Introduction
Intel won the run and is the first to ship a chipset that is running at 100 MHz system or front side bus clock. People who want to have the fastest Pentium II CPUs will now have to take advantage of this chipset, because 333 MHz is as far as Intel will let you get in a board with 440LX chipset. 350, 400 and 450 MHz will require 100 MHz FSB and so 440BX. 100 MHz FSB alone are hardly giving any performance increase over 66 MHz, as you know from my article about the performance differences between 66 to 100 MHz FSB Slot 1 systems, but 350 or 400 Mhz CPU clock will make sure that a BX system will indeed be faster than even an LX chipset with a 333 MHz CPU.
Ram Timing Negligible
You certainly remember the fact that the higher data bandwidth of main memory at 100 MHz FSB is only responsible for a mere 3% performance increase over 66 MHz FSB in most software, which is due to the L2 cache of a Pentium II CPU, that runs always at half the CPU clock, unchanged by the FSB. This may be surprising to a lot of you, because things are so different in Socket 7 systems. However, the reason for this difference is the direct relation between bus clock and L2 cache speed in Socket 7 systems. Now although it is surprising that a 50% bandwidth increase from 528 MB/s to 800 MB/s has only a little 3% impact on overall system performance, there is a good side to it as well. In the past I advised all people who wanted to get the most out of their system to tune their memory timings in the BIOS setup. This could improve performance particularly in Socket 7 systems, but even Slot 1 systems with 440LX chipset could notice a difference. This time has changed with 440BX and the 100 MHz FSB. The data bandwidth is now so high, that even if it was double it would have hardly any impact on overall system performance anymore. Thus little differences in memory timing have no influence on system performance at all anymore, if the data bandwidth is 830 MB/s or only 760 MB/s, the difference in performance is minimal. The consequence is that many motherboard manufacturers removed the memory timing options from the BIOS setup, but made sure that every SDRAM is running safely and reliably instead. Others cut down the timing adjustments to only one and the ones who still offer 4 different adjustments for memory timing must be asked if they want to impress or confuse the user. In this case you are still not doing anything wrong when choosing the slowest and therefore safest settings, the system will run the same as if you would use the fastest but most dangerous settings.
It Doesn’t Have to Be PC100 SDRAM !
The memory issue is anyway one of the most important ones with 440BX. 100 MHz bus are asking 50% more performance of the SDRAM modules and there is a good chance that you can throw away the PC66 SDRAM you bought for your Socket 7 or 440LX board. However, I was very surprised that although some boards wouldn’t run some PC100 DIMMs, almost all ran some high quality Samsung PC66 DIMM without the slightest problems. Thus you have a chance of being able to use your old PC66 SDRAM. As long as you bought very high quality. Quality is not the only thing that is required, one other important requirement is the serial presence detect SPD-EEPROM on the DIMM, which holds some important information about the setup and hold cycle times of the module and which is particularly responsible for the auto adjustment of the RAM timing in the board. Intel’s PC100 and PC66 specification allows 9 ns tAC for PC66 modules, but only 6ns for PC100. It is possible that PC66 SDRAM already applies to 6ns time for getting the data ready, but this is of course pure luck, although certainly worth a try in case you own PC66 SDRAM. My Samsung PC66 SDRAM ran fine in all boards that would let you use PC66 modules even up to 112 MHz and it’s now used in my very own system at 100 MHz FSB, the system where I am just now writing this article. Unfortunately are there quite a few BX boards that wouldn’t neccesarily run each PC100 DIMM without problems, the Toshiba PC100 SDRAM I used for testing had problems in three boards, all of them were dual CPU boards. The boards from Intel and Siemens wouldn’t let you run any SDRAM that doesn’t say ‘I am PC100’ in the SDP-EEPROM. The Intel board tells you this in particular whilst booting before it simply halts, the Siemens board just keeps a black screen.
Jumperless Becomes Fashionable
Two things are obviously becoming pretty common now. 15 of the 20 boards in this test were equipped with a hardware monitor that would inform you about CPU temperature, power supply voltages and fan speeds and 8 boards can call themselves ‘jumperless’, offering the CPU multiplier and FSB adjustment in the BIOS rather than the fiddely jumper job. This is certainly a satisfying situation and I hope that jumpers will finally become history. There are board manfacturers who say that their customers prefer a hardware over a software adjustment, but these ones I may ask why they can’t at least use tidy dip switches rather than messy jumpers.
All 20 boards ran with all AGP graphics cards I used, which are no less than a card with NVIDIA’s RIVA128, 3DLabs Permedia2, Intel’s i740, Matrox MGA-2064, Matrox MGA-G100, Matrox MGA-G200 and ATI’s Rage Pro (Turbo). What I found out though was that some boards get very touchy with the used SDRAM when using a RIVA, i740 or MGA-G200 card and these boards got a fail with these RAM in case the instabilities happened.
The ‘B21-Issue’ – Automatic FSB Adjustments
One important thing which sets BX boards apart from each other is the way the FSB is adjusted. Intel’s specification says that the board should check the state of CPU contact ‘B21’ and adjust the FSB to 66 or 100 MHz accordingly. 9 of the 20 boards apply to exactly this spec, 4 boards would let you either choose between 66, 68, 75 or 83 MHz in case you’ve got a 66 MHz FSB CPU or between 100, 102, 112, 133 for 100 MHz FSB CPUs. 7 board would leave the adjustments entirely up to you, so you can choose between 66 up to 133 MHz FSB, regardless which CPU is plugged in. In case you still should want to run at 100 MHz FSB or above with the boards that read out ‘B21’ you need to do the ‘B21-procedure’ I described in my article. The 7 boards that are not applying to Intel’s spec are listed under my ‘overclocking recommendations’.
AMI WinBIOS Is Big Disappointment
The compatibility test brought it to light that BIOS isn’t BIOS. AMI’s WinBIOS showed that not every BIOS manufacturer is able to see the sign of the times. Whilst every board with Award BIOS is able to distribute any available IRQ to any PCI and ISA slot, AMI’s WinBIOS is still in the BIOS middle age and assigns a fixed IRQ to each PCI slot, making my compatibility test a real painful experience. Thus all boards with this wonderful AMI WinBIOS failed the compatibility test and only one could make it into the recommendation list because the manufacturer (Tyan) had done some serious work against this occurance and because this one board has already everything onboard, making the compatibility test almost obsolete. However, not all boards with Award BIOS were the same as well. Asus, AOpen and Chaintech are some typical examples for companies who put a lot of effort in their own BIOS development, resulting in the most bug free Award BIOS available. Particularly Asus claims to have 40% own code in the BIOS and this is noticable when you plug in different PCI or ISA cards. Elitegroup is also using Award BIOS, but for some crazy reason the BIOS wouldn’t allow you to disable the IRQ for the USB. This resulted in a compatibility test failure, because there is simply no sense in USB sitting on an IRQ when USB isn’t used at all. Siemens showed how cool a BIOS can really be, since the D1064 with Phoenix BIOS was the best product in the compatibility test. This will certainly be due to a lot of BIOS development work within Siemens. Intel’s SE440BX is also using Phoenix BIOS, but it didn’t look quite as good in the compatibility test at all, especially when enabling the onboard sound it was a huge pain getting other cards to work.
ISA PnP Network Card Caused a lot of Problems
There was an interesting occurance using an ISA PnP network card. In no less than 11 cases the card was reported as running fine in the Windows 95 system manager, but it was still impossible to connect to any network. This problem could be solved as soon as the used IRQ was manually changed to IRQ5. Any IRQ higher than 7 would show this described inability to connect to a network and that without the slightest error message. Microstar’s MS-6119 is the only board with this failure that made it into the recommendations list and I hope that MSI will sort out this problem asap. Chaintech showed that it can be done. After I reported this error to Chaintech it took 2 days and they supplied me with a BIOS that didn’t have this problem anymore.
The performance of the boards was within a 3% range, which I consider as negligible. Thus even the ‘slowest’ board certainly offers satisfying performance.
Abit BX6
Abit’s BX6 was designed with overclocking in mind, but unfortunately there wasn’t enough time spent on compatibility issues. Nevertheless is the BX6 a step in the right direction if you compare it to its predecessor LX6. The compatibility is a lot better than it was with the LX6, which had problems with many PCI and ISA cards and also a serious keyboard problem. Abit is trying to make a big deal out of the data buffers, but I can’t help and stay unimpressed by this. Data buffers are normally required as soon as the board comes with 4 instead of only 3 DIMM slots and most of the other boards have them as well. The other point of the data buffers is that they can slow the board down, whilst covering timing problems with the RAM. Some good designs can do without this and actually run faster.
Abit forgot a few features, like e.g. the ‘SBLink’ connector.
The BX6 shares the ISA network card incompatibility problem with a lot of other boards and with some work of Abit’s BIOS engineers this problem should be fixed pretty soon. On the RAM side the board did very well. You can plug in pretty much any PC100 SDRAM and it even runs some PC66 SDRAM at 112 MHz. This doesn’t even require any settings change in the BIOS setup, because there are no RAM timing settings to be adjusted.
The board does not only offer the selection of FSB clocks from 66 to 133 MHz in its well known ‘CPU SoftMenu’, but also the option of adjusting the CPU voltage. Abit proudly presents this as ‘SoftMenu II’, forgetting the fact that this new feature is a great tool for killing your CPU. The BX6 is a board for people who like experiments and who like overclocking to insensible areas. The missing RAM timing adjustment in the BIOS setup makes it easy enough to crank up the FSB to crazy heights.
All in all the BX6 isn’t a bad board at all and for overclockers it could be the number one choice. However, if overclocking doesn’t play the major role, you can find better BX boards.
The BX6 didn’t make the overall recommendations list because of its compatibility problem in combination with low to average performance and the lack of any special feature.
AOpen AX6B
AOpen’s AX6B is a solid BX motherboard that is able to satisfy the needs of most people. It did not have any compatibility problems and it runs fine with any SDRAM I used for testing, including the PC66 SDRAM. The performance is average, the feature list is pretty full. The board offers ‘suspend to disk’ in case you like to do the same as what you may be used from your notebook.
The board offers the full overclocking range. With a convenient adjustment from the BIOS setup menu you can choose FSB clocks form 66 to 133 MHz.
This moterboard did not reveal any particularly outstanding findings, but its lack of weaknesses let it get right into the recommendations list. However, If you are particular about good DIMM slots you may dislike the AX6B.
Get the AX6B if you want to have a stable and reliable motherboard with a lot of features.
Asus P2B
Asus’ P2B made the top of my recommendation’s list due to a good combination of very good compatibility, a decent list of features and very good performance. The board could be even better if the CPU settings could be adjusted from the BIOS setup menu or at least via dip switches. Jumpers are really getting out of fashion, even Intel uses a software approach. The P2B is unfortunately equipped with only 3 DIMM sockets, which is the other critique I have.
I recommend the P2B as a well performing ‘low trouble’ solution. If you fancy onboard U2W SCSI or LAN or both you don’t have to look far as well, because boards with these features are available from Asus as well. I have already received the P2B -LS, a single board monster with U2W SCSI, Intel 100/10 MBit ethernet onboard and four instead of three DIMM sockets. I will test this board as soon as I’m back from Taiwan.
Chaintech 6BTM
Chaintech is proudly saying that their board is particularly compatible with any PC100 and a lot of PC66 SDRAM and my tests could do nothing but confirm that. The board comes with a whole lot of features, 4 DIMM sockets, all the ‘Wake up’ features, Power on Keyboard, fan off in suspend mode, … The BIOS was changed so that you can choose any FSB clock from 66 to 133 MHz regardless which CPU you are running and the compatibility tests were passed successfully too.
Chaintech had a good idea on how to make the CPU temperature sensor more accurate by placing it closer to the CPU, as shown in this little picture:
As you can see, the sensor is not at board level, but about an inch above the board, thus close enough to give you accurate CPU temperature information.
The performance is about average, but this doesn’t change the fact that this is one of the best single boards in the test.
DFI P2BXL
I actually received this board as a ‘Yakumo Y686BX’ as distributed by Frank&Walther, Braunschweig, Germany. The only difference to the ‘original’ DFI P2BXL is a ‘Yakumo’ sticker in front of the AGP slot.
The P2BXL passed all compatibility tests and ran with all the SDRAMs I tried. The CPU multiplier is adjusted via a dip switch on the board, the FSB clock can either be simply detected automatically by setting the FSB clock to ‘default’ in the BIOS setup, or you can choose from 66 to 133 MHz manually, reagardless which CPU is used.
The performance was at an average level.
All in all the P2BXL is a good BX motherboard with all the common features. The RAM timing can not be adjusted in the BIOS setup, which can only save a lot of headache whilst hardly having any disadvantage from the performance point of view.
This board is definitely worth being featured in the recommendations list.
DFI P2BXL/S
The brother of the P2BXL is called P2BXL/S and comes with an additional onboard ultra 2 wide SCSI host adapter, offering one U2W, one UW and one U-narrow port. The chip used is Adaptec’s new 7890AB.
The features apart form the SCSI adapter are identical to the P2BXL, you are adjusting the CPU settings via a mixed dip switch/BIOS setup solution. The dip switches are responsible for the multiplier, the BIOS setup menu gives you the option to choose between ‘default’, which detects the FSB automatically or from FSB clocks between 66 and 133 MHz.
Unfortunately there was a slight compatibility problem due to the IRQ grabbed by the 7890 chip. You should avoid using PCI slot #3 for a card that requires an IRQ. Voodoo2 cards or other devices which either don’t require an IRQ or which can share it can still be used in this slot.
The P2BXL/S is a good motherboard and the onboard U2W SCSI adapter makes it even more attractive. Without the compatibility problem the board would certainly have made the recommendations list.
Elitegroup P6BX-A+
Elitegroup’s P6BX-A+ board caused quite a few problems until I could start testing it. With the first BIOS the board was unable to run 66 or 100 MHz FSB clock, instead it only ran 68 or 103 MHz. c’t-Magazine had the same problem, but their BIOS allowed to manually switch to 66 or 100 MHz. The BIOS I had was stubbornly sticking to the ‘turbo mode’, so that I first couldn’t do any benchmarking tests. The second BIOS I received solved this problem, but I ran into another one. In the process of trying to adjust the BIOS for best performance I had enabled ‘AGP Bus Cacheing’. This leads to a complete inability of using the 3D hardware accelerator of any AGP card. You can imagine that it took me a while to find what the reason for this strange occurance was. Thus keep your hands off ‘AGP Bus Cacheing’, leave it DISABLED!
Finally the board could be tested and it showed above average performance. It also didn’t have the slightest trouble with any of the SDRAMs I used for testing, including a successful pass of the PC66 SDRAM test at 112 MHz. The P6BX-A+ comes with some really fancy connectors for the IDE and floppy cable. Clamps on both sides ensure a save connection.
The compatibility test turned out to be quite a bit of a pain however and the reason for it is so sad. Elitegroup decided that it is not necessary offering a BIOS setting for the disableing of the USB IRQ. Thus the USB port is blocking one IRQ permanently, regardless if you use it or not. As you can imagine, this caused quite a bit of trouble in my 5 card configuration. After 2 hours of trying I finally got all the cards to work. You can read what it took in the compatibility report.
Summarizing the occurances leaves a mixed impression. The P6BX-A+ does certainly have what it takes to be a good motherboard. However, unnecessary BIOS features as ‘AGP Cacheing’ and the lack of ‘Disable USB IRQ’ are currently the reason why this board is a potential trouble maker. If Elitegroup should change the BIOS, the board could easily make the recommendations list, because it runs stable and it runs fast.
Elitegroup has sent me a BIOS that will now let you disable the USB IRQ. ‘AGP Cacheing’ was removed from the setup menu. I will retest this board soon, but for now it looks as if the problems were resolved.
Intel SE440BX
Intel’s SE440BX is an interesting motherboard. It comes with onboard sound, using the Crystal CS4611 and CS4236B.
Since I received the board without the usual manual, I first had to find out how the CPU settings are changed. After looking around for a while I was sure that this board has to have some software mechanism for changing the CPU multipliers, since there are hardly any jumpers and certainly no dip switches on the board. Instead of this there is a special jumper that is called J8A1 ‘BIOS CONFIG’. This jumper has to have 2-3 shorted and after starting the system you are in a special configuration mode of the BIOS setup. Now the CPU multipliers can indeed be changed via software. After saving the config you have to short 1-2 and the board will start normally. I think that this is a very secure way of using a BIOS adjustable CPU configuration.
My tests with the SE440BX showed that the board is very stable with any of the tested PC100 SDRAM, but it would give you an error message and halt the system if you should dare plugging PC66 SDRAM into it. This is another pretty sensible security check, which can avoid a lot of possible trouble when using the wrong SDRAM.
The performance of the SE440BX was average and you can certainly imagine that this board doesn’t offer any chance of changing the FSB clock. Instead of this it detects the FSB clock automatically via CPU contact ‘B21’ and can hence still be overclocked using the ‘B21 manouver’ described on my website. However, no FSB frequencies besides 66 and 100 MHz are supported.
The compatibility testing was causing a lot of trouble. The IRQ for the USB cannot be physically disabled, although there’s a setting in the BIOS setup that doesn’t help, causing the first problem. The onboard Crystal sound chip is very inflexible regarding the IRQs it is using, so that I had to play PCI card jokey until I could finally get all the cards to work. There was also a problem with the ISA PnP network card. I couldn’t find any chance of getting this card to work properly (including a connection to the network) as long as the Crystal chip was not disabled.
The SE440BX is all in all an interesting board and I would have liked recommending it if I wouldn’t have had so many problems in the compatibility test. Not unlike the Siemens board it shows a lot of unique and intelligent features.
Iwill DBS100
Iwill’s DBS100 dual BX motherboard is a massive piece of electronics. It comes with onboard dual UW SCSI host adapter, using Adaptec’s 7895 chip and offers a RAIDport II, as it seems to become habit with dual boards now.
The 5 card compatibility test was passed without any problems, but unfortunately the DBS100 stumbled over the IS PnP network card as many other boards as well.
Unfortunately the board would refuse running with the Toshiba PC100 SDRAM, but it ran stable with the Corsair PC100 SDRAM as well as with the Samsung PC66 SDRAM. I will do further tests to see which SDRAM can be used in this board.
The performance was average, but this is really not of any particular importance in case of a dual board.
The DBS100 is certainly an interesting dual Pentium II BX motherboard which requires a few flaws sorted out. I really liked the board and it is certainly an interesting alternative for people who think that Tyan’s Thunder 100 Pro is overkill.
Who ever played Quake II on the net whilst recording a CDROM at the same time will know why I prefer dual CPU systems under Windows NT over any single CPU system.
Microstar MS-6119
Microstar could carry on its success in building fast and reliable Pentium II boards. Microstar’s LX board already received my recommendation and the MS-6119 could have made the top of the recommendations list if it wouldn’t have a few minor problems. The board is first of all not as easily overclockable as e.g. the Asus P2B. It recognizes the CPU FSB clock, so that you can either choose between 66/68/75/83 or 100/103/112/133 MHz in the BIOS setup menu. The good thing is that you won’t require any fiddleing with jumpers, the CPU setup is completely done via the BIOS setup.
MSI included all the popular new features, offering an optional hardware monitor as well. Also using this fancy heat sensor that’s closer to the CPU:
The compatibilty testing went fine in the ‘5 card test’, however, the PnP ISA network adapter didn’t want to connect to a network unless it was using IRQ 5.
Performance wise the MS-6119 was about the fastest board in the test, so that it all in all could still make second in the recommendations list.
QDI Brilliant I
I received a few mails from people who seem to be very religious about QDI, claiming they’d make ‘the best motherboards in the world’. Well, I would never dare saying this about any company, but I do admit that QDI is certainly a well organized Chinese motherboard manfacturer that procuces good products.
The Brilliant I however is pretty much an average board. It comes with only 3 DIMM sockets, uses the ‘SpeedEasy’ jumperless CPU setup, which has become very fashionable these days and it performs average too.
The features are also pretty much the same as what its competitors come with, including all the Wake Up features as well as Hardware monitoring.
The board ran fine and stable at up to 112 MHz with the three different memories I tested and it offers FSB clocks of either 66/68/75/83 or 100/103/112/133 depending on the FSB speed of the CPU. If you wish to overclock a FSB 66 MHz CPU to 100 MHz FSB or more, you’ll require the ‘B21 procedure’ to do that.
Compatibility wise the Brilliant I was doing fine in the ‘5 card test’, however it had a problem with the PnP ISA network card, which wouldn’t connect to a network as soon as a SCSI card was plugged in as well and the IRQ was not 5.
The average performance, the compatibility problem, only three DIMM sockets and the difficulty in overclocking are the reasons why the Brilliant I could not make the recommendations list.
Siemens D1064-E
Siemens’ last motherboard ‘Made in Germany’ is quite a special one. It’s the only board in the test which optionally comes with Matrox’ G100 chip onboard, which means that it could possibly even be equipped with the G200 chip, since this chip is pin compatible to the G100. Other special features include a port for a chip card reader, a special dip switch for ‘BIOS restore’, which can save your life after an unsuccessful BIOS flash update. Additionally to the normal Wake up features it also offers ‘Wake Up on Chip Card’ and ‘Keyboard Power On’.
The board ran stable with each of the tested PC100 SDRAM and it refuses to work with any SDRAM that identifies itself as PC66 SDRAM in its SDP EEPROM. The compatibility test was a very pleasant experience, since the D1064 didn’t require any BIOS settings to pass all the tests.
This motherboard is certainly targeted to the corporate market, where overclocking features are not of any interest, but where easyness of use is the most important topic. This is what the D1064 does very well.
The board did not make it into the overall recommendations list because of its lack of overclocking features, its inability to run even good PC66 SDRAM and because of its only average performance. However, this motherboard could be very interesting for corporate use, especially in combination with the onboard MGA-G100 VGA chip.
Soyo SY-6BA
Soyo’s 6BA motherboard is another example for stability, reliabilty and flexibility. The board is equipped with all the important features, although lacking a few ‘Wake UP’ ones, but coming with the hardware monitor, which is important to quite a few people nowadays. It also has four instead of the common three DIMM sockets.
The stability with the three SDRAMs I tested was excellent up to 112 MHz and the compatibilty test also showed that Soyo as one of the largest motherboard manufacturers can afford a BIOS team that gets the job done right.
Performance was average, which I consider as a minor problem. All in all the 6BA fully deserves making it on the recommendations list.
Supermicro P6DBS
Supermicro’s P6DBS dual BX motherboard is another impressive piece of hardware, big and well equipped. It comes with onboard dual UW SCSI and a RAIDPort II, making it a nice worksation or server motherboard. It’s of course equipped with 4 DIMM sockets as well.
Unfortunately there was a little problem with Toshiba SDRAM chips, the board only ran it at slow settings and even that wasn’t too stable. It seems to be a common problem, because the Iwill and one of the Tyan dual CPU boards ran into the same trouble. It worked without any problems when using the Samsung PC66 SDRAM, which shows that the RAM timing is not a really serious issue. However, you could be facing problems when not buying the right PC100 SDRAMs.
The compatibility testing turned out to be some kind of nightmare, but I can’t really blame Supermicro in the first place, since the problems were mainly caused by AMI’s wonderful WinBIOS. The WinBIOS has got a serious problem with the IRQ assigning for the different PCI slots, making my default configuration completely impossible to run. Instead of this I had to play PCI card jokey. There is an option of disabling the USB IRQ in the BIOS setup menu, but USB still grabs IRQ 9 regardless what setting you choose. The same is valid for the onboard SCSI, which has a ‘disable’ setting, but after that IRQ 10 is still used by it. Additionally, AMI WInBIOS would not let the ISA PnP network card connect to a network unless it’s using IRQ 5.
Although this board showed the best perfomance amongst the dual boards, the problem with the SDRAM as well as the huge compatibility problem were strong enough destroying any chance for the P6DBS of entering any recommendations list.
Supermicro P6SBA
The P6SBA is some kind of ‘low end’ version for a product with the name ‘Supermicro’ on it. Only 3 DIMM slots and no special onboard components make it look like it’s a starved brother of the P6DBS.
While the board shares all the compatibility problems with its brother P6DBS, the SDRAM test showed that the P6SBA is a lot more stable. It ran fine with all memories I used for testing.
Let me again mention the pain I had to go through when doing the compatibility testing. The ‘5 card test’ was a nightmare and the ISA PnP network card wouldn’t do as it should as well.
Performance wise the P6SBA was one of the fastest boards in the test, but it doesn’t offer any chance for overclocking. The FSB is adjusted automatically between 66 and 100 MHz.
Overall the P6SBA is a board with too many problems whilst offering too few advantages. Thus you won’t be able finding it on any of the recommendations lists.
Tekram P6B40-A4X
Tekram is only about entering the motherboard market after making a name with their SCSI adapters. The P6B40 single CPU BX motherboard is small but it still comes with 4 DIMM sockets, which I consider as remarkable.
Onboard features include all that it takes, like all the ‘Wake Ups’ as well as hardware monitoring.
Compatibility testing was to my surprise quite a pleasure, the P6B40 did not run into any problem at all. Memory testing showed the same good result, the board ran with all three test SDRAMs stable at up to 112 MHz FSB.
Overclocking is possible but not quite that easy, since the FSB can only be choosen between either 66/68/75/83 or 100/103/112/133 MHz depending on the FSB clock the CPU is designed for, so that you need the ‘B21 procedure’ for getting to 100 MHz FSB or above when using a Celeron or a PII 233-333.
The performance of the board was at a low average level which is the reason why the P6B40 didn’t quite make it into the recommendations list.
Tyan Tiger 100
The Tiger 100 is Tyan’s dual BX board ‘version for poor people’. Unlike the Thunder it doesn’t have any special onboard components, keeping it a lot smaller than the Thunder 100.
Compatibility testing of this board was as painful as it was with the Supermicro P6DBS, mainly due to the AMI WinBIOS they both use. Please refer to the compatibility evaluation page for further details.
The board does not offer any overclocking features and the performance was worst in the whole field. However, the Tiger 100 was the only dual board in the test which did not have any problems with PC100 SDRAM that’s equipped with Toshiba chips.
Tyan is already working to improve the BIOS, the stability of a dual board is more important than the performance, so that this motherboard could still become an interesting alternative for people who want to have a dual Pentium II system, but who don’t require onboard SCSI.
Tyan Tsunami
Tsunami is the name of Tyan’s simple single BX motherboard. It doesn’t come with any special onboard stuff and it has only got 3 DIMM sockets as well. However, it has got a wonderful green LED and an onboard speaker. I discussed with Tyan about the green LED and they assured me that for people who don’t like this LED they will happily supply a red or yellow LED instead.
The compatibility test was also under the threat of AMI’s WinBIOS, making the tests pretty painful. Please refer to the compatibility report for further details. The board didn’t have any problems with the different SDRAMs I used for testing though.
Overclocking can be done wen using the latest AMI Text BIOS. This BIOS will let you choose higher FSB clocks than the automatically detected 66 or 100 MHz.
The performance of the Tsunami was pleasing, making it one of the fastest boards in the test field. Tyan is working on the BIOS problems so that I hope the board will be good enough for making it into the recommendations list the next time I test it.
Tyan Thunder 100 Pro
Tyan’s new mega-board ‘Thunder 100 Pro’ offers almost everything the heart of a dual motherboard freak like me could ask for. Onboard dual UW SCSI, onboard 100/10 MBit Intel ethernet and onboard Vibra VX sound don’t leave much to desire and of course this board comes with 4 DIMM sockets as well.
Compatibility testing seemed like a farce to me, since there’s hardly anything you would have to plug onto this board anymore after equipping it with an AGP graphics card and a Voodoo2 3D accelerator. Nevertheless was the Thunder 100 Pro the only board in the test with AMI WinBIOS, that was able running the ISA PnP network card with any IRQ, although it’s extremely unlikely that anyone would put such a cheap card in this precious motherboard. Another feature that I really liked is the ability to enable each single onboard component via a jumper.
The performance of the Tiger 100 was average in the field, which is still pretty remarkable for a dual board. Of course there aren’t any overclocking features supported, something that’s also typical for dual motherboards. The FSB is adjusted automatically between 66 or 100 MHz and if you should indeed want to run your PII 233-333 at 100 MHz FSB you’ll require the ‘B21 procedure’.
The only annoying thing i found was the Thunder’s inability of running PC100 SDRAM with Toshiba chips. So far i’ve tested two different vendors, both with the same negative results. Samsung’s PC66 SDRAM runs fine however, so does Kingston and Corsair PC100 memory.
The Thunder 100 Pro was good enough making it into my own system, but I still hope that Tyan will sort out the memory problem.
Zida Tomato 6ABX
I can’t remember when I stumbled across the name ‘Tomato Board’ for the first time, but I think it was my ctcm motherboard survey some two years back. Zida is still building motherboards with this funny name and the boards aren’t as strange as the name would make you think.
The 6ABX is the only motherboard in this test using AMI’s normal good old text BIOS, which I prefer a lot over this fancy WinBIOS stuff. However, AMI is still a warrant for trouble regardless if text or windows BIOS. The compatibility test was another painful experience, pretty similar to my experiences with AMI WinBIOS. Please refer to the compatibility report for more details.
The stability of the Tomato board was good with all the SDRAMs I tested.
Overclocking is not the field of the 6ABX, FSB is adjusted automatically between 66 and 100 MHz, any higher FSB clocks are unavailable.
The performance of the 6ABX was low average and thus another reason for not getting into a recommendations list.
Compatibility Report
The most important new feature of my new motherboard tests is certainly the compatibility testing. To do this I tried to be as nasty as I could, making a successful pass as difficult as possible. This test shall reflect situations which come close to what my readers could see in their own system after plugging a new add-on card in their system.
Card Compatibility Test 1
The first and most important test shows the following situation:
I decided that a modern system should not only run with an AGP graphics card, but also with a Voodoo2 3D add-on card. This is not asked too much, since the Voodoo2 cards don’t require an own interrupt and only some mapped memory area which should be easy to supply. I also expect a PCI UW SCSI card to run in this system and I used the pretty common Adaptec 2940UW SCSI host adapter, but also the DPT PM-2144UW and the Tekram DC-390F. All cards showed the same, they either ran all or none of them did. Additionally to this I used a cheap PCI network adapter which you can get for $30 to $40. The used card is the Realtek RTL8029 PCI card, which is just fine for the usage in multiplayer gaming and some basic file transferring. To be really nasty I also used an old Soundblaster16 ISA legacy card, which does not do PnP.
Hence the system had to host five components at the same time:
- an AGP graphics card, the Canopus Total3D 128V AGP in the AGP slot
- a 3Dfx Voodoo2 add-on 3D card, the Creative 3D Blaster Voodoo2 in the PCI slot next to the AGP slot, which usually is PCI Slot #1 unless it’s a Supermicro board that counts the other way around.
- a PCI UW SCSI host adapter, either the Adaptec 2940UW PCI or an onboard SCSI adapter if available in the second next PCI slot after the AGP slot, usually #2
- a PCI LAN adapter, either the Realtek RTL8029 or an onboard LAN adapter in the third, usually #3
- an ISA sound card, either the Creative Soundblaster16 ISA legacy or an onboard sound device in any ISA slot, there’s no difference
The device were plugged into the board in the following succession:
- AGP graphic
These five devices were supposed to run all without any problems, which means that you had to be able to use the RIVA’s 3D features, the Voodoo2’s 3D features, a SCSI hard drive had to run at acceptable speed, I had to be able to connect to my network and the sound card was supposed to play sound. The Windows 95 system manager was not supposed to report any problems.
On top of that I expected COM1, COM2 and LPT1 to work properly as well as both EIDE ports, running an EIDE hard drive on the one port and a IDE CD-ROM at the other.
This may sound tough, but it should certainly be possible as you can see in this list:
- COM1 taking IRQ4
- COM2 taking IRQ3
- LPT1 taking IRQ7
- Soundblaster16 taking IRQ5
- primary IDE port taking IRQ14
- secondary IDE port taking IRQ15
- AGP graphics card taking IRQ9-11
- SCSI adapter taking IRQ9-11
- network adapter taking IRQ9-11
As you can see, there are definitely enough IRQ’s available, the AGP, SCSI and network card only have to distribute IRQ9, IRQ10 and IRQ11 amongst each other properly and everything would run fine. There is one problem though. Since the RIVA doesn’t like sharing its IRQ with any device and the SCSI and network cards don’t work with IRQ sharing either, there wasn’t any IRQ available for the USB anymore. Thus the board ran into trouble if the USB IRQ could not be disabled. In case of the Award BIOS, IRQ5 had to be declared ‘ISA legacy’, so that it wouldn’t be distributed amongst the cards that are asking for an IRQ. This was not counted as a test failure, because the Soundblaster16 is an ISA legacy card indeed. Only the Siemens D1064 didn’t require this procedure.
Card Compatibility Test 1 – The Results
The comments in RED reflect a successful pass of the test.
Compatibility 5 Card what had to be done to run them |
IRQ Table (video-network-SCSI-sound) | |
Abit BX6 | making IRQ 5 ISA Legacy, manual IRQ table is very difficult to access in the BIOS setup | 9-10-11-5 |
AOpen AX6B | making IRQ 5 ISA Legacy | 9-10-11-5 |
Asus P2B | making IRQ 5 ISA Legacy | 11-9-10-5 |
Chaintech 6BTM | making IRQ 5 ISA Legacy | 11-9-10-5 |
DFI P2XBL (Yakumo Y686BX) | making IRQ 5 ISA Legacy | 11-9-10-5 |
DFI P2XBL/S | if PCI network card is plugged into PCI Slot #3, system will not work because onboard SCSI adapter uses the same resources as PCI Slot #3. Using any other PCI Slot only requires IRQ 5 set to ISA Legacy and everything works fine. Thus PCI Slot #3 should only be used for cards that don’t require an IRQ, like e.g. a Voodoo2 card.On board SCSI cannot be disabled. | 11-10-9-5 |
Elitegroup P6BX-A+ | making IRQ5 ISA Legacy, USB IRQ cannot be disabled, hence test was only successful after manually disabling USB port in Windows 95 system manager and moving PCI network card from PCI slot 3 to PCI slot 4. Otherwise network and SCSI card get the same IRQ which results in a resource conflict. | 11-10-9-5 |
Intel SE440BX | USB IRQ cannot be disabled, hence test with SB16 soundcard (Crystal sound disabled) was only successful after manually disabling USB port in Windows 95 system manager, for using the onboard Crystal sound PCI network card had to be moved to PCI slot 4 (from PCI slot 3) and SCSI card had to be moved from PCI slot 2 to PCI slot 3, USB had to be disabled in Windows 95 system manager also. | 11-9-10-5 |
Iwill DBS100 | making IRQ 5 ISA Legacy | 11-9-10-5 |
Microstar MS-6119 | making IRQ 5 ISA Legacy | 11-9-10-5 |
QDI Brilliant I / P6144BX | making IRQ 5 ISA Legacy | 11-9-10-5 |
Siemens D1064-E | no problems | 11-10-9-5 |
Soyo SY-6BA | making IRQ 5 ISA Legacy | 10-9-11-5 |
Supermicro P6DBS | board shows interesting phenomenon. WinBIOS would not let you disable USB properly, which grabs IRQ9. To use this IRQ you have to plug the PCI card into PCI Slot 4 (the one next to the AGP Slot), PCI Slots 1 and 2 will only get IRQ11, which is already used by AGP VGA and it leads to crashes with several AGP cards if this IRQ is shared. Hence the network card needs to be plugged into PCI 4 to get the last free IRQ (IRQ9), whilst the onboard SCSI adapter always grabs IRQ10, otherwise the system will not run with e.g. a RIVA card. To make this all work you will have to disable the USB manually in the Windows 95 system manager. This board shows a very inflexible way of distributing IRQs over the 4 PCI slots, which cannot be improved by any BIOS setting. If you are using an AGP card and need the EIDE ports as well as both COM ports and the printer port, you can *only* use PCI Slot 4 for *one* card that needs its own IRQ. | 11-9-10-5 |
Supermicro P6SBA | board shows interesting phenomenon. WinBIOS would not let you disable USB properly, which grabs IRQ9. To use this IRQ you have to plug the PCI card into PCI Slot 4 (the one next to the AGP Slot), PCI Slots 1 and 2 will only get IRQ11, which is already used by AGP VGA and it leads to crashes with several AGP cards if this IRQ is shared. Hence the network and the SCSI card need to be plugged into PCI Slot 3 and 4, otherwise the system will not run with e.g. a RIVA card. To make this all work you will have to disable the USB manually in the windows 95 system manager. This board shows a very inflexible way of distributing IRQs over the 4 PCI slots, which cannot be improved by any BIOS setting. If you are using an AGP card and need the EIDE ports as well as both COM ports and the printer port, you can only use PCI Slot 3 and 4 for cards that need their own IRQ. | 11-9-10-5 |
Tekram P6B40-A4X | making IRQ 5 ISA Legacy | 11-10-9-5 |
Tyan Tiger 100 S1832DL | board shows interesting phenomenon. WinBIOS would not let you disable USB properly, which grabs IRQ9. To use this IRQ you have to plug the PCI card into PCI Slot 4 (Tyan is counting the correct way, so PCI4 is the fourth PCI slot starting from the AGP slot), PCI Slots 1 and 2 will only get IRQ11, which is already used by AGP VGA and it leads to crashes with several AGP cards if this IRQ is shared. Hence the network and the SCSI card need to be plugged into PCI Slot 3 and 4, otherwise the system will not run with e.g. a RIVA card. To make this all work you will have to disable the USB manually in the windows 95 system manager. This board shows a very inflexible way of distributing IRQs over the 4 PCI slots, which cannot be improved by any BIOS setting. If you are using an AGP card and need the EIDE ports as well as both COM ports and the printer port, you can only use PCI Slot 3 and 4 for cards that need their own IRQ. | 11-9-10-5 |
Tyan Tsunami S1846 | board shows interesting phenomenon. WinBIOS would not let you disable USB properly, which grabs IRQ9. To use this IRQ you have to plug the PCI card into PCI Slot 4 (Tyan is counting the correct way, so PCI4 is the fourth PCI slot starting from the AGP slot), PCI Slots 1 and 2 will only get IRQ11, which is already used by AGP VGA and it leads to crashes with several AGP cards if this IRQ is shared. Hence the network and the SCSI card need to be plugged into PCI Slot 3 and 4, otherwise the system will not run with e.g. a RIVA card. To make this all work you will have to disable the USB manually in the windows 95 system manager. This board shows a very inflexible way of distributing IRQs over the 4 PCI slots, which cannot be improved by any BIOS setting. If you are using an AGP card and need the EIDE ports as well as both COM ports and the printer port, you can only use PCI Slot 3 and 4 for cards that need their own IRQ. | 11-9-10-5 |
Tyan Thunder 100 Pro S1836DLUAN | no problems, everything is already onboard | 11-9-10-5 |
Zida Tomato Board 6ABX | although USB is disabled in BIOS, it still tries grabbing an IRQ. Since IRQs 9-11 are taken by network, SCSI and AGP VGA card, it grabs IRQ3, disabling COM2. To avoid this, both COM ports and the printer port need to get actual port number assignment rather than the ‘auto’ default setting. Then USB uses IRQ11, the same as AGP VGA card. VGA card doesn’t care about this if USB is disabled in Windows 95 system manager. This BIOS would be a lot better than the WinBIOS if USB would be disabled properly, since it distributes the IRQs correctly and avoids the disastrous IRQ-sharing between VGA and another card. | 11-9-10-5 |
Card Compatibility Test 2
The second test was a lot more simple. I only expected an ISA PnP network card to run properly and be able to connect to my network. The used card is an extremely cheap Realtek RTL8019 card, which you can get for less than 20 bucks, but which is just fine for basic file transferring at up to 900 kB/s and multiplayer games. The card was reported as running fine by the system manager in most of the cases. However, in no less than 11 cases I couldn’t connect to the network unless I manually changed the IRQ to less than IRQ8. Of course only IRQ5 is available in this area unless you want to disable COM or LPT ports and if you have got a sound card, IRQ5 s usually in use already as well.
Compatibility ISA PnP Network Card | |
Abit BX6 | problem when SCSI PCI card in system, only connects to network when IRQ 3-7 is used |
AOpen AX6B | no problems |
Asus P2B | no problems |
Chaintech 6BTM | no problems |
DFI P2XBL (Yakumo Y686BX) | no problems |
DFI P2XBL/S | no problems |
Elitegroup P6BX-A+ | no problems as long as USB is disabled in Windows 95 system manager |
Intel SE440BX | ISA PnP network card needs to get the resources assigned manually, because of the USB grabbing an IRQ, however, IRQ5 cannot be released from Crystal chip and network card doesn’t connect to network with IRQs higher than 7. Hence test *unsuccessful*! SCSI card would not work in PCI slot 2 when Crystal sound chip is enabled!! |
Iwill DBS100 | problem, only connects to network when IRQ 3-7 is used |
Microstar MS-6119 | problem when SCSI PCI card in system, only connects to network when IRQ 3-7 is used |
QDI Brilliant I / P6144BX | problem when SCSI PCI card in system, only IRQ 3-7 work |
Siemens D1064-E | no problems |
Soyo SY-6BA | no problems |
Supermicro P6DBS | card can only connect to network if IRQ is within 3-7 |
Supermicro P6SBA | card can only connect to network if IRQ is within 3-7 |
Tekram P6B40-A4X | no problems |
Tyan Tiger 100 S1832DL | no problems, everything is already onboard, but with onboard network card disabled it works as well |
Tyan Tsunami S1846 | card can only connect to network if IRQ is within 3-7 |
Tyan Thunder 100 Pro S1836DLUAN | card can only connect to network if IRQ is within 3-7 |
Zida Tomato Board 6ABX | USB is grabbing IRQ9 although disabled in BIOS, therefore resource conflict of ISA PnP network card when SCSI card is present, IRQ requires to be assigned to network card manually in Windows 95 system manager and only IRQs 3-7 let you connect to network, IRQ9 doesn’t, although system manager doesn’t report any problems |
SDRAM Compatibility
The next test was the RAM compatibility test. I was using a 64 MB Corsair PC100 w/parity module which uses Samsung SDRAM chips (KM48S8030BT = 2Mx8x4, 4k refresh, mapping 14/11) and a 64MB Toshiba PC100 module (TC59S6408FTL-80H = 4Mx4x4, 4k refresh, mapping 12/9). I also used two 32 MB Samsung PC66 SDRAM modules (module: SEC KMM366S403BTN-G0, chips: SEC KM48S2020BT-G10 = 2Mx8x2). It’s my experience that modules with 2Mx8 chips are running better than with 4Mx4 architecture. Corsair has always used 2Mx8.
These are the results:
PC100 Memory Compatibility | Samsung SPD PC66 running at 100 MHz? | |
Abit BX6 | both memories stable up to 112 MHz | yes, even at 112 MHz !! |
AOpen AX6B | both memories stable at fast settings up to 112 MHz | yes, even at 112 MHz !! Change RAS-CAS delay to 3 |
Asus P2B | both memories stable at fast settings up to 112 MHz | yes, even at 112 MHz !!, Memory Timing to ‘by SPD’ |
Chaintech 6BTM | both memories stable at fast settings up to 112 MHz | yes, even at 112 MHz !! Change RAS-CAS delay to 3 |
DFI P2XBL (Yakumo Y686BX) | both memories stable at fast settings up to 112 MHz | yes, even at 112 MHz !! |
DFI P2XBL/S | both memories stable up to 112 MHz | yes, even at 112 MHz !! |
Elitegroup P6BX-A+ | both memories stable at fast settings up to 112 MHz | yes, even at 112 MHz !! Change RAS-CAS delay to 3, SDRAM Leadoff to 4, other 2 values to 3 as well |
Intel SE440BX | both memories stable | no! Board detects SPD and reports wrong spec, then system halts. |
Iwill DBS100 | Samsung only! | yes, at slowest memory timings |
Microstar MS-6119 | both memories stable at fast settings up to 112 MHz | yes, even at 112 MHz !! RAS to CAS delay ‘3’ |
QDI Brilliant I / P6144BX | both memories stable at fast settings up to 112 MHz | yes, even at 112 MHz !! RAS to CAS delay ‘3’ |
Siemens D1064-E | both memories stable | board wouldn’t even boot |
Soyo SY-6BA | both memories stable at fast settings up to 112 MHz | yes, even at 112 MHz !! RAS to CAS delay ‘3’ |
Supermicro P6DBS | Samsung at fast settings, Toshiba only at slow settings | yes, slow settings 3-3-3-4 |
Supermicro P6SBA | both memories stable at fast settings | yes, slow settings 3-3-3-4 |
Tekram P6B40-A4X | both memories stable at fast settings up to 112 MHz | yes, even at 112 MHz !! RAS to CAS delay ‘3’ |
Tyan Tiger 100 S1832DL | Samsung only | yes, at slowest memory timings |
Tyan Tsunami S1846 | both memories stable at fast settings | yes, even at fast memory timings |
Tyan Thunder 100 Pro S1836DLUAN | both memories stable at fast settings | yes, even at fast memory timings |
Zida Tomato Board 6ABX | both memories stable at fast settings | yes, even at fast memory timings |
Benchmark Results
Winstone 98 was ran in the following configuration:
- OS Windows 95 OSR 2.1
- Screen Resolution 1024×768, 16 bit color, 85 Hz vertical refresh
- SCSI Hard Drive IBM DGVS 09U
- SCSI Adapter Adaptec 2940UW
- RAM 64 MB Corsair PC100 SDRAM
- Graphics Card Diamond Fire GL 1000 Pro AGP, driver built 2255
- Intel Pentium II 400 CPU
- L2 Cache ECC was turned off in all boards
As you can see, the differences are minimal.
Ran with additional Creative Labs 3D Blaster Voodoo2.
Ran with additional Creative Labs 3D Blaster Voodoo2.
Recommendations
As explained in the new Motherboard Guide, I’m now not only focussing on performance but particularly on stability, features and compatibility. Thus there will be an overall recommendation as well as a recommendation in each of these areas.
Overall Recommendations
The seven motherboards that showed the best combination of stability, compatibility, features and performance are:
- Asus P2B
This board is an example of stability as it runs with each SDRAM I threw against it, up to 112 MHz FSB. The compatibility tests were all successful, I only had to block IRQ5 for the ISA legacy soundcard following the official specs. The board is very flexible for overclockers as it offers FSB clocks from 66 to 133 MHz regardless which CPU is plugged in. On the negative side the board does not offer a software adjustment of FSB and multiplier, you have to use old fashioned jumpers. The board offers all the ‘Wake ups’, ‘Power on Keyboard’, SBLINK and a hardware monitor. From the performance point of view the P2B is one of the fastest boards. All in all this board is fast, stable, reliable, very compatible and it carries a well known name. There are also versions with onboard U2W SCSI and onboard LAN adapter. - Microstar MS-6119
The MS-6119 would almost have ruled this review, if it wouldn’t have stumbled over an ISA network card. This card ran only with IRQ5, IRQs 9-11 would result in an inability to connect to any network. Otherwise the board is as stable as the Asus P2B and it offers more features. It comes with SBLink and all the new ‘Wake up’ features, you can get it with an optional hardware monitor and it can switch off the CPU fan in suspend mode for the ones who hate noise. Multiplier and front side bus can be adjusted from the BIOS setup, so that you don’t need to fiddle around with jumpers or dip switches. People who want to run 66 MHz FSB CPUs at 100 MHz FSB or more will need to do the procedure of covering up ‘B21’, because otherwise the BIOS wouldn’t let you choose any FSB above 83 MHz. Performance wise it’s the fastest board in the test, a tiny bit in front of the Asus, Supermicro and Tyan board. The MS-6119 is certainly a motherboard to be considered and as soon as Microstar’s BIOS engineers have fixed the problem with the ISA network card it will even surpass the Asus P2B. It is also a very attractive board from the price point of view. - AOpen AX6B
The AX6B is also running rock stable up to 112 MHz FSB with any SDRAM I used and it offers the adjustment of the CPU multiplier as well as the FSB from 66 to 133 MHz in the BIOS setup, thus also no jumper setting is required. The features include not only all the different ‘Wake ups’ and SBLink, but as the only one in the test it offers ‘suspend to disk’, as known from notebooks, which enables you to start your system at the very same state as you left it before. There was no flaw in the compatibility tests and the board is equipped with 4 DIMM slots instead of only three as the two boards above. However, I have to criticise the cheap DIMM slots, which don’t offer any fixation for the plugged in SDRAM DIMMs. The performance of the AX6B is in the middle of the test field. This board is a stable and very compatible solution, which will also apply to overclockers. - Chaintech 6BTM
Chaintech’s 6BTM is another example for motherboard stability, it also ran flawlessly with the Samsung PC100 SDRAM from Corsair as well as with the PC100 Toshiba SDRAM and with the PC66 Samsung SDRAM. The CPU settings can all be adjusted in the BIOS and since Chaintech made some adjustments due to my Celeron article, you can now choose from 66 to 133 MHz FSB regardless which CPU is plugged in. The features of this board include almost everything except of ‘suspend to disk’, like all ‘Wake up’ modes, hardware monitor, SBLink, Keyboard Power On, fan off in suspend mode. It also comes with 4 DIMM slots instead of the common 3 and these slots are of good quality. After Chaintech improved the BIOS the 6BTM passed all compatibility tests without a glitch. The performance is equal to the AOpen AX6B in the middle of the test field and if you can live without ‘suspend to disk’ I would prefer it over the AX6B even. - DFI P2BXL (Yakumo Y686BX)
The P2BXL is another board that proves the fact that quality doesn’t have to be expensive. The board ran rock stable with all SDRAMs, it offers all common features and it didn’t have any problems in the compatibility tests. It recognizes the FSB of the CPU, but you can choose 66-133 Mhz FSB in case you do not choose ‘default’ in the BIOS setup. The FSB is adjusted in the BIOS setup, the multiplier is adjusted via dip switches on the board. The performance is average, which means it’s only 1.5% slower than the fastest board in the test. This board is good and comes at an attractive price. - Soyo SY-6BA
Soyo’s SY-6BA made only place six in this list, because it’s a tiny bit slower than the other five above. It also runs absolutely stable with all tested SDRAMs, has no compatibility problems and is nice for overclockers, because it offers FSB clocks from 66 to 133 easily adjustable in the BIOS setup, no jumper setting is required. I was missing the features ‘Wake up on LAN’, ‘Wake up on Clock’ and ‘suspend to disk’, but the SY-6BA is coming with ‘Wake Up on Ring’, a hardware monitor and SBLink. It also offers 4 DIMM slots, however of cheap quality without any fixation for the DIMMs. It is the only one of the six recommendations, which has 5 PCI and only 2 ISA slots, so if this should be important to you … I read somewhere that Soyo would be a ‘small player in the motherboard arena’. This is completely wrong! Soyo is currently the 7th largest motherboard manufacturer in the world. - Tyan Thunder 100 Pro S1836D
The main reason why the Thunder 100 is last in this list is simply its price. People who don’t care about the costs will probably just love this board. Its onboard dual channel SCSI, 100 MBit Intel ethernet and Creative Vibra 16XV don’t leave much to desire. This board needs two Voodoo2 cards and a Matrox G200 AGP card and you’re all set. There wasn’t much to test from the compatibilty point of view, since this board has everything already integrated onboard. The only bad thing was its inability to run Toshiba’s PC100 SDRAM, however it runs rock stable with Samsung PC66 SDRAM, currently in my own Roadrunner. It shouldn’t surprise you that I had to get exacly this board into my own system. Please consider that you need the ‘B21-procedure’ if you want to run a 66 MHz FSB CPU at 100 MHz FSB. It’s not necessarily an overclocker board. The performance is average, which means it’s about 1.5% behind the fastest in this test. If you want it all, don’t take less than the Thunder Pro 100.
Dual Board Recommendations
The Supermicro P6DBS could unfortunately not make it inside the recommendation list here, although it shows the best performance of the tested dual CPU boards. This is due to incompatibility problems due to AMI’s WinBIOS as well as to the inability to run Toshiba PC100 SDRAM. Thus I consider this board as a definite trouble maker in its current state and cannot recommend it.
- Tyan Thunder 100 Pro S1836D
Of the four dual boards I tested the Thunder was the one with the best features and due to this equipment the compatibility test was an easy thing to pass. The only annoying thing with this board is its inability to run with Toshiba’s PC100 SDRAM, which could mean that it won’t run with other PC100 SDRAM as well. As long as you go for Corsair or maybe other PC100 SDRAM that’s using Samsung PC100 chips, you should be save though. - Iwill DBS100
Iwill’s DBS100 is the only dual board in the test with Award BIOS, resulting in only few compatibility problems. However I would have had appreciated if the ISA PnP network card would run with IRQs higher than 7 as well, so I hope that Iwill will do some work here. Chaintech has already proven that this is possible. This board would also not work with Toshiba’s PC100 SDRAM, so this is anoter thing to worry about. The performance is in the middle of the test field and the dual UW SCSI adapter and RAIDport II are some features to make you happy. - Tyan Tiger 100 S1832DL
I did a lot of thinking if I would include this board in my recommendation list, since it has the typical AMI WinBIOS incopatibility problem and its performance is at the bottom of the list. However, it’s the only dual board in my test that would run fine with any SDRAM I tested and this should make it worth a recommendation. Unfortunately this board does not come with onboard SCSI, sound or ethernet.
Compatbility and Stability Recommendations
Compatibility Test Results
Only motherboards that passed the compatibility tests without problems could make it into this list. Siemens’ D1064 is a shiny example for a ‘no-trouble’ board, since it seemed as if you could plug into it whatever you wanted and Windows 95 or NT would start without any complaints. The 6 boards in second place required to set IRQ5 to ‘ISA legacy’ for the ISA legacy soundcard I used in the test. This is still according to official spec and hence the test can be seen as successfully passed.
- Siemens D1064-E
The only board that didn’t require any BIOS setting for the successful passing of all the compatibility test was the D1064. This board comes with Phoenix BIOS and this BIOS seems to be smart enough recognizing that the ISA legacy soundcard is simply grabbing IRQ5 so that it wouldn’t try giving this IRQ to any other device. The board ran stable with both PC100 SDRAM brands and according to Intel’s spec it would refuse to work with PC66 SDRAM. You will see the D1064 again on top of the recommendation list for motherboard features. It is a very good board, particularly for corporate use and that’s what it was designed for. - AOpen AX6B, Asus P2B, Chaintech 6BTM, Soyo SY-6BA, Tekram P6B40-A4X, DFI P2BXL (Yakumo Y686BX)
These six motherboards only required the BIOS setting ‘IRQ5 – ISA Legacy’ to run the old ISA Soundblaster 16, which is according to spec. Otherwise there were no problems noticed, everything was running perfectly and without any problems. Stability was no issue for these six boards either, they all ran any SDRAM inclusive the PC66 Samsung DIMMs at even 112 MHz without the slightest glitch.
Feature Recommendations
Features List:
- Siemens D1064-E
Siemens equipped the D1064 with a lot of unique features. It offers:- Onboard Chip Card Reader Port
- Dip Switch for BIOS recover
- optional Onboard VGA, Matrox G100 chip, 4 MB, upgradeable to 8 MB
- Quiet Boot
- Wake Up on Chip Card
- Power on Keyboard
- Wake Up on LAN
- Wake Up on Ring
- Wake Up on Clock
- Hardware Monitor
- DFI P2XBL/S
The P2XBS is the only single CPU board in the test that has not only got a SCSI adapter onboard, but it’s even the latest one from Adaptec (AHA-7890AB), with the new Ultra 2 SCSI interface. The board also offers all te ‘Wake up’ features, SBLink and a hardware monitor. If there wouldn’t have been a slight compatibility problem which made it unable to run a PCI network card in PCI slot 3 without assigning resources manually, this board would definitely have made it into the overall recommendation list. - Tyan Thunder 100 Pro S1836D
Some of you may wonder why this board isn’t on top of this features list, but I decided to list single CPU boards first. All in all the S1836D offers more than most can dream of. It’s a dual Pentium II BX board with onboard dual channel UW SCSI (AHA-7895P), RAIDport II, onboard Intel 82558 10/100 Mbit Ethernet and onboard Creative Labs Vibra 16XV. The coolest thing is that you can disable each of these components with a jumper so well, that these components are indeed invisible to an OS as e.g. Windows95. It’s no question that all these features have its price, but if you want to have one of the most perfect all-in-one solutions, then you will fall in love with this monster board.
The chip card reader port may be a little bit too fancy, but I really like a sensible feature like the dip switch for recoverage of the original BIOS, in case disaster struck when you tried flashing a new BIOS. ‘Quiet Boot’ is a feature typical for corporate use, it keeps the screen black until the OS pops up. The optinional MGA-G100 VGA chip is also a very sensible solution for companies and here the chip card can be used very well for identification of the user. This last board that Siemens ‘Made in Germany’ shows how much thought a manufacturer can put into a product. It’s horribly sad that Siemens sold the motherboard manufacturing facilities to Acer.
Overclocker Recommendations
- Abit BX6
Abit focussed on overclocking again, but this time they almost took it over the top. The ‘SoftMenu’ doesn’t only offer the adjustment of the CPU multiplier and the FSB in a large range, it also gives you the ability to change the CPU voltage. In the first BIOS revisions the range was so big, that you could easily fry your CPU, as successfully accomplished by nobody less than Georg Schnurer, c’t-Magazine’s star motherboard tester, who had all reasons to be very upset about this crazy feature which killed a brand new Pentium II 400 CPU. Abit reduced the range, but I still want to make clear that it’s fairly unlikely to be more successful in overclocking after raising the CPU voltage to levels beyond Intel’s spec, whilst having a good chance of sending your shiny new Intel CPU into the eternal hunting grounds. Anyway, for the gamblers of you this board may be the number one choice. The BX6 does not offer any adjustments of the memory timing, which I consider as pretty good. It runs fine with all the memory I used for testing, but it had a compatibility problem with an ISA network card. The performance of the BX6 is average, the feature list is average too. - AOpen AX6B, Chaintech 6BTM, Soyo SY-6BA
These three boards let you adjust the FSB in a range from 66 to 133 MHz regardless which CPU is used. This as well as the multiplier setting is adjusted comfortably form within the BIOS setup. The performance of these three is equal as well, so you’ve got to look into the features list to find out which board applies to you most. - DFI P2XBL/S, DFI P2BXL (Yakumo Y686BX)
The two DFI boards have a dip switch for the multiplier setting whilst the FSB can be adjusted in a range from 66 to 133 MHz from within the BIOS setup. Whilst the P2XBL is a tad faster than the P2XBL/S, the latter comes with onboard U2W SCSI adapter. - Asus P2B
The P2B is certainly no bad board for overclockers, but it’s the only one in this list which requires fiddeling with jumpers. If you haven’t got a problem with that, you will certainly remember why this board is on top of the overall recommendation list.
Performance Recommendations
- MSI MS-6119
The MS-6119 is the fastest board in the test, which doesn’t mean that it’s a lot faster than the others in the list. It comes with pretty good compatibility, great stability and good features. - Asus P2B, Supermicro P6SBA, Tyan Tsunami S1846
These three boards are a tiny bit less fast than the MSI board. Unfortunately, the Tyan Tsunami board suffers from the incompatibilities caused by AMI’s pretty useless WinBIOS and so does the Supermicro P6SBA. - Elitegroup P6BX-A+
The P6BX-A+ is also a board from the fast league and if the BIOS setup would allow to disable the IRQ for the oh-so-wonderful USB, it would even have been mentioned in the overall recommendations list