Hating Asus A8N-SLI Premium - need alternative

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Superguy

Member
Sep 12, 2005
132
0
0
Ouch.

I bought the original mobo A8N-E and the Premium from ZZF. Never installed the A8N-E ... realized it didnn't have everything I wanted.

RMA policy isn't as good as Newegg's ... only 10 days instead of 2 weeks, but I've never had a problem with an RMA from them.
 

virtualrain

Member
Aug 7, 2005
158
0
0
I forgot to mention in my paraphrased oc'ing guide that be sure to lower the HTT multiplier so that your total HTT bus speed never exceeds much over 1000. i.e. if you are testing with HTT at 250Mhz use the 4x multiplier. I'm getting away with 4x260 so you can oc the HTT a bit at least.

 

Pl0p

Member
Sep 14, 2005
25
0
0
Same problems here.

I'll have to get into detailed testing as described in this thread here, but i've seen too many posts with similar kinds of problems that i'm afraid there is no proper solution. I think they have messed up their voltage regulators.
Are there any marking son the board showing a hardware revision or alike?
 

Superguy

Member
Sep 12, 2005
132
0
0
I'm happier with the 1008-003 beta bios ... vCore at least seems to work and I've been able to tweak a bit more. I started a thread on that, so check that out.

I do have a question. I'm still currently testing my Mushkin Extreme Performance PC4000 memory speed max. It's rated for DDR500 at 2.8v, and my chip seems to run stably at 2700MHz (9 x 300). My memory can't handle DDR600, so for best performance, would I be best running at 270 x 10 (DDR540 at 2.95v), or 300 x 9, with DDR running at the 333 ratio for around DDR440?

I might be able to tighten the timings at a lower speed.

Any thoughts?

Thanks,

Super
 

virtualrain

Member
Aug 7, 2005
158
0
0
The better choice will be to run your RAM at 270 even if you have to loosen the timings. You can run some benchmarks at both speeds if you like.
 

Superguy

Member
Sep 12, 2005
132
0
0
My RAM seems to peter out at 533. Can't pass mem test at 266.

My proc will do 285 x 9. Memory is running at DDR466.
 

virtualrain

Member
Aug 7, 2005
158
0
0
Originally posted by: Superguy
My RAM seems to peter out at 533. Can't pass mem test at 266.

My proc will do 285 x 9. Memory is running at DDR466.

If your processor will do 285x9 = 2565 then you would be better off running at 256 or 260x10 and running memory without a divder at DDR520.

However, above, you said your proc could run at 2700 MHz... Was that unstable?
 

Superguy

Member
Sep 12, 2005
132
0
0
Yeah, I used Prime95 as the test.

Started at 300 x 9 and it errored pretty quickly.

Backed it down in 5 MHz increments. Made it a little further each time.

285 was where it would run for hours.

I'll try 256 or 260 tonight when I get home. I haven't tried my ram at DDR520 yet.
 

virtualrain

Member
Aug 7, 2005
158
0
0
Originally posted by: Superguy
Yeah, I used Prime95 as the test.

Started at 300 x 9 and it errored pretty quickly.

Backed it down in 5 MHz increments. Made it a little further each time.

285 was where it would run for hours.

I'll try 256 or 260 tonight when I get home. I haven't tried my ram at DDR520 yet.

You said earlier that your RAM petered out at 533? If that's the case, it will run at 520 no problem right?

Perhaps you are confusing things... I think that based on what you've said, the best results will be to run your CPU Freq. at 260 with a CPU multiplier of x10, an HTT multiplier of x4 (to keep it under 1000) and use the DDR400 setting for RAM which is the 1:1 divder which will mean your RAM speed is 260 as well or DDR520.

If 2600 is too fast for your CPU, try 256 or 255 instead as you said your max CPU clock was 2565.
 

Superguy

Member
Sep 12, 2005
132
0
0
256 x 10 is what I ended up with.

Bandwidth difference between that and 286 x 9 but at the 333 ratio was about 300MB/s in favor of the DDR400 ratio. Other benchies in Sandra were roughly the same.

Couldn't believe the difference between 2T and 1T. Nearly 1GB/s!

So memory is running at DDR512 right now. Timings are (from CPU-Z) 3-4-3-12-16 at 1T. Any suggestions on what I should try to tighten down next?

Thanks!

Super
 

virtualrain

Member
Aug 7, 2005
158
0
0
Originally posted by: Superguy

So memory is running at DDR512 right now. Timings are (from CPU-Z) 3-4-3-12-16 at 1T. Any suggestions on what I should try to tighten down next?

Thanks!

Super

If your memory is anything like my OCZ DDR500 memory (which is rated for 3-4-4-8 but will run at 3-4-3-8), you could try the following timings... (as copied from A64Tweaker on my system)

CAS 3 (you might even try 2.5)
TRC 11
Trfc 14
RAS to CAS Trcd 4
RAS to RAS Trrd 2 (can't be set in BIOS)
Tras 8
Trp 3
Twr 2
W to R (Twtr) 2 (can't be set in BIOS)
R to W (Trtw) 3
Write CAS (Twcl) 1 (can't be set in BIOS)

Idle should be 16 clocks as reported by CPU-Z or A64 Tweaker

I suggest you make the changes incrementally in BIOS and then boot a Memtest CD and running test 5 and 7... If that works ok, then try booting Windows and running Prime. You run the risk of screwing up your Windows install if you make a bunch of timing changes and boot straight to Windows.

Keep in mind though that you are into diminishing returns now... your gains will barely be measurable and I actually found that I could set Trtw to 2 instead of 3 but it benchmarked worse.
 

niggles

Senior member
Jan 10, 2002
797
0
0
this is a really interesting thread. I am in the market for a mobo myself, but I'm looking for an older AGP because I'm not ready to upgrade my GPU. My current mobo is the Asus A7N8X Deluxe. Go back and look at the official thread on this puppy on Anandtech, it's huge! I think the first 10 pages were a discussion on what was the best bios to run and what was being fixed with each Bios. the A7N8X is now a great board, but the first year it was out the bios was crap and drivers were buggy as well. ASUS used to be known as the rock solid board supplier, but as long as we continue to play the role of test department for ASUS they will not change their communications strategies, or their production processes.