For Vincent Tong
Missed your earlier message.

Suggest you

1 Check the Tyan site to confirm MPX is MP2000+ capable

2. try it with one processor at a time with single memory stick etc try both CPUs in each slot seperately to eliminate processor problems

Otherwise speak to your distributor there is enough on this thread to suggest there have been some manufacturing problems
Regards
brendan falvey

Thank you guys for your advice. I've managed to exchanged the board after finally convincing the shop that the board is faulty. However, I was so unlucky to get another board with a wrong bios. Anyway, got all these sorted out and I've happy "playing" with my new toys. So far so good.... but it's extremely hot!!
Thanks.

Hi, guys
After replacing the BIOS to the correct one, my USB is working but my NIC is causing problems. There's always a yellow ! and error code 10 in the device manager no matter how many times I reinstall the driver. Suspected a conflict. Took out every single card and tried installing XP, failed. Tried 2000 again, it completed the installation but there the yellow ! again!!! Reinstalled other cards caused windows startup failure. Disabling the NIC via jumper allows XP to completed installation.

Configurations are:
Tiger MPX 2466N-4M
2xMP 2000+
2x512MB Registered TwinMos PC-2100 DDR RAM
Adaptec 39160 SCSI
Advansys 3940AU SCSI
Creative Sound Audigy
Adaptec DuoConnect
ATI Radeon 64MB VIVO

Can anyone help??

Check your BIOS-Settings if the NIC-Slot is a MASTER (thats important for some cards, for example the intel cards)

I hope this was successfully.

Malte

Hi,
Sorry not to mention that the NIC that I referred here as NG is the onboard NIC. All PCI slots are enabled with Bus-Mastering. Anyway, since it didn't work with the Microsoft, I tested with Linux and everything was working fine.... *sigh*
I'm pretty sure it's something to do with the IRQ or memory conflict but in windows, it doesn't allow me to reassign them. Tried reshuffle the cards around to force system to reassign IRQ but the result is still the same.
Anyone has idea how to force the onboard NIC to take another IRQ or memory range?

You may need to turn all except the SCSI slot to slave and turn them one one at a time I used the 39160 without any problems. Failing that turn off the onboard LAN and try a bog standard NIC no fancy names and see if it is a hardware or software problem.

If you load W2k ensure you ooad the drivers supplied from Tyan
regards
Brendan falvey

I've trying stripping everything and install each hardware piece by piece. Installed the XP and 2000 after disabling the onboard NIC. Once NIC enabled, XP doesn't boot. 2000 shows the yellow !. Another PCI 3Com NIC works fine. Installed Slackware Linux 8.1 and everything was working perfectly fine, including the onboard NIC. Tried to remove ACPI and force windows to redetect all hardware IRQs. NG.
What's wrong with Windows??? Anyone knows any other way to change the IRQ manually??

Just put together a Tyan MPX S2466N, 512MB of Crucial registered, 2xIDE hard drives, 1xCDRW, 1x DVD


I am using a Radeon All in Wonder 64 MB card.

When I turn on the system it hangs after the POST. But I can see the memory test and everything up to that point.

It works with the cheapo card from one of my work pcs.

Am I going to have to flash the BIOS?

How can I flash the BIOS if I can't see the screen?

HELP I've been up 18 hours working on this now!!

Hi, John Hill
You need to flash the bios to version 101. It addresses your current problem with Radeon card. Boot up your PC with the cheapo card and flash the bios.

Just MY two cents about the Tiger MP:

I just got the board a couple of days ago, and I have not ínstalled a Microsoft OS yet. I just migrated my old debian installation to the new system and the Machine has been running 48h+ straight, encoding mpeg4 and as a SMB and NFS server.
The only problem I have had so far is directly related to cooling, I have to leave the case open at the Moment, but an additional case fan and better CPU Coolers are already on the way. Even then it was not the CPUs that got too hot, but the graphics card could not take the heat around it.

For anyone interested, this is the Setup:

Tiger MP (of course) Bios 1.04
2x XP1700
1x512MB Infineon Reg. ECC PC2100
SBLive! 1024
Promise Ultra66
3Com Vortex 10/100 PCI NIC
Geforce2 Pro
Maxtor 80G 5400U/min
IBM IC35... 30G 7200U/min
Enermax 350W
Toshiba DVD 14x
Plextor 12x CD-RW

After reading about all your problems with M$ OSs I don't think I will even try installing one. I thought about win2k pro, just to be able to play something other than quake and ut, but after reading all thios it doesn't seem woth the hassle.

maybe you should all just install a OS with decent SMP support...

like I said, just my two cents.

regards

Felix

Felix

Apart from initial setup problems in the earlier 1.03 version bios I had very few Win2k problems. The residual problems could be attributed to cooling and power supply (fixed on the 2466 MB) problems. The residual problems were more of a nuisance hanging probably every day or so. They do like air conditioned comfort. I have similar problems on a non Tyan/Intel box.

Better power supplies provide a large fan to extract the hot air more effectively from the CPU area of the board. This was more effective than three case mounted fans

The new 2466 MB (MPX) has much more robust power, and seperated, connections to the MB.

Regards
Brendan Falvey

My system specs:

Dual Athlon MP 1900+ (1.6GHz)
Cooling: ThermalRight SK6 with YSTech Ultra 7000RPM fan
Chaintech 7KDD motherboard
Corsair Value Select ECC Registered 512MB PC2100
two Western Digital 800JB HDD
Enermax 431W PSU

Problems:
1. the onboard Promise mbFastTrack133 has some problems detecting the proper IRQ assignment. It always gives insufficient IRQ. I have to do all sort of resetting and/or manual select. See #3 for the error message. If I did manually, when it boots up, I'll either get step #2 or the CPU utilization is extremely high (i.e. Kernel times at CPU2 is always 100% and sometime the kernel times is impacted at the CPU1); I have to change this setting again.

2. When it works, it will do the following two things:
a> the CDRW drive will read along with the HDD. Then, the system will lock up under Windows XP. And, there's a system slowdowns as well.
b> the system will lock up when the drive is transferring mass amount of data under Windows XP.

3. After rebooting, I would get the following message:

Warning - FastTrack doesn't detect proper interrupts.
Please check your IRQ setup, and make sure the PCI slot support Bus Mastering Option.

----

please, I have been tinkering with this problem for days, and I still can't resolve it. It seems that the motherboard doesn't look defective and that the drives as well as the cables are in working condition.

Michael
You did not specifiy whether you had XP Home or Pro I will assume Pro. Not knowing the details of the RAID implementation I would suggest the following possible approaches. Depending upon the way the IDE RAID is implemented it may or may not work. If you can seperate the CD drive from the HDD ports on multi port PCI IDE RAID card, do so.

Check BIOS settings re PCI slots since it is possible that the BIOS has
mastering for that slot disabled and it would seem reasonable for the HDD to have a mastering access to the buss

Check you have all the latest drivers

Fresh reinstall of XP Pro
When asked for third party mass storage drivers use install driver disk provided it should be similar to other NT based systems
This should enable XP to recognise the beast that is there
hopefully manage the IRQ problem

Alternatively

Install XP on a single drive
Install the latest driver for the IDE busses (in WIN2k/NT4 this was done in SCSI devices - I know a bit like Start to shut down a system)
Install second drive and use the proper driver to configure RAID function

Brendan Falvey

Thanks for the quick response. Yes, it's XP Professional and the same problem persists with Windows 2000. The BIOS doesn't seem to have a bus mastering option for a specific slot. All it has is PCI/PNP configuration and other stuff. I tried disabling all the serial and parallel ports in the bios, so the auto detection of IRQs seems to work better. I did try reinstalling Windows XP on 160GB striping as well as redid it on one of the 80GB drive; and it doesn't seem to work at all! Now, I disconnected the 80GB from the mbFastTrack133 "lite" RAID controller, and put it on the standard AMD ATA100 controller, and it appears to work okay (I just verified that the system doesn't crash when transferring 1GB of data). If it does work, I'll leave it on the AMD IDE controller, but in the future, if I decided to get another large drive (possibly Western Digital 200GB drive), then I'll need to use the fastTrack controller (version 2.00 BIOS supports 48-bit addressing) since it's the only place on 7KDD that supports over 137GB limit. I believe that I will get this problem, again.

Furthermore, before getting Chaintech 7KDD, I had Gigabyte GA-7DPXDW (RAID). Both of my drives are connected to the promise RAID controller on the motherboard. Once in a while, the drive will stop responding and Windows will give me a "Kernel_Stack_Inpage_Error" message. I know it can't possibly be my drives (since they are new) nor my cables (I tested all four ATA100 cables). Microsoft support site suggests that the motherboard may be defective since none of the resolutions matches my descriptions (the RAM is good, the drives are in great condition, the cables are properly connected, etc). Nevertheless, I never tested the hard drives on the AMD IDE controller when I had this board. I then RMA'd the product (Gigabyte GA-7DPXDW) back for another motherboard (Chaintech 7KDD) due to that reason and other options like ATA133 and USB 2.0.

If both motherboards were giving me similar problems, are the following possible?

1. Enermax 431W PSU. The power supply seems to work just fine powering up the system, but at burst time (i.e. turn off and on without waiting), the system will not boot up unlike many other systems that I had tried this before. It's already too late for me to exchange this one to Enermax 550W PSU, if that will help at all.

2. Memory, drives, cables, CPUs are all functional, so I can't really think they are bad.

3. However, if it's only the promise raid controllerS (i.e. drives connecting to AMD IDE controller don't cause the system to lockup) that appears to be problematic (i.e. system lockups or crashes when transferring only 1GB of data), then what other factors are possible??? Besides, I used the drivers from the motherboard CD and did try from the Promise website.

Thanks for your time in responding and reading my post. I'm really appreciated. Besides, I have this system for a month, and I never have any luck with this system! This is my first time actually building my own dual processor system; I did build quite a few single processor systems without any difficulty, however.

Michael
I ended up with that model Enermax PSU and that reduced many of the heat related problem I had with th MP board. The MPX board has fan power to the board via a seperate connector that removed all instabilities.

However back to your RAID problem. The fact that the system works without the controller if I understand you correctly points the finger at the PCI RAID controller. Under NT systems (once up and running) generally require the appropriate drivers to be loaded but not necessarily so for XP. You did not state if but I presume it is that the RAID board is PnP. Once you loaded XP in bare bones reinstall the board and fire up. If it is PnP then it should be detected and XP looks for and installs or asks for the required driver. If this does not happen then the finger seems to point to a defective RAID board since the inital IRQ assignment is detected during the ESCD phase of the boot sequence.

Suggest you seek more information on fastrack. Fast track as I recall first enabled larger drives to function onolder DOS based systems. XP does not need any such support. just checkfor any restriction on the ability of BIOS to recognise the larger drives
Brendan Falvey

Hi,

The system seems to be able to transfer more files when connecting one of the drive to the standard ATA controller. However, I did see a "Kernel_Stack_Inpage_Error" in the past 6 hours but maybe the CPU was overclocked to 1656MHz from 1600MHz but I don't see why the CPU will fail at only 56MHz clock higher. Afterwards, the Promise RAID has some difficulties starting up. I have to reset the ESCD and wait for a while in order to boot to Windows XP. The XP fastTrack drivers are all taken from the motherboard CD to a floppy disk when installing XP.

From the motherboard manual, I don't even know how to disable this controller, so that I don't have to wait for the controller to assign to an IRQ or detect any drives that I have. I believe that somewhere in the line, the system seems to react with the RAID controller for no reasons at all causing the device to fail... but maybe this observation is invalid. Perhaps, will disabling the XP's FastTrack driver help to resolve this problem? I had similar problems with the RAID controller for both motherboards (Gigabyte GA-7DPXDW [I had it for more than three weeks] has onboard Promise FastTrack100 and Chaintech 7KDD [I had it for about a week] has MBFastTrack133 "Lite"). Therefore, it can't be the controller. What have I done wrong??? Currently, I already gave up using the RAID as long as the system can be stable in 24/7 on-time. Well, since Chaintech 7KDD costs around the same price as Asus A7m266-D, Tyan Tiger MPX, and MSI K7D, I don't mind getting the system to work without a RAID. I regretted that I want a RAID option in this rig after spending restless hours in forcing this system to work the way that I wanted to work.

Michael
Cannot remember where but when I started on my mini odessey on the TYyan MP board i recall seeing something about problems with Promise Raid controllers even though it was not applicable to my problem at the time. If i remember correctly the problem was with XP/Win2K . check the net for the latest drivers.
Brendan Falvey

Hello:

Thanks for the response, but .

I'm not sure if it were the promise raid. The system still crashed when connecting the drives to AMD controller. I have to wait for about ten minutes before starting the computer normally. Is there a way to test is the processor(s) defective? I believe that I got this kind of problem when I plugged in either one of the processor. What other factors are there? If the promise raid is a problem, then explain why the drivers and the motherboard are shipped to consumers? And, from the reviews, nobody except me are having this problem! This includes Gigabyte GA-7DPXDW. I haven't yet seen any Chaintech 7KDD review, but surely am I the only one with this problem? I believe so. I checked Promise site for drivers, etc. but they don't help at all! There's no way to flash the Promise bios since it's nonprogrammable ROM.

Now, I decided to place a regular fan next (turned on Hi) to my opened case to insure that the temperature isn't an issue. From the Gigabyte board, I got a core temperature readout of 60-70C but from the Chaintech, I only get around 40-50C on both CPUs when using the Copper heatsink and fan with Arctic Silver III paste. This is taken when the system temperature is around 40C-47C. Do you think that this system temperature and core temperature is a little bit too high? I think the Gigabyte board has a better measurement of core temperature since the thermal diode sticked out very closely to the CPU core, unlike the chaintech board. However, I know that the Athlon CPUs should be operational even at 65C. Should the temperature be a problem even if the system doesn't crash by any other factors except disk failure? Since I would need to wait another 10 minutes before the system actually works, I would have to believe that one of the chip is overheating. Then, if the fan idea really works, I'll get several powerful 120mm fans (~114CFM) and place them right on the to be modded left case panel (it's a Lian-Li Aluminum case, so the Aluminum sheet should be very easy to cut). What do you think will the temperature be an issue, even if the hardware is not defective? Or, maybe even the promise RAID chip needs a nice heatsink and fan since the chip might not even stand some heat, and that 10 minutes interval between shut off time and turn on time is around the time for the overheated chip to cool down. I even have a Celeron Tualatin 1.2GHz (not overclocked) that crashes very often when the core temperature is only at 53C. Well, while replying this message, my drive successfully copied 2-3GB of data when connecting to FastTrack133, and a regular fan is placed next to my rig, covering the entire motherboard region. It would fail within 1GB of data transfer before, but now it takes more of that. I hope that my system remains cool for the next day or so.

I'll do anything just to get rid of this problem and leave my system on for days. I'm fresh out of ideas, otherwise. Again, thanks for helping me. I'll let you know if you have any other ideas or the same problem persists.

After reading all your posts, I suspect defective MB, where the problem is thermal. Seems you've pretty well eliminated the CPUs by trying them as singles, HD, and HD controllers through various combinations you've used. Also, the case and CPU temps are fine, at least with open case and fan blowing in. So, if the large fan and open case make a difference, but the CPU and system temps are fine, this suggestions some other component on the MB with a thermal problem to me. I don't suspect drives as long as you've tried to use both as boot drive and had same problem. If you haven't tried each drive as boot drive on the standard IDE controller, do that as a trouble shooting step. Would also try another stick of RAM as trouble shooting measure. Otherwise, you have great components (Lian Li, WD JB drives, etc.), that also points me to the MoBo.

Perhaps try the RAM and drive trouble shooting steps, then I'd look into exchanging the MoBo at this point. My Tyan 760MP works fine, and can run for days using some similar components (you can see my post earlier in this thread for details). Of course, seems everyone has had some trouble with each of the dual proc boards, but I'd also consider someone like Asus, I think it is very hard to beat Asus reliability.

Good luck!

Michael
The Aluminium cases usually dissipate heat better try running with the covers off to test the theory. With the early Tyan board the CPU seems to run hot according to the BIOS in excess of 65 a real worry. try swopping the CPU and use only one stick of RAM. Does the ASUS board have an auxiliary power connection I believe most of my early stability problems came down to the single supply to the board.

Brendan