks. I guess this is the factor that distinguishes the AX63Pro from the MS-6163 Pro.

Back to top

---

CPU-Zilla's Conclusion

Overall, I find the AX63Pro to be a wonderful board for 133MHz bus speeds. However, due to the non-existent nature of CPUs with 133MHz specs, I doubt this board would ever get very popular. Still, we should not underestimate Intel, as they may just introduce 133MHz CPUs in a desperate bid to divert the market's attention away from AMD's forthcoming K7 processors. I think we will just have to wait and see. But if you are dying to get your hands on a PC133 chipset with UDMA66 capability, then I guess this board is something that you could consider. Otherwise, your existing BX motherboard would have been more than suffice your computing needs, since we see very little difference between the Apollo Pro133 and Intel 440BX chipsets at 133MHz operation.

Back to top

---

MOTHERBOARD RATING

This Product is provided courtesy of,

Copyright © 1998 by Singapore Hardware Zone. All rights reserved.

None of the above shall be reproduced, copied and/or
modified without the permission of the WebMaster.