• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

ASUS A8N SLI Deluxe... overclocking success / oddities

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Originally posted by: sreams
I remember people saying that Asus stated the PCI was always locked at 33MHz when you overclocked, even in the 1001-1002 bios (it just didn't show it in the BIOS). Is there any reason it wouldn't?

It definitely was not locked in eaerlier BIOS revisions. Using a PCI Firewire card (for more ports) would always result in corrupted FW hard disks whenever my system was overclocked (all other things being 100% stable).

-S


Umm it is definately "lockable" in the earlier BIOSes. I am running on 1002 right now (back from 1003.005 because it would recognize my DiamondMax 10 )

All you have to do is manually set the fsb to 201 then boot into windows and use Clockgen to raise the HTT.

 
Originally posted by: jlambvo
Is there any utility that shows the PCI clock? Everest doesn't seem to.

I just reverted to the 1002 BIOS and am running with a 240FSB, 1:1 memory, 1.465vcore. My Audigy 2 seems to be working fine (currently playing Mp3's off my iPod connected to the Audigy firewire port) and I just ran SiSandra, 3dMark05, etc a few times. I think I'll leave it like this for a while, maybe my problem before was with trying to run the bus over 240.


Clockgen detects the PCI Clock.

http://www.cpuid.com/clockgen.php

And it looks like the only people who have been able to get the mem speed over 240 at 1t have used chipset vmods. Although from what I have heard BH5 can give you a little over 250 at 1t (howver these results seem to be 512mb total)


 
GuitarDaddy,

Please, how in the !%*# did you get over the 240 range with 1T timings? I have the same RAM, and everyone else who has it, including me, cannot get over about 242 mhz. What, exactly, are you doing?

Thank You,

Dave:Q
 
I used the DDR400 memory setting, loosen up timings to 2.5-3-3-8 and gave it 2.90 volts.
And dropped the HTT multi to 4x, and the CPU multi to 9.
9 x 268 = 2412mhz

I'm using the beta bios 1003.05, and using Clockgen NF4 to increase the CPU volts to 1.58 per CPU-Z and AI booster.

*IMPORTANT*
With Clockgen and AI booster both open, I use AI booster to increase the FSB. With each increase in FSB using AI booster, I immediately click the "Apply VID" button on Clockgen to reapply the higher voltage. Because each time you use AI to increase FSB it drops the voltage. (prior to bios 1003.05 rebooting also dropped the voltage). And once you get the OC you want, reboot and "don't" touch AI booster or bios settings again or it will drop the voltage.

I beleive that because of the A64's on chip memory controller, the higher CPU voltage is required to run your ram at high speeds. I think most people miss the connection here. Just upping your ddr volts is not enough to get ram to run stable over 240. I even experimented with dropping the multi's and underclocking the CPU, and I still can't get it stable with ram running over 240 without using the method described above to max the CPU volts. This highlights the connection created by the on chip controller. When you OC ram you are indirectly OC'ing the CPU regardless of CPU mhz.

I believe this connection between ram and cpu via the on chip controller, also means unless future bios revisions allow a higher voltage you will never be able to get a maximum CPU overclock while running ram 1:1. The best I can run 1:1 is just over 2.4mhz, no matter which multi's I use.

Although I can run prime stable overnight with the above mentioned settings. I am currently running the following, because I get better bench's this way. The extra 150mhz of CPU seems to more than compensate for the 35mhz drop in ram speed.

284.2mhz x 9 = 2557.6mhz w/1.58v
ram 232.5mhz 2-3-3-5 w/2.85v (DDR333 setting)

I primed this setting for 16hrs, and have been running it for 4days. It is rock stable and screaming fast.
If I run the ram in Auto mode I can get 2620mhz prime stable, but the 2558 setting with ram 233 beats it in benchmarks.

I think the biggest factor in being able to get these results is I was lucky to get a CPU that overclocks very well. Its a 3200+ week 41 chip
 
Originally posted by: Panorama
I dont understand...asus supports Rams upto 400Mhz and some of you are using 433 and 500mhz?! :s

I don't understand what you don't understand. "Official" support is DDR400 (200Mhz), but a lot of people use memory rated to run higher, like DDR500 (250 Mhz).

Anyways, I think the problem is some of us had AMD 64 boards before, including me, and have never had problems running above 240 Mhz with 1T, but now can't with the A8N-SLI Deluxe.
 
Originally posted by: GuitarDaddy
Although I can run prime stable overnight with the above mentioned settings. I am currently running the following, because I get better bench's this way. The extra 150mhz of CPU seems to more than compensate for the 35mhz drop in ram speed.

284.2mhz x 9 = 2557.6mhz w/1.58v
ram 232.5mhz 2-3-3-5 w/2.85v (DDR333 setting)

I primed this setting for 16hrs, and have been running it for 4days. It is rock stable and screaming fast.
What temperatures are you getting with that setup?
 
Another thing, if you've used another AMD64 board before, then you probably had no problems with the labeling of memory speeds -- memory speeds of DDR400 and under adjust your memory speed, assuming nominal speed of DDR400. Since it's impossible to run ram faster than the HTT frequency, speeds above DDR400 naturally raise the HTT frequency at the same time. Makes sense if you've used an AMD64 board before.
 
Originally posted by: superbooga
Originally posted by: GuitarDaddy
I can run my mem as high as 268 1T with 2.9v using 1gb OCZ plat. rev2

Is it TCCD? Seems like TCCD and the A8N-SLI Deluxe don't go that well together?

Yes, mine are the TCCD chips. I'm convinced that how high you can get your ram doesn't depend as much on the ram as it does on the overclockability of the onboard memory controller on the CPU
 
Yes, mine are the TCCD chips. I'm convinced that how high you can get your ram doesn't depend as much on the ram as it does on the overclockability of the onboard memory controller on the CPU

If that were completely true, then, my results should be pretty simialar accross platforms. However, using the same memory and CPU that gave me a stellar memory overclock before (showing that the "overclockability of [my] onboard memory controller" is fine), I am getting much less out of the A8N SLI. I can overclock the CPU to near as far, but the memory speeds leave much to be desired.

So, the conclusion in this situation, given that certain factors have remained constant, is that the problems is with the motherboard, or its BIOS, or whatever. And, since others have reported the same problems with this motherboard, I have to conclude that my CPU and and memory are still fine.

Dave
 
Originally posted by: GuitarDaddy
I can run my mem as high as 268 1T with 2.9v using 1gb OCZ plat. rev2


I am curious what your chipset voltage is. Do you have a multimeter? It might be that your chipset is getting higher voltage then ours. I measured mine with a MM and it is running at 1.52v, if yours is higher thta could explain why you are getitng a better clock.

 
The good news is that A64s are not nearly as dependant on memory timings/bandwidth as P4s, so running async at a lower divider will not hurt much.

Still, this board never should have been shipped with such a glaring problem. I guess ASUS thought no one was going to attempt to go past 240. Hopefully this can be corrected with a BIOS update. If not, I'll go to KT890 or Xpress200 later on.
 
Originally posted by: DavidHull
Yes, mine are the TCCD chips. I'm convinced that how high you can get your ram doesn't depend as much on the ram as it does on the overclockability of the onboard memory controller on the CPU

If that were completely true, then, my results should be pretty simialar accross platforms. However, using the same memory and CPU that gave me a stellar memory overclock before (showing that the "overclockability of [my] onboard memory controller" is fine), I am getting much less out of the A8N SLI. I can overclock the CPU to near as far, but the memory speeds leave much to be desired.

So, the conclusion in this situation, given that certain factors have remained constant, is that the problems is with the motherboard, or its BIOS, or whatever. And, since others have reported the same problems with this motherboard, I have to conclude that my CPU and and memory are still fine.

Dave

Great point! This is my first A64 board, so I don't have anything to compare to. I assume your previous platform was an A64? What voltages were you running on your previous board? I don't have a multi-meter, but I will try and get my hands on one, and test the chipset voltage as Jmag suggests
 
I assume your previous platform was an A64? What voltages were you running on your previous board?

My previous platform was, in fact, an A64. The 3000+ 90nm on, surprise, and MSI Neo2 Platinum. My OCZ EL Plat Rev.2 would overclock to 285 at 2.8 volts (3 4 4 10), or 275 at 2.8 volts (2.5 4 4 10). I was finally running it all at 272 x 9, the highest stable OC I could get with my 3000+ on air. The CPU needed about 1.54 volts. I'm probably going to upgrade to the 90nm 3500+ because I like that it uses lower voltage, but only when the price drops another $100.

Dave
 
So far I managed to get my 3000+ cpu (1.8ghz stock) to about 2.4ghz. I am disillusioned about the chances to get to 2.6 ghz, as mentioned in the anandtech article about overclocking winchesters. Even at the highest voltage setting (1.65V), prime95 fails within a few minutes at 275x9 = 2475mhz, and that's with the memory on auto, which ends up setting it to about 180 mhz. So it's definetly not a memory issue. The highest stable overclock (about 4 hours of prime-95) is 265x9 = 2385 with memory at 178 mhz. I haven't yet tried to set the memory divider to "DDR333", which should bring it to about 218mhz. I hope my kingston hyperx pc3200 can handle that with good timings.

My question is if 2.4ghz is typical for 3000+ winchester maximum overclock ? I mean it's not too bad, but I was hoping for 290x9 = 2610 like anandtech claimed to have reached with this cpu. btw, my cpu voltage is showing 1.568V in cpu-z, and all my cpu's stepping/revision/brand/family/model codes shown in cpu-z are exactly like anandtech's. Also, the temps at full load are 45-50C with the stock HSF.

Did anyone get a 3000+ to 2.6 ghz ?
 
First of all, thanks to all that post here!

I´m having great problems trying to OC a Amd 64 3200+ with this board.

Processor: amd 64 3200+ 90um
Board: Asus A8n-Sli Deluxe
Cooler: Zalman 7000B-Alcu
Mem: 2x512mg Samsung TCCD chips CL2
Power supply: Revoltec Chromus 450w
Graphic Card: Sparkle 6600GT PCI-E


The thing is that I can´t even get 2.2Ghz stable!
My current mem settings are:

2.5/3/3/8 2T @ 2.7v

Hyper Transfer 4x

I have bios 1003 final, and with Clockgen I can´t get Vcore past 1.450v. I´ve tried different combinations using multiplier 9x,9.5x and 10x but without success... the system reoboots if I play a game.

What am I doing wrong?

Thanks to all, and keep the good work! 🙂




 
Originally posted by: sreams
I'm currently running a 90nm 3200+ at 2500MHz (500MHz/25% overclock) with a 250MHz FSB and DDR500 with 2.5-3-3-7 timings.

Lastly... I'm pretty sure this board/BIOS has problems with Maxtor DiamondMax 10 SATA drives (the ones that support Native Command Queuing). A previous poster had said he was having trouble getting the board to consistently detect his DM10 drive, and I'm having the same issue. My boot drive is a 200GB DM10, and the system loses communication with the drive at the WinXP Welcome screen very often. Hit the reset button... and on the next boot, the drive is not detected at all. Powering the system off and then on allows the drive to be detected again, but who knows if you'll make it past that Welcome screen.

-S

I was under the impression there is no FSB with an Athlon 64 with HTT on this motherboard. Am I misinformed?

Also, yes there are definitely many problems with Maxtor SATA drives and this motherboard and the nVidia unified drivers at this time seem to be CRAP as well. There are many posts about the Maxtor drive problems if you do a search. Including this one rant I just posted:

http://forums.anandtech.com/me...=1504418&enterthread=y
 
've been reading through this thread and others on AnandTech about OCing my 3000+ but I can't seem to be able to get it over 2000 Ghz, when I go to 2100 my computer freezes and I have to reboot. Using the 1002 BIOS and ClockGen to overclock.

I've disabled Cool'n'Quite in the BIOS, but the ASUS Cool'n'Quite software is still installed.

I locked the PCI-e bus buy moving it to 101 in the BIOS and moved the HT to 201 in the BIOS to lock the PCI.

I've tried setting the RAM timing to DDR 400 and DDR 333, using both the auto settings as well as lose latency settings.

I'm using the 9x multi and tried both 4X and 3X HT settings.

Running it stock the MB and CPU idle at around 32C and after playing games or running 3DMark it raises up to about 42C my 6600 GT stay at around 41C and gives an ambient temp reading of about 31C

Do I need to flash the BIOS? Am I missing something? I'd be happy to just get it to 2.2 and have it stable.

Thanks for your help in adavnce
 
My first post!

I bought this system, with the intentions of not overclocking it. But after reading for a while on this and other forums, I decided to see what I could pull.

Thus far, I have only pushed it to 2007 MHz (9x223). Not a really hard push, but it is 200 MHz, and stable. I am interisted in getting 2.1 - 2.2 GHz results, but thats just for kicks.

Settings:
DDR333
4X HTT
223 Clock
2.5-4-4-7 (Default for this RAM)

Results (going by memory here):
PCMark04: 4000 (3500 no-OC)
3DMark05: 3200 (3100 no-OC)
3DMark03: 7300 (7100 no-OC)

I'm hoping to get it to 2.1 when I get home, and run prime95 on it for the night.
 
Don't worry about damaging your system. Push that puppy! When you push to far 95% of the time it just reboots or hangs the system and you have to hard reboot it, and lower your settings. Worst case scenario is something gets corrupted on your hard drive and you have to repair or reinstall window. And the theoretical reduced life of the cpu from overclocking say 10yrs to 7yrs shoudn't concern most people. Even if your cpu burns out in 3yrs, what will a 3000+ cost in 2008? not much. And chances are you will have upgraded by then anyway. And you got to love the Idea of running FX55+ speeds on a <$200 chip.

That 3000+ is the best overclocking winchester, I'm using the 3200+ running 2557mhz, and most of the guys I see posting 2.6+ are using 3000+'s

Heres my benchs with the 3200+ @ 2557 and 1 6600GT

Benchmarks:
PCMark04 = 4998
3DMark03 = 9405
3DMark05 = 3708 cpu=5382
Aquamark03 gfx=8727, cpu=11947, Tot=63,938
Nbench3 = 3456
 
So what's the key to get this board running 1T ram above 240mhz? It seems above 240 I have to drop to 2T. Also I have to set the DDR to 333 or whatever. So it seems I can OC the CPU no prob, it's the dang memory. I'm not sure what voltage to use so I'm using auto. Any advice would be appreciated. Sorry I'm new to the overclocking thing, but fooling around with it, and rebooting, and rebooting, and rebooting - has been very educational.
 
Back
Top