• 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.

Unofficial ASUS K8N-E thread

Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Originally posted by: otispunkmeyer
also will a 6800GT with a arctic cooling VGA silencer fit as the chipset heatsink is pretty close to the agp slot
Click to expand...

I was worried about that too. Fortunately the heatsink isn't that big - it doesn't stick out more than a PCI slot or the AGP slot itself (basing on the pictures - I don't have this board myself).
However, I wonder if there won't be any problems when replacing the heatsink with something more efficient...
From an Asus board I'd expect better layout :/
 
My 6800 GT sits on my cooler when it has a fan on it. So unless that cooler makes the rear of the card wider (not taller, aka taking up 2 slots) then I don't see it being a problem.
 
Originally posted by: xDanielx
We have exactly the same ram, and almost exaclty the same setup I don't see why you can't reach DDR 500.. I see your running 2.5-2-2-5... why? Especially if TRAS 10 has been proven to be faster.. Did you try it with 2.5-2-2-10? I might have had to use 2.5-2-3-10 to be stable.
Click to expand...

I just tried Tras=10 and benched memory speed using: memtest86+ v1.20, cachemem and SiSoft Sandra's memory benchmark.

I am finding NO IMPROVEMENT from Tras=5 or 6!

Are you guys getting any improvement?
 
Err, hold on. The AT article says Tras is RAS Precharge. But CPU-Z gives Tras and Ras Precharge as two different things...
 
From what I read elsewhere: Tras = Active Precharge Delay = Row-Active Precharge. I think the author called TRas the "RAS Precharge" by mistake, since it is most certainly a different thing.
If the performance doesn't change between Tras=5 and Tras=10 does it at least improve stability?
 
"DRAM Active" is, I belive, what the ASUS bios calls that setting which has a range from 5 to 15. That must be what you're referring to. Since none of the other settings go to 10.

Since my system is rock solid @ Tras = 5, I don't know what use Tras=10 will be. Synthetic benchmark results are exactly the same. Stability, not important, i'm already stable. 😉

You could say that I can now increase my memory mhz even more with Tras=10, but my memory just sucks (compared to other good overclocking memory) -- above 210mhz it craps out no matter what the timings are set to.

Right now I'm sitting tight at:

+375mhz (250 * 9.5)
HT multipler = x3
mem @ 5/6 (208mhz)
2.0-3-2-5-1T
 
I was really refering to Kador about running the Tras (Cycle Time) setting at 10. Because he has the same ram I do, Cruicial Ballistix 3200.

LocutusX, the difference in performance was negligable to start with (1-2 fps in games, if that), but since it was a lower setting that provides no real difference in performace (in mots cases better performance, and I have read several articles that say this) then who cares, but it is worth a shot to test. I think I first heard of this setting at hardocp.com
 
So maybe only certain RAM brands/types benefit from Tras of 10?

As for me, I can't even reproduce a 1-2 fps difference in games. Both memtest86+ and cachemem 2.65 MMX - neither of which are Windows programs, both run in their own self-contained "OS" - are showing absolutely NO difference from Tras=11, 10 to Tras=5,6. My RAM is OCZ.
 
Originally posted by: xDanielx
We have exactly the same ram, and almost exaclty the same setup I don't see why you can't reach DDR 500.. I see your running 2.5-2-2-5... why? Especially if TRAS 10 has been proven to be faster.. Did you try it with 2.5-2-2-10? I might have had to use 2.5-2-3-10 to be stable.
Click to expand...


I think I already tried 250 with such settings, but without success. Maybe should try again, but it should be a voltage issue on my side ...
 
i've been following this thread for quite a while, and everybody's comments i find helpful and informative. 🙂

i also have a 3200 NC, and just recently purchased 1GB of Winbond BH5 Mushkin 3200 Level II ram to use on my AMD 64. however, i'm a bit worried that that may not have been the right decision to make, especially since most of the people here have Crucial Ballistix or OCZ EB, all sticks that have the new Micron chips in them (and seem to review very well on AMD 64 rigs) and that seem to be made for AMD 64 (whereas the BH5s seem to be more comfortable on Athlon XP rigs).

so a couple of questions:
1. has anybody else on this thread had any experience with the Mushkin Level II RAM on the K8N-Deluxe? if so, how well does it work?

2. i understand this BH5 ram requires a lot of voltage (i've seen people had to have 3.1-3.2 to get stable OC) - should that be a major concern with this Asus board, which seems to have a wall at 2.7? that will probably get better, right?

3. should i cancel my order from mushkin and just get the ballistix from crucial? (probably could not afford the extra for the OCZ EBs...) i'm wavering back and forth on this...

thanks anybody for any help!
 
Originally posted by: Kador
Originally posted by: xDanielx
We have exactly the same ram, and almost exaclty the same setup I don't see why you can't reach DDR 500.. I see your running 2.5-2-2-5... why? Especially if TRAS 10 has been proven to be faster.. Did you try it with 2.5-2-2-10? I might have had to use 2.5-2-3-10 to be stable.
Click to expand...


I think I already tried 250 with such settings, but without success. Maybe should try again, but it should be a voltage issue on my side ...
Click to expand...


Just tried again : at 250 1:1 the bios gets mad (wrong checksum and all the stuff). I definetly need more voltage !!
 
Is anyone aware of some affordable (ie: under 175.00) RAM that will work with this MB? I have been considering the Corsair VS GB Kit 2*512 at Cas2.5.... and suggestions? Thanks for all the great info...
 
Hey everybody.

I've been following this thread long before the latest BIOS came out, back when 220 was still the wall. Well I have some questions now that things are sorta straightened out with this board.

I have Kingston Hyper-X DDR500 ram. I tried all the overclocking techniques everyone in this thread has gotten to work. The highest result I got that was stable was 2546 (268*9.5x) I had the HTT at x3 and the ram at 333Mhz. I didn't alter the timings at all. I ran SiSoft and it gave me an extremely high CPU rating. I topped out over the 3.6ghz P4 90nm. The thing I am trying to figure out is my ram got a rating of 1700 something for float/int. When I clocked the whole system down to 2200 and set the HTT to x4 and the ram to 400Mhz wiht AUTO timings, I noticed soemthing odd. I get the same CPU rating, well about .5% lower, but my Memory benchmark is at 2500 or so compared to the 1700 I got before.

Anyone know what the best timings are for this ram on this board? I am gonna try 2.5 3 3 10 1T when I get home tonight. If it turns out that this ram doesnt work well with the Athlon64 and board I am using I'll dump it to someone. I kept it from my 2.8 Prescott box when I sold it. I don't care how much it goes for really, I got it from BestBuy when the instant rebate made it 112 bucks per stick so no big loss....

Also, I'm wondering from what everyone else is doing, what is the benefit of overclocking the CPU frequency if your ram is being dumbed down to do so? Is there a performance benefit at all?

Just wondering.

Thanks in advance.
 
Mustanggb659,

I have no personal experience with Corsairs on this board (just using some cheap OCZ) but word on the street is that this board doesn't mix well with the Corsairs.

Get the good OCZ that others in this thread have reported success with (or even the ones I have, which are non-EB) or maybe the Ballistix.

Would be interesting to see how much success that one poster has with the BH-5 he's planning on trying..
 
I have OCZ High Performance PC 3500 DDR, and an AMD Athlon 64 3200+. I'd like to run my ram at a 1:1 ratio, what would be the ideal setting for my FSB and multiplier?
 
hard to say, I would set fsb to 210 and keep mutliplier at 10 (or 11 if yours is 512cache) and use clockgen to continualy raise till you hit a stability issue. After which drop memory to 333 and see if you still have issues, if you do then lower multiplier by 1 and continue raising. Or drop multiplier to 8 or 9 keep raising by a few mhz and checking stability (use prime and do the memory intense one) fine the highest fsb that you are stable (lower chipset mult to 3 from 4) then compare to stock clockspede (ie if you are 11 * 200 and reach 8 * 240 then you can probably go to 10 * 240, if you hit 250-260 then probaly will max at 9... Though memory at 1:1 does not seem to provide that much of a benefit to z64's so if you can hit a much higher clockspede with memory at 5/6 then at 1:1 that may be the better option. Which is quite likely as long as voltage is stuck at 2.7
 
Originally posted by: LocutusX
Mustanggb659,

I have no personal experience with Corsairs on this board (just using some cheap OCZ) but word on the street is that this board doesn't mix well with the Corsairs.

Get the good OCZ that others in this thread have reported success with (or even the ones I have, which are non-EB) or maybe the Ballistix.

Would be interesting to see how much success that one poster has with the BH-5 he's planning on trying..
Click to expand...

i'll let you know...the BH-5's was shipped today and should be here monday or tuesday. although i'm still wavering on either returning it and getting crucial ballistix...

argh...if only i could get them all at once and try them all out! hah. 🙂
 
How exactly do I turn up the ram frequency on this board? And what is a stable setting at 2.7 in your experience?
 
I forget what the option is called, might be "FSB Frequency" or "HTT Frequency". It has a range from 200 to 300 and you use + and - to adjust the number. Your RAM speed will be *based* on this frequency.

In the "chipset config" menu, there is an option for RAM speed. this actually controls which divider is used, nothing more. 400mhz means 1:1, 333mhz means 5:6, etc.

So if you choose HTT freq of 250 and 333mhz ram speed, actual ram speed will be = 250 * 5 / 6 = 208mhz
If you choose HTT = 250 and 400mhz ram speed, actual ram speed will remain at 250mhz
 
Originally posted by: xFusioNx
Hey everybody.

I've been following this thread long before the latest BIOS came out, back when 220 was still the wall. Well I have some questions now that things are sorta straightened out with this board.

I have Kingston Hyper-X DDR500 ram. I tried all the overclocking techniques everyone in this thread has gotten to work. The highest result I got that was stable was 2546 (268*9.5x) I had the HTT at x3 and the ram at 333Mhz. I didn't alter the timings at all. I ran SiSoft and it gave me an extremely high CPU rating. I topped out over the 3.6ghz P4 90nm. The thing I am trying to figure out is my ram got a rating of 1700 something for float/int. When I clocked the whole system down to 2200 and set the HTT to x4 and the ram to 400Mhz wiht AUTO timings, I noticed soemthing odd. I get the same CPU rating, well about .5% lower, but my Memory benchmark is at 2500 or so compared to the 1700 I got before.

Thanks in advance.
Click to expand...

Should be your 333 settings (means 5/6 FSH) along with the x.5 multiplier, I don't remember the rule about not using whoile numbers, but for sure you should avoid.

On sandra I get abot 3500 right now. Went up to 3800 but was unstable (too low voltage on ram leads to errors) on superpi.
 
the thing about x.5 multipliers *appears* to be that it reduces RAM speed (mhz) slightly. it may still work out to your advantage with respect to your particular CPU/RAM combination.

here's what happened to me:

(HTT) 250 x 9.5 = 2375mhz
Mem @ 5/6 divider

BUT... ClockGen and CPU-Z say my memory is running at 198mhz instead of 208mhz. even though memtest86+ says its at 208mhz! weird. but the sisoft mem b/w test results indicates that its more likely to be 198mhz.

different CPU to RAM divider I suppose...


edit: deleted some foolishness ;P
 
Yes, should be something like that for x.5 multipliers : FSB / 10 x multiplier x ratio. But what happens if the miltiplier is above 10 (10.5 for instance) ?
 
I beleive your formula is off by a little it is fsb/stock * mult *ratio
and you cant have the mult greater then stock (without an fx chip anyway).
 
Originally posted by: Kador
Got the ballistix Gig. Cuttently at FSB 240, HTT 720 (x3), RAM 240 (1:1) at 2.7V and 2.5-2-2-5 and some A64Tweaker opts, CPU 2520 (10.5, should try 11 to puh this babe) at 1.6V. I'd say this ram is very nice for the price. Couldn't do 250 2.5-2-2-5, I need more voltage, ASUS it's in your hand (I'll write a second email I think !!). Still its very good for me, taking in account I use 2 sticks.
Click to expand...

i know i'm replying to an old message, but how fast if you run your CPU at stock voltage? and is 240 still currently the wall for your ballistix at 1:1? that's really nice for the ballistix for those timings...

sigh. i'll test out this mushkin PC3200 level 2 BH-5 coming my way next tuesday, but i may have to RMA if there are compatibility issues or if i really need a higher vdimm threshold to do anything worthwhile. the ballistix chips just sound like so much more room to overclock, and they aren't even at their spec 2.8...

i can see that locutusx is using 5:6 FSH - but how about everybody else? is everybody else using 1:1 FSH?
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top