A8N-SLI Owners

Vega

Supreme [H]ardness
Joined
Oct 12, 2004
Messages
7,115
Is it just my comp or in it's current state is this MB unstable and really bad at overclocking? In SLI mode I could not even get my 6800 GT's to Ultra speeds of 400mhz/1.1Ghz. I went back to single graphics card because the main Game I play, World of Warcraft does not support it. Anyways, I've found that even with one graphic card this board is pretty unstable and trying to overclock is a nightmare.

If I had a dime for everytime I had to reach in for the CMOS clear jumper I'd have a nice steak dinner. I have tried all versions of the bios's on Asus website including the 03 beta revisions. I can't even get the machine to post running STOCK 2-2-2-5 1T timmings on 2GB of some decient ram: http://www.newegg.com/app/ViewProductDesc.asp?description=20-220-033&depa=1. Seems everything I do to get the damn Ram out of auto DDR 333 locks something up and I have to clear the CMOS. The other half of the time the machine refuses to send out a video signal. Machine locks up, power button doesn't work, only thing that gets it back is a CMOS clear. Only way I can get it to run with any stabilty is setting 2T with 400Mhz 2-2-2-5. This obviously is a performance hit.

If the only thing I adjust OFF of Auto is 1T, it will Post a rediculous DDR 200. I've just never seen a MB so prone to just locking up constantly with memory changes, no errors no beeps just clear CMOS time. Another strange thing, if I leave HT on auto it sets it to 4x. What crap is that. I don't want to set a FX-55 5x HT CPU at 4x. If I leave everything in the BIOS on Auto and use NTune, it gets some decient overclock but it always sets it too high, it will always lock up in Prime 95 or 3DMark 05. I am confirming that the PCI bus is locked to 33Mhz.

Now something recently happened, god knows what, Windows XP Pro is hanging after the WInXP splash at a black desktop with a working white mouse cursor as the only thing appearing. I've tried system restore, tried everything possible under safe mode to get normal mode back up and running and nothing. I suspect all this constant crashing did something. Time to re-install the OS.

As far as overclocking in SLI mode with two 6800 GT's it pretty much didn't get far before it locked up no matter the speed/vcore etc etc.

I doubt also it is a PSU cause, its a 550Watt with 40AMP 12v rail keeping a minimum of 11.7V under the most stress.

So, anyone else with A8N series seeing anything similar and have comments/suggestions?
 
Is your PSU a 20 pin or 24 pin? this mobo is 24 pin, and if your overclocking i hear 20 pin doesnt work. its yours is 20 pin, get a 20 to 24 pin converter. thats what i have to do... you can get them on egg for under 10 with shipping
 
At least you can see your board in action lol Im without a cpu and a videocard :(

Vega said:
Is it just my comp or in it's current state is this MB unstable and really bad at overclocking? In SLI mode I could not even get my 6800 GT's to Ultra speeds of 400mhz/1.1Ghz. I went back to single graphics card because the main Game I play, World of Warcraft does not support it. Anyways, I've found that even with one graphic card this board is pretty unstable and trying to overclock is a nightmare.

If I had a dime for everytime I had to reach in for the CMOS clear jumper I'd have a nice steak dinner. I have tried all versions of the bios's on Asus website including the 03 beta revisions. I can't even get the machine to post running STOCK 2-2-2-5 1T timmings on 2GB of some decient ram: http://www.newegg.com/app/ViewProductDesc.asp?description=20-220-033&depa=1. Seems everything I do to get the damn Ram out of auto DDR 333 locks something up and I have to clear the CMOS. The other half of the time the machine refuses to send out a video signal. Machine locks up, power button doesn't work, only thing that gets it back is a CMOS clear. Only way I can get it to run with any stabilty is setting 2T with 400Mhz 2-2-2-5. This obviously is a performance hit.

If the only thing I adjust OFF of Auto is 1T, it will Post a rediculous DDR 200. I've just never seen a MB so prone to just locking up constantly with memory changes, no errors no beeps just clear CMOS time. Another strange thing, if I leave HT on auto it sets it to 4x. What crap is that. I don't want to set a FX-55 5x HT CPU at 4x. If I leave everything in the BIOS on Auto and use NTune, it gets some decient overclock but it always sets it too high, it will always lock up in Prime 95 or 3DMark 05. I am confirming that the PCI bus is locked to 33Mhz.

Now something recently happened, god knows what, Windows XP Pro is hanging after the WInXP splash at a black desktop with a working white mouse cursor as the only thing appearing. I've tried system restore, tried everything possible under safe mode to get normal mode back up and running and nothing. I suspect all this constant crashing did something. Time to re-install the OS.

As far as overclocking in SLI mode with two 6800 GT's it pretty much didn't get far before it locked up no matter the speed/vcore etc etc.

I doubt also it is a PSU cause, its a 550Watt with 40AMP 12v rail keeping a minimum of 11.7V under the most stress.

So, anyone else with A8N series seeing anything similar and have comments/suggestions?
 
Vega said:
Is it just my comp or in it's current state is this MB unstable and really bad at overclocking? In SLI mode I could not even get my 6800 GT's to Ultra speeds of 400mhz/1.1Ghz. I went back to single graphics card because the main Game I play, World of Warcraft does not support it. Anyways, I've found that even with one graphic card this board is pretty unstable and trying to overclock is a nightmare.

If I had a dime for everytime I had to reach in for the CMOS clear jumper I'd have a nice steak dinner. I have tried all versions of the bios's on Asus website including the 03 beta revisions. I can't even get the machine to post running STOCK 2-2-2-5 1T timmings on 2GB of some decient ram: http://www.newegg.com/app/ViewProductDesc.asp?description=20-220-033&depa=1. Seems everything I do to get the damn Ram out of auto DDR 333 locks something up and I have to clear the CMOS. The other half of the time the machine refuses to send out a video signal. Machine locks up, power button doesn't work, only thing that gets it back is a CMOS clear. Only way I can get it to run with any stabilty is setting 2T with 400Mhz 2-2-2-5. This obviously is a performance hit.

If the only thing I adjust OFF of Auto is 1T, it will Post a rediculous DDR 200. I've just never seen a MB so prone to just locking up constantly with memory changes, no errors no beeps just clear CMOS time. Another strange thing, if I leave HT on auto it sets it to 4x. What crap is that. I don't want to set a FX-55 5x HT CPU at 4x. If I leave everything in the BIOS on Auto and use NTune, it gets some decient overclock but it always sets it too high, it will always lock up in Prime 95 or 3DMark 05. I am confirming that the PCI bus is locked to 33Mhz.

Now something recently happened, god knows what, Windows XP Pro is hanging after the WInXP splash at a black desktop with a working white mouse cursor as the only thing appearing. I've tried system restore, tried everything possible under safe mode to get normal mode back up and running and nothing. I suspect all this constant crashing did something. Time to re-install the OS.

As far as overclocking in SLI mode with two 6800 GT's it pretty much didn't get far before it locked up no matter the speed/vcore etc etc.

I doubt also it is a PSU cause, its a 550Watt with 40AMP 12v rail keeping a minimum of 11.7V under the most stress.

So, anyone else with A8N series seeing anything similar and have comments/suggestions?

Okay, I have an A8N-SLI, and until recently had similar headaches, but at a lesser degree. I'll go through the list of what I had, and what seemed to solve them:

1) Windows sometimes locking up right after the splash screen: Had this problem, at first a lot, then once in a while, now hasn't hit for over three days. First initial fix, one that got me from "a lot" to "once in a while" was to make sure each video card got their own power plugs. Initially I had a 20 pin standard PC Power & Cooling 510 Deluxe. Like you, that should be more then enough, but I found I had a lot less lockups when each six pin to dual four pin power plug for my video cards had their own molex plugs. Final solution was to order a new 24 pin PC Power & Cooling 510 SLI, other then once I haven't had the splash screen lockup. The SLI model, above having 24 pin for the mainboard power, has two six plug video card plugs for PCI-E video cards. Other then once, it hasn't hit since I put in the new PSU. I've seen someone with fairly similar lock up problems switch to an OCZ 520 watt unit, I think it has a 24 pin motherboard connector and at least one 6 pin video card connector, and apparently they are now fine too.

2) Motherboard no longer sees drives on the NVidia SATA ports if I use any of the 1003.00x series beta BIOS. No solution, other then to stick with 1002 for now. I've tried 1003.002, 1003.004 and 1003.005. I would assume 1003.001 and 1003.003 would do the same, not going to try a new BIOS until 1004 or whatever the next standard release will be.

3) Motherboard doesn't like memory at RAM timings of 2-2-2-5, 1t. My memory is 2 x 512mb of Corsair XL Pro, PC 3200. Stock timings are 2-2-2-5, 1T, I can post at that, but I fail Memtest 86+ if I use that. Solution, use 2-3-2-5, 1T. My MSI K8N Neo 2 Platinum, a Nforce 3 250 ultra socket 939 board was the same, the memory came from that board. I'd call it slightly bad memory, other then I remember seeing an article on Mushkin's site suggesting that all memory, their brand or other brands, on AMD 64 Socket 939 boards may run more stable if the second number is set for N+1, where N is the factory rated latency for that number. I know some people can run at stock timings, but I can't with my memory. Anyway, your solution might be trying 2-3-2-5, 1T. Also, maybe obvious but are you undervolting your memory and don't know it? My memory takes 2.75 volts for instance, I even set to 2.8 to be sure. If I go to "auto" on voltage it will be a bit off. Again did this on my MSI board also.

4) Iffy on CPU overclock, board will fail to post, but can be set for the overclock I want and then post, just won't post again next time. Solution...err...just play around to see what I can hit. Example: Like you I have a FX-55. I tried 216 * 13, 1.6 volts, an OC I had on my MSI board. Half the time, the board would tell me it can't post, to hit "del" to enter startup, or F1 to ignore and contiune. If I hit F1, the board boots at stock speed and voltage, if I het Del, and just go to save and exit, the board then posts at that. I dropped to 210 * 13, stock voltage and everything is fine. Went to 211 * 13, same error. Now if it doesn't like over 210, or if I dumped in some extra voltage it would suddenly be happy, I don't know, haven't tried yet.

5) Can't get decent OC on my video cards: Solution: Can now get good OC with 24 pin power supply. Note that wattage in and of itself or amperage on the 12 volt isn't the answer if your PSU isn't 24 pin, and preferbly each video card doens't have it's own plug(s). I switched from a different make of the same PSU, the 510 SLI has the same wattage and amperage rating for each rail. On the older 20 pin 510 deluxe, I could maybe get 380/1000 on my two 6800GT's. With the 510 Deluxe SLI, I'm at 400/1005 and still counting.

One final comment, if you're OCing your CPU, then you may want to run the HT at 4x anyway, the general idea is to keep that at 1000 or lower. Even a mild overclock at 5x will throw it above 5x, and make things more unstable. I actually never set mine to auto, since I knew I'd do at least a mild CPU overclock, I set mine to 4x from the beginning.
 
Arvig said:
Okay, I have an A8N-SLI, and until recently had similar headaches, but at a lesser degree. I'll go through the list of what I had, and what seemed to solve them:


3) Motherboard doesn't like memory at RAM timings of 2-2-2-5, 1t. My memory is 2 x 512mb of Corsair XL Pro, PC 3200. Stock timings are 2-2-2-5, 1T, I can post at that, but I fail Memtest 86+ if I use that. Solution, use 2-3-2-5, 1T. My MSI K8N Neo 2 Platinum, a Nforce 3 250 ultra socket 939 board was the same, the memory came from that board. I'd call it slightly bad memory, other then I remember seeing an article on Mushkin's site suggesting that all memory, their brand or other brands, on AMD 64 Socket 939 boards may run more stable if the second number is set for N+1, where N is the factory rated latency for that number. I know some people can run at stock timings, but I can't with my memory. Anyway, your solution might be trying 2-3-2-5, 1T. Also, maybe obvious but are you undervolting your memory and don't know it? My memory takes 2.75 volts for instance, I even set to 2.8 to be sure. If I go to "auto" on voltage it will be a bit off. Again did this on my MSI board also.

Why are both of you trying for 2-2-2-5 1T timing? The best timings for an A64 are 2-2-2-10 1T.
 
1. TccD memory above 220 run at 2.5-3-3-6 at 2.65V
2. Get a big amperage (at 12V rail) psu, such as Zippy 700W with 45A
3. Open the chipset hsf and see what revision of nforce4 it is. Rev.01 and 02 are buggy.
I tried A8N SLI with almost the same config. as yours (vega) with phase changer on fx55 and got the same problems. 1 day was enough and went back to neo2.
 
CZ100B said:
Why are both of you trying for 2-2-2-5 1T timing? The best timings for an A64 are 2-2-2-10 1T.


I've tried both 2-2-2-10, 1T and 2-3-2-10, 1T. 2-2-2-10, 1T was like 2-2-2-5, 1T, unstable in Memtest 86+. I know I've seen reports that running the last number at 10 on an AMD 64 system produces better results, but on my end, at least on every memory benchmark I ran, this didn't prove to be the case, so I went back to 2-3-2-5, 1T, again because I'm not stable if the second number is at "2".

Me
 
I saw another post recently at Anandtech I think it was, from a guy having RAM speed detection and general instability issues with all 4 RAM slots populated. Just for the sake of troubleshooting, my suggestion would be to try running with just 1 gig, reformatting and see how stable that is.

Vega said:
Is it just my comp or in it's current state is this MB unstable and really bad at overclocking? In SLI mode I could not even get my 6800 GT's to Ultra speeds of 400mhz/1.1Ghz. I went back to single graphics card because the main Game I play, World of Warcraft does not support it. Anyways, I've found that even with one graphic card this board is pretty unstable and trying to overclock is a nightmare.

If I had a dime for everytime I had to reach in for the CMOS clear jumper I'd have a nice steak dinner. I have tried all versions of the bios's on Asus website including the 03 beta revisions. I can't even get the machine to post running STOCK 2-2-2-5 1T timmings on 2GB of some decient ram: http://www.newegg.com/app/ViewProductDesc.asp?description=20-220-033&depa=1. Seems everything I do to get the damn Ram out of auto DDR 333 locks something up and I have to clear the CMOS. The other half of the time the machine refuses to send out a video signal. Machine locks up, power button doesn't work, only thing that gets it back is a CMOS clear. Only way I can get it to run with any stabilty is setting 2T with 400Mhz 2-2-2-5. This obviously is a performance hit.

If the only thing I adjust OFF of Auto is 1T, it will Post a rediculous DDR 200. I've just never seen a MB so prone to just locking up constantly with memory changes, no errors no beeps just clear CMOS time. Another strange thing, if I leave HT on auto it sets it to 4x. What crap is that. I don't want to set a FX-55 5x HT CPU at 4x. If I leave everything in the BIOS on Auto and use NTune, it gets some decient overclock but it always sets it too high, it will always lock up in Prime 95 or 3DMark 05. I am confirming that the PCI bus is locked to 33Mhz.

Now something recently happened, god knows what, Windows XP Pro is hanging after the WInXP splash at a black desktop with a working white mouse cursor as the only thing appearing. I've tried system restore, tried everything possible under safe mode to get normal mode back up and running and nothing. I suspect all this constant crashing did something. Time to re-install the OS.

As far as overclocking in SLI mode with two 6800 GT's it pretty much didn't get far before it locked up no matter the speed/vcore etc etc.

I doubt also it is a PSU cause, its a 550Watt with 40AMP 12v rail keeping a minimum of 11.7V under the most stress.

So, anyone else with A8N series seeing anything similar and have comments/suggestions?
 
Thanks for all of the replies. I will try some of the suggestions and see how they work!
 
This looks like an excellent start to an official "A8N-SLI Problems/Solutions" thread. I vote that this should be moved to the AMD MB section and stickied. :)
 
I would have to aggree with the guy above, I have the board coming in on tuesday with my new rig.
 
KillerTomato said:
Is your PSU a 20 pin or 24 pin? this mobo is 24 pin, and if your overclocking i hear 20 pin doesnt work. its yours is 20 pin, get a 20 to 24 pin converter. thats what i have to do... you can get them on egg for under 10 with shipping


My PSU is a 20-pin but it did come with the pin converter. Accoring to the MB Manual, as long as you have the ez-plug molex connected and a decent wattage PSU having a 20-pin line is not an issue.
 
1) Windows sometimes locking up right after the splash screen: Had this problem, at first a lot, then once in a while, now hasn't hit for over three days. First initial fix, one that got me from "a lot" to "once in a while" was to make sure each video card got their own power plugs. Initially I had a 20 pin standard PC Power & Cooling 510 Deluxe. Like you, that should be more then enough, but I found I had a lot less lockups when each six pin to dual four pin power plug for my video cards had their own molex plugs. Final solution was to order a new 24 pin PC Power & Cooling 510 SLI, other then once I haven't had the splash screen lockup. The SLI model, above having 24 pin for the mainboard power, has two six plug video card plugs for PCI-E video cards. Other then once, it hasn't hit since I put in the new PSU. I've seen someone with fairly similar lock up problems switch to an OCZ 520 watt unit, I think it has a 24 pin motherboard connector and at least one 6 pin video card connector, and apparently they are now fine too.

This is a strange one indeed. When I was running SLI I had two 12v lines dedicated to just going into the dual 6-pin PCI-E adaptors and never had a problem. I only had this Windows lockup when I was running one graphics card. I don't see how one graphics card could stress my 550W PSU to the point of WIndows not loading, especially when it didn't happen during overclocked dual graphics cards. I was looking at the 510 SLI Deluxe PSU you mentioned, it only has 34 AMP's on the 12V rail which isn't better then the one I currently have. They have it at $280 retail on their website, a bit steep.
 
centvalny said:
1. TccD memory above 220 run at 2.5-3-3-6 at 2.65V
2. Get a big amperage (at 12V rail) psu, such as Zippy 700W with 45A
3. Open the chipset hsf and see what revision of nforce4 it is. Rev.01 and 02 are buggy.
I tried A8N SLI with almost the same config. as yours (vega) with phase changer on fx55 and got the same problems. 1 day was enough and went back to neo2.

In nTune I believe it picks up what chip revision you have without actually looking at the chip. Mine says A03.
 
for $280, might as well get the zippy 700W with 45A at 12V ;) Pic. here with Antec430 and Enermax660W(gold case in the middle)
psus1.JPG
 
Vega said:
This is a strange one indeed. When I was running SLI I had two 12v lines dedicated to just going into the dual 6-pin PCI-E adaptors and never had a problem. I only had this Windows lockup when I was running one graphics card. I don't see how one graphics card could stress my 550W PSU to the point of WIndows not loading, especially when it didn't happen during overclocked dual graphics cards. I was looking at the 510 SLI Deluxe PSU you mentioned, it only has 34 AMP's on the 12V rail which isn't better then the one I currently have. They have it at $280 retail on their website, a bit steep.


Odd, I was the opposite, and I've seen other posts of a similar nature. Yes the price is a bit steep, but I don't remember it being $280. When I ordered it was $225.00...yes still steep, but not quite $280.00

All I can say is that I think a lot of the cure isn't even gaining wattage though, IMO you possibly missed my point. I do think my cure was going from a 20 pin unit to a 24 pin unit, and possibly going to having a six pin power plug for each video card. Remember, in wattage and amperage I went sideways so to speak, all I did is go from a 510 Deluxe to a 510 SLI, the Deluxe and SLI models have the same wattage and amperage. For what it's worth, I have a four pin molex plugged into the motherboard, and did before too. I haven't had any lockups for almost a week now, the symptom seems gone completly.
 
BTW, if any of you FPS game online, be sure to not install nvidia's firewall. It added a minimum of 30ms lag to my connection (look up a program called ftrace and use the udp protocol) pings looked normal but everything udp was lagged. Even using the other ethernet connection, it was still lagged until I removed the software.
Just thought I would let everyone know. :cool:
 
I also do not like the nVidia Firewall. I had to set it to Anti-Hacking only or it would not show me my fileserver. After the 3rd time I had to reinstall Windows because of SATA HDD data corruption I just skipped the firewall all together.

I do not seem to have any issues with my video card, but I am running an Antec NeoPower. It comes with the 24pin power connector, and my 6800nu does not have a PCIe power connector.
 
Sib said:
I also do not like the nVidia Firewall. I had to set it to Anti-Hacking only or it would not show me my fileserver. After the 3rd time I had to reinstall Windows because of SATA HDD data corruption I just skipped the firewall all together.

I do not seem to have any issues with my video card, but I am running an Antec NeoPower. It comes with the 24pin power connector, and my 6800nu does not have a PCIe power connector.


I never even installed the NVidia firewall, but I played around with it on my MSI K8N Neo 2, and didn't like it then, so assumed I wouldn't like it now. ;)

I think you'll be more then fine, the lack of power plugs on your 6800's will make the video card issue moot, and the 24 pin of the Antec Neopower obviously fits the bill. I think it's just us with power plugs in their video cards running 6800GT's or Ultras that have a problem unless they get a pretty hefty PSU.

If I may ask, why the drive corruption? I haven't gone for a real high FSB at all, but I'm air-cooling a FX-55. What port(s) did you put your hard drives in? I put them in SATA 3 & 4 of the NVidia chipset ports, the upper two...well, upper two if my motherboard wasn't upside down, I have a Lian-Li V1000...but upper two for almost anyone else with a tower case. ;)

Anyway, reason I picked these ports was because on a lot of the NForce 3 250gb/Ultra boards, the upper two SATA ports, the one near the AGP slot were the locked ones, the lower ones next to the chipset are not. I know all four ports on our board sit next to the chipset, but I figured why not go with that, and use those ports on the off chance the bottom two aren't locked.
 
CZ100B said:
Why are both of you trying for 2-2-2-5 1T timing? The best timings for an A64 are 2-2-2-10 1T.

I disagree. I have found that 2-2-2-9 is superior to 2-2-2-10 on the a8n-sli (but only by a few mb/se cin bandwidth which could be explained by statistical error, but I have rerun the test a few times)
 
A8N SLI/OCZ 4200/FX55/2x6800 MSI/Ultra 550watt PS/VapoLS
1.Motherboard will not reboot. Have to enter bios,save and exit on reboot. Problems fixed by switching to 1003.005 beta bios. I haven't done an install with this bios to see if there are any sata raid detection problems
2. Some locks and instability not due to ocing... Fixed by going to what is supposed to be a lesser powersupply. Went from an Antec TrueControl 550W(12v rail=24 amp) to an Ultra 500watt 34amp 12v.rail.
3. Low performance on raid using nVidia controller. I'm comparing to scores I was getting with the nVidia raid on the MSI nforce 3 board.
4. Not enough Vcore for decent ocing on a Vapo
5. I have no idea how to use the pci clock settings. it's default at 100 and goes to 145 in bios. What do you all set to in relation to other settings.
6. I am hearing the sound of found hardware or installed hardware during the 3D tests. Usually after the first test in 3D 03. Also in 3D 05 but can't remember where. You know like when your board finds a printer. Anyone else get this?
 
Shogan191 said:
A8N SLI/OCZ 4200/FX55/2x6800 MSI/Ultra 550watt PS/VapoLS
1.Motherboard will not reboot. Have to enter bios,save and exit on reboot. Problems fixed by switching to 1003.005 beta bios. I haven't done an install with this bios to see if there are any sata raid detection problems
2. Some locks and instability not due to ocing... Fixed by going to what is supposed to be a lesser powersupply. Went from an Antec TrueControl 550W(12v rail=24 amp) to an Ultra 500watt 34amp 12v.rail.
3. Low performance on raid using nVidia controller. I'm comparing to scores I was getting with the nVidia raid on the MSI nforce 3 board.
4. Not enough Vcore for decent ocing on a Vapo
5. I have no idea how to use the pci clock settings. it's default at 100 and goes to 145 in bios. What do you all set to in relation to other settings.
6. I am hearing the sound of found hardware or installed hardware during the 3D tests. Usually after the first test in 3D 03. Also in 3D 05 but can't remember where. You know like when your board finds a printer. Anyone else get this?

Replying to each of your items: 1) 1003.00x have all caused my SATA ports to not see my hard drives, as stated before. I'm stuck on 1002 for now. But I haven't had any problems rebooting, just the lockups I mentioned earlier, which I haven't had in a long while now, so I do assume that is completly solved with swapping to a 24 pin PSU.

2) Could be the jump in amps on the 12 volt rail? The PC Power & Cooling is 34 amp on the 12 volt also. Maybe 24 amps just wasn't enough. I was almost stable with a 20 pin PSU with 34 amps, and I'm runing 2x 6800GT's. Might be one can use a 20 pin if using video cards that don't need an extra power plug, as long as the amps are around 34 on the 12 volt?

3) Haven't seen poor performance. If anything, it's slightly higher then expected on my end.

4) Not cooling with a Vapochill, but yes the Vcore max is 1.65 volt, so I can see how that might not be enough.

5) Err...I just set it to 101 on a lark, and went from there. Why 101? Well...seen some boards where the PCI lock wasn't in place unless set for 1 FSB higher then default, so I did the same thing with this board's PCI-E FSB setting. Otherwise no logical reason.

6) Same thing happens on my end. I noticed that the second video card temps are pretty low, unless I have been gaming or running a couple benchmarks over and over again to stress it out and confirm that my video card OC is stable. . So I'm guessing that the second video card sits almost idle until needed, and then when needed comes on...and thus the tone you mentioned, it's also the same tone one gets if one plugs something into a USB port, like a thumb drive, right? That's the tone I hear.
 
J-Mag said:
I disagree. I have found that 2-2-2-9 is superior to 2-2-2-10 on the a8n-sli (but only by a few mb/se cin bandwidth which could be explained by statistical error, but I have rerun the test a few times)


I never tried 2-2-2-9, I may try that in a while, since I didn't see any difference with 2-2-2-10...well okay, 2-3-2-10, like I said I seem to be a bit unstable setting my memory to 2-2-2-x, 2-3-2-x is fine.
 
I'm using 1003.005 and I don't have any issues with my SATA drives not being detected.

I'm not in RAID, but I'm using a WD740 and an XA52P
 
Arvig said:
I never tried 2-2-2-9, I may try that in a while, since I didn't see any difference with 2-2-2-10...well okay, 2-3-2-10, like I said I seem to be a bit unstable setting my memory to 2-2-2-x, 2-3-2-x is fine.

Yeah I saw some german website that had he same situation going. I run 2-3-2-9 @ 225 24/7 right now... I seem to be able to hit ~235mhz with 2-3-2-9 but with the memory divider I am using, it gets my proc a little to high for running doom3...

However if I am running at stock 200mhz, 2-2-2-9 works fine.
 
Xrave said:
I'm using 1003.005 and I don't have any issues with my SATA drives not being detected.

I'm not in RAID, but I'm using a WD740 and an XA52P

I have noticed my older 120gb WD SATA drives are detected in 1003.005 along with my cinvereted IDE drive, but my main drive which should have NCQ (that doesn't work) is a diamondmax10 300gb and it is not detected by 1003.005, so I run 1002.

However, I can achieve the same max overclock with both BIOSes so there is really no reason to use the Beta verisons.
 
Xrave said:
I'm using 1003.005 and I don't have any issues with my SATA drives not being detected.

I'm not in RAID, but I'm using a WD740 and an XA52P

I have noticed my older 120gb WD SATA drives are detected in 1003.005 along with my cinvereted IDE drive, but my main drive which should have NCQ (that doesn't work) is a diamondmax10 300gb and it is not detected by 1003.005, so I run 1002.

However, I can achieve the same max overclock with both BIOSes so there is really no reason to use the Beta verisons.
 
J-Mag said:
I have noticed my older 120gb WD SATA drives are detected in 1003.005 along with my cinvereted IDE drive, but my main drive which should have NCQ (that doesn't work) is a diamondmax10 300gb and it is not detected by 1003.005, so I run 1002.

However, I can achieve the same max overclock with both BIOSes so there is really no reason to use the Beta verisons.

Okay, I'm using two 250gb Diamond Max 10's w/NCQ. That must be it then, the 1003.00x beta bioses don't see a drive with NCQ.
 
I will be building a system with a A8N and a 300gb Maxtor NCQ drive in the next week and was wondering whether the problems you're experiencing w/ the 1003.x BIOSes are just on the nvidia SATA controller or on the Silicon Image controller also, since that's what I was planning on using for my HDDs (for the RAID 5, eventually).
 
I've got a Winchester 3000+ at home cooking on Prime95 at 240x9. 240mhz is about as far of an OC that the board will allow (with 1:1 memory ratio) in its current state. You may see others that say they have gotten more but may be confused since the memory divider seems to have a mind of it’s own with most bios revisions. I’ve pushed the HTT clock to 315 with reasonable stability but garbage memory performance as you might expect.

Not so well known O/C tips with this board:
1.) Use ClockGen NF4 http://www.cpuid.com/download/CG-NVNF4.zip
You can get your CPU voltage up to 1.58 using this combined with the AI tool. First set 1.60v in AI tool (Actual will be 1.5), Then set 1.5 in ClockGen NF4. Cpuid will show you the voltage of 1.58.

2.) When using AI or ClockGen Up speed in increments of 10mhz only, anymore seems to crash the board.

3.) If using a bios without a PCI lock option, enter a 201mhz in the bios before booting to lock the PCI to 33mhz. If you don't, you'll regret it. I think this will also lock your memory 1:1 too.
 
mercilor said:
I will be building a system with a A8N and a 300gb Maxtor NCQ drive in the next week and was wondering whether the problems you're experiencing w/ the 1003.x BIOSes are just on the nvidia SATA controller or on the Silicon Image controller also, since that's what I was planning on using for my HDDs (for the RAID 5, eventually).


I never tried the Silicon Image controller personally, so that might be an option right there. I THINK the silicon image controller doesn't support NCQ, so maybe that in and of itself might ironically save you if nothing else...it will be just like a normal drive to the Silicon Image SATA ports I suspect.
 
Muhadib said:
I've got a Winchester 3000+ at home cooking on Prime95 at 240x9. 240mhz is about as far of an OC that the board will allow (with 1:1 memory ratio) in its current state. You may see others that say they have gotten more but may be confused since the memory divider seems to have a mind of it’s own with most bios revisions. I’ve pushed the HTT clock to 315 with reasonable stability but garbage memory performance as you might expect.

Not so well known O/C tips with this board:
1.) Use ClockGen NF4 http://www.cpuid.com/download/CG-NVNF4.zip
You can get your CPU voltage up to 1.58 using this combined with the AI tool. First set 1.60v in AI tool (Actual will be 1.5), Then set 1.5 in ClockGen NF4. Cpuid will show you the voltage of 1.58.
.

Doesn't the bios support a VCore of 1.65 anyway? Or are you talking about making the starting defualt voltage 1.58 instead of 1.5, so that a setting of 1.65 in the bios will really be about 1.73 or so?

Me
 
Bump...err...I do have a vested interest in this thread, and I wouldn't mind it becoming at least the semi-official A8N-SLI thread. ;)
 
I have been tinkering around with this mobo (or should i say mofo) trying to get 2 NCQ Drives to run a RAID 0 and doesnt really matter what i do i tried pretty much anything i could think of. currently running 1002. everything is fine till after i upload the 3rd party RAID drivers for windows. if i uplaod the mass storage controler (or something close to that name) i get a nice BSD. if i use just the NVIDIA RAID drivers i get to the partition part of the OS install where it recognizes two drives of the same label and size. after the load and reboot to finish install, then after it POST it says drive read error. all errror messeges i get always refer to the drive. my guess is that there are some issues with some NCQ drives.
 
the nvidia raid killed my raptor!!!, was working fine then all of a sudden the machine kept locking up and rebooting, would not even run at stock, replaced the drive and reloading windows now. Right now I am feeling some buyers remorse in getting this board. When the abit one comes out I will probably jump on that now cause of all the issues this board is having.
 
still more and more problems, should i use the other serial ata controller then the nvidia one??
 
AthlonXP said:
still more and more problems, should i use the other serial ata controller then the nvidia one??


I never had the RAID kill a drive, other then the problems I mentioned with NCQ drives and the beta BIOSes 1003.00x. I did have a lot of problems initially installing Windows, RAID array or as two seperate drives before I figured out that it was due to using a 20 pin PSU, and not isolating my video cards power cords. When I isolated them with a 20 pin PSU, I only had occasional problems either in the initial install, or lockups on the Windows splash screen going into Windows. 24 pin PSU, even though it's the same wattage and amperage rating, no problems at all.

I guess my point is if you're getting problems with a Raptor, a drive without NCQ, it could possibly be your PSU? What power supply do you have?

EDIT: Just got an answer back from ASUS about NCQ drives. Well okay, they replied to my e-mail, wouldn't call it an answer.

For the record, this is what I asked:

>Whenever I try using BIOS 1003.005 or BIOS 1003.006, my motherboard no longer sees my hard
drives. It's literally as if they don't exist as far as my motherboard is concerned. I am using two Maxtor
Diamond Max 10 hard drives in a RAID 0, each hard drive supports Native Command Queuing
(NCQ). I am connected to the NVidia SATA ports 3 and 4, but this happens if I switch to NVidia
SATA ports 1 and 2. If I switch back to BIOS 1002 I am fine. Does BIOS 1003.005 and BIOS
1003.006 not support hard drives with NCQ? If so, will the next standard BIOS support hard drives
with NCQ?<

Okay, simple enough. Otherwise, trying to find out if the beta BIOSes don't support NCQ, and if the next final BIOS will.

The "answer" I got...

>Thank you for contacting ASUS Customer Service.
My name is Ailon and I would be assisting you today

Sir, Because of 1003.005 and 1003.006 version bios are beta bios . So they may have some faulty point . But i 'm sure our R&D engineers will improve it in the next normal version . Please use 1002 bios before this , Sorry for giving any inconvenience ! <

Err...yes, I know they are beta BIOS, and I know they have "some faulty point", I can't use my NCQ drives with them! I want to know if they anticipate support for NCQ with the next "normal version".

:rolleyes:

Oh well. Again, not flaming ASUS per-se, just not shocked that I got this sort of reply, seems like if one writes any mobo maker, one gets a non-answer. Oh well, back to doing what I was doing, waiting for the next "normal version" to see if I can use it and NCQ drives.
 
well I think the board I have is a flakey one, I am returning it and getting a new one. The issue I have is the machine seems to work fine but sometimes after a cold boot it will hang. I ram memtest, ran at both stock and OC, dft ram fine, and I have a neo power 480 psu.


BTW when it was working I had my winchester 3200+ at 10*260= 2.6 ghz :D
 
AthlonXP said:
well I think the board I have is a flakey one, I am returning it and getting a new one. The issue I have is the machine seems to work fine but sometimes after a cold boot it will hang. I ram memtest, ran at both stock and OC, dft ram fine, and I have a neo power 480 psu.


BTW when it was working I had my winchester 3200+ at 10*260= 2.6 ghz :D

I am not sure as to your other system components, but I would always be careful with a dual 12v rail PSU...

Not because I don't think it has enough amps in the 12v rails, but because if you happen to hook all your 12v needing components up to just one of the rails it could cause probs.

However I KNOW that 18amps is enough to power 1 Hard drive, 1 CDRom and 1 6800GT (plus all the standard components)

Your PSU has: +12V1@18A, +12V2@15A

All I am saying is be careful that you aren't putitng all the 12v requirements onto the 15A Rail...

I have tried a few differnet PSU's with my A8N-SLI setup:

w/ my 480W thermaltake (18A 12v) I see some major droops int he 12v line when the 6800GT starts kickin into gear, it goes from 12.1 to 11.7 under game load (measured with a multimeter)

w/ my 400W Zippy (30A 12V) The lowest I have seen it droop is 11.99v under game load.

It is kinda interesting that my 400W PSU can Whip my 480W PSU ( NEVER go by their wattage rating...)

Actually the 400w Zippy costs $30 more than the 480w Thermaltake, but thats besides the point.
 
Muhadib said:
I've got a Winchester 3000+ at home cooking on Prime95 at 240x9. 240mhz is about as far of an OC that the board will allow (with 1:1 memory ratio) in its current state. You may see others that say they have gotten more but may be confused since the memory divider seems to have a mind of it’s own with most bios revisions. I’ve pushed the HTT clock to 315 with reasonable stability but garbage memory performance as you might expect.

Not so well known O/C tips with this board:
1.) Use ClockGen NF4 http://www.cpuid.com/download/CG-NVNF4.zip
You can get your CPU voltage up to 1.58 using this combined with the AI tool. First set 1.60v in AI tool (Actual will be 1.5), Then set 1.5 in ClockGen NF4. Cpuid will show you the voltage of 1.58.

2.) When using AI or ClockGen Up speed in increments of 10mhz only, anymore seems to crash the board.

3.) If using a bios without a PCI lock option, enter a 201mhz in the bios before booting to lock the PCI to 33mhz. If you don't, you'll regret it. I think this will also lock your memory 1:1 too.


I agree with the use of Clockgen... However the 10mhz thing is no longer applicable. In v1.02 I can go from 1.8ghz to 2.6ghz no prob... It was the older version that was giving me crashes when using more than 7mhz.

Also this board CAN do 295mhz memory 1:1 (although only at 2t w/ 512mb (x2)).

At 1t I can only get up to 239.6 before I get memory errors... Even when running at really crappy timings.
 
Back
Top