Asus P5E3 Premium Wi-fi-AP

[Razmatazz]

Last night I set voltages all over the place trying to get FSB 450 and CPU multiplier 8.0 I would suspect that I should be able to do this with this CPU. At a multiplier of 7.5 all was well. After I bumped it to 8 for 3.6GHz I kept getting prime errors anywhere from 1 to 13 minutes. Currently FSB termination voltage at 1.3 and CPU pll at 1.56 NB 1.35 I've varied Vcore over a wide range and it didn't seem to help. Even had it up to over 1.3635 (set in BIOS) and that didn't help. All of the errors are in core 0 and 1. Dram frequecy is low and timings are relaxed in order to isolate CPU. More Vcore or other? Any suggestions to improve stability?

[voodo]

You may be overheating. Set everything to auto. Put a fan on the NB. Raise the NB voltage until it gets stable. Don't be surprise if you are 1.6 or over.

[voodo]

450 fsb is for running ram at 1800. 400 fsb is for running ram a 1600. You should be able to run 10x400=4000 easy. Maybe only thing you may have to do is give it a little move vcore volts. You will have a very fast machine be happy.

[nunor]

Many of the penryn quads seems to have fsb wall at 400fsb... i have the same problem when i set more than 400fsb

[SoulsCollective]

Some random observations -

RAM

• In general, a delay skew was needed, with more delay needed on channel A than channel B. Between 400 and 420 FSB best stability comes from Normal skew on both A and B channels. 420-440 needs a 25ps delay on channel A. 440-470 needed a 50ps delay on channel A and a 25ps delay on channel B, my board won't successfully post at higher than 470 so I can't comment above.
• Memtest is useless for testing skew. As long as the RAM will boot, Memtest will run successfully. Best test appears to be a SuperPI run, the longer the better.
• Random reboots at idle is a symptom of bad skew.
• Performance level and clock twist is limited by both FSB and RAM. My RAM (the approved 2*1GB Corsair XMP1600 CAS-7) is happy to do PL-6 at up to 420FSB, PL-7 at up to 440 and PL-8 at up to 470. Clock twist is less easy to narrow down, I seem to be able to run various settings at various FSB and RAM settings. Not much help I realise.
• The P5E3 Prem in general seems happy to run 1T timings at all FSB settings.

Board

• vTT (called FSB Term) and vPLL are needed to push high FSB on a quad, with vTT having a greater effect on stability. Between 400 and 420FSB a vTT of 1.2 and a vPLL of 1.6 were needed. Between 420 and 440FSB a vTT of 1.26 and a vPLL of 1.66 were needed. Up to 470 FSB a vTT of 1.3 and a vPLL of 1.7 were needed. I did not feel comfortable running higher voltages than this on a 45nm chip, especially as Intel lists the max safe vTT as 1.35V and many users report killing 45nm chips with vPLL settings of 1.8V.
• Setting strap to "Auto" is useless, set it manually. The board does not seem to realise that at a higher FSB setting a higher strap is needed, and instead changes strap according to the RAM speed setting you've selected. 333 strap was stable up to about 440FSB, 400 was needed above 440. Be aware that RAM dividers are linked to strap settings so your RAM OCing may suffer somewhat.
• Board is quite heat sensitive. Aftermarket cooling or directed airflow is recommended - the X48 chipset has the highest rated TDP of all Intel chipsets at ~35W.
• NB voltage appeared to have little affect on results after 1.4V was reached running 2*1GB. This may change at 4*1GB or at FSB speeds above 470.

Hope that helps

[Booj]

Hey SC

Interesting observation there regarding the skew and Superpi. With my E8500 I can get stable @ 500fsb, 1000mhz ram 8-8-8-21 1T at 2.04v vdimm, 1.52 vNB in all stablity test except superpi. I'll have a play with the skew and see what effect it has there.

[SoulsCollective]

Oh, sup mate! Any luck fiddling about with the skew? Also, are you getting random reboots at idle?

[Booj]

Oh, sup mate! Any luck fiddling about with the skew? Also, are you getting random reboots at idle?

Yeah going good mate

No I haven't had any random reboots. Do they happen with other BIOS? I've tried all the BIOS releases so far and keep going back to 0151, primarily because the 266 strap is working. At least for me

I spent about an hour on the skew today but didn't have much luck. I started fiddling with CLK skew leaving CMD skew on auto. I've read a couple of Tony's post on the issue, it'd take a lot of testing.. Anyway those clocks I listed are for benching pretty much, the brute force approach still works Still, I find it curious as to why i'd would be prime, OCCT and Wprime stable but fail in the first loop of superpi? Using XP SP3.

But it doesn't hurt to crank it a little



I would like to see finer TRFC adjustment in a future BIOS.. 10 clock increments are too large when DFI and others are offering 1 clock incremental adjustment.

[voodo]

Says in the manual in order to use the Dram Skew you need to be running Dram Command Rate at N1. If you want more confusion the beta bios 503 has added more options. We now have a Dram command skew a Dram clock skew and a Dram control skew. Now just try and figure that out. LOL Sounds more complicated that finding the right Gtl voltage. BTW i am running 466 fsb on a QX9650 with a a vtt of 1.18. My GTL Ref is 0/2 is .660 and 1/3 is .675.

[Mykou]

has anybody tried lowest value for cpu gtl's ?

I'm wondering how this could be 0.37 for first die ???

[Infa]

that wont workk... below 50 it wont probably post.

and you cant control what DIE u want the GTL on , you just got to options of GTL 0/2 and 1/3 , first are the bus gtl other are adress something, or vice verse cant really remember... but its not DIE controlled. And to low or to high values will not post the system.

[Mykou]

yes just tried and no post below 0.48 .

why asus let those values though .. to give more confusion or so what

which bios is most stable for you Infa ?

[Infa]

I'm with the 401. others are good too, but this are the most stable one, though, dividers dont work so well, just 1:2 im afraid, atleast for me.
204 work great with other dividers but dont clock, or are stable as good.

i want to try this 503 beta above post was about...seems interesting.

Im not so updated right now, have moved to a new place, and was a bit off internet before and after i got here.
and now im back on a 100Mbit connection
been a couple of weeks, and meantime my board that i've been waiting for has been released, the DFI UT X48 T3R , and i want that board so badly, but dont have the founds right now, but i will get it! this premium has been good AND bad to me, almost been pulling my hair of understanding why it doesnt to logical stuff, but it seems that not just me have this, almost all x48 boards are like this, and maybe that was the point from the manufactures, they did say x48 was to become a "enthusiast" board... so why not make it hard! like the idea....

ohhh man im wondering away with my thoughts heheh... good night.

[sauria]

I'm with the 401. others are good too, but this are the most stable one, though, dividers dont work so well, just 1:2 im afraid, atleast for me.
204 work great with other dividers but dont clock, or are stable as good.

i want to try this 503 beta above post was about...seems interesting.

Im not so updated right now, have moved to a new place, and was a bit off internet before and after i got here.
and now im back on a 100Mbit connection
been a couple of weeks, and meantime my board that i've been waiting for has been released, the DFI UT X48 T3R , and i want that board so badly, but dont have the founds right now, but i will get it! this premium has been good AND bad to me, almost been pulling my hair of understanding why it doesnt to logical stuff, but it seems that not just me have this, almost all x48 boards are like this, and maybe that was the point from the manufactures, they did say x48 was to become a "enthusiast" board... so why not make it hard! like the idea....

ohhh man im wondering away with my thoughts heheh... good night.

So this 0401 clocks your e8500 better than 0204? Tried 0401 on a Quad CPU?

[Infa]

its kind of hard to explain...

on 401, the system are more stable, but the dividers arent working correctly, therefore i cant OC my cpu to max couse, i have to go ddr2000+ on my mem with 1:2 divider, and they cant take it, they have degraded, and are now quite bad..

on 204, it behaves kind of strange on the straps, so i cant get it up in FSB, dunno whats up, i got my cpu to almost 4.8 and 540 in fsb in a quick test just when i got the board, with 151 bios.

so i cant really recommend any bios spec.. try yourself

[sauria]

Thank you Infa -- I'm on BIOS 0204 and my e8400 doesn't like high fsb with any multiplier aboove 6.5.

[djstx2003]

please help me!!! i have this problem.
i have this system:
mobo:Asus P5E3 Premium with X48 chipset,
CPU Intel QX9650(batch L744B156)
and memory: Corsair Dominator DDR3 at 1800Mhz.
i don't want to make something extreme with this.
i only want to put the memory to play at 1800mhz. so in bios i put
manual,
multiplier to 7.0,
FSB Strap to Northbridge 400,
FSB at 450,
PCI-E:100,
the memory 1800,
timings : 7,7,7,20,
command rate: 2N
and memory voltage 2.0V
all the memory settings are the stock settings.
all the other settings i leave them to auto.
also the DRAM Static..and DRAM Dynamic...i put them to disabled and load line calibration i checked to both normal and performance and CPU Spread Spectrum and PCIE Spread Spectrum also disabled.
so i have the CPU at 3.15GH and memory at 1800Mhz.
Now the problem is that the system is not stable.
if i run CPU Stability Test after 1-2 minutes says problem in TPrime module. also in Prime95 multicore after a while fails in 2 cores.
please can you advice me what settings may i put to make my system stable?
also i noticed in Real Temp that from cpu the core 1 sensor is stuck at 40c always.only if i overclock to 4Gh the sensor starts to work. in my system i have a watercooling solution from Thermaltake. but the cpu temp is idle at
45c. is it right this temp for this cpu? also if i underclock,the temp stays at
45c.
i think something is wrong here.i have bad luck with my cpu? should i RMA it?

running 0503 bios(also tested 0401,0402)

this is what i get from Cpu Stability test and Real Temp:



thank you!!!

[Mykou]

I'm with the 401. others are good too, but this are the most stable one, though, dividers dont work so well, just 1:2 im afraid, atleast for me.
204 work great with other dividers but dont clock, or are stable as good.

i want to try this 503 beta above post was about...seems interesting.

Im not so updated right now, have moved to a new place, and was a bit off internet before and after i got here.
and now im back on a 100Mbit connection
been a couple of weeks, and meantime my board that i've been waiting for has been released, the DFI UT X48 T3R , and i want that board so badly, but dont have the founds right now, but i will get it! this premium has been good AND bad to me, almost been pulling my hair of understanding why it doesnt to logical stuff, but it seems that not just me have this, almost all x48 boards are like this, and maybe that was the point from the manufactures, they did say x48 was to become a "enthusiast" board... so why not make it hard! like the idea....
ohhh man im wondering away with my thoughts heheh... good night.

thanks Infa for your thoughts lol

I do agree with your that all X48 boards seems tough to clock

sometimes , makes me wanna take back the P5K premium I gave to my son

[soapman72]

Guys,


I have a Question pertaining to Northbridge voltage and stability. If my system is stable for 4 hours under prime95, is it still possible that if I am running the absolute minimum NB voltage for stabilty, that 3D gaming might require slightly more?

I ask, because yesterday it seemed that going from 1.51v to 1.55v to my NB, got rid of the error that I was getting under prime after 1 to 2 hours. I confiremd that by dropping the NB voltage back down to 1.51v today, and I got an error after an hour. I went back up to 1.55v today, and prime was able to run for over 4 hours, just like yesterday. So far, so good, right?

Well, I finally got Crysis, and I figured that is about as good a stability test as any, since it also stresses the video card while the cpu and memory are being stressed. Well, I played it for wuite a while earlier today, but this evening, Crysis crashed, and I got the infamous "driver has stopped responding, but has recovered" error message.

I have the latest patch installed, so can you guys tell me if Crysis generally runs problem free on a stable computer? By now it should, but I don't want to keep cranking up the volts. For the heck of it, I did bump up the NB voltage to 1.57v and played for a while longer. I did not get any more crashes. I guess maybe I should lower the NB back down to 1.55 to see if I continue to have issues with that game.

I really could use some advice on this one. Since I am very close to the limit on voltage for the NB (at least what I have heared is safe for 24/7 use), is there anything else i could do to try to get the NB stable, like change the NB GTL? Could that make a difference?

[voodo]

Just add more NB voltage until it gets stable

[soapman72]

Just add more NB voltage until it gets stable

I hear you.

I guess I am wondering if there is an alternative to adding more voltage to the NB. I believe the NB is the reason the game crashed. It is stable when priming, but just on the edge. Demanding 3D games on top of loading the CPU & ram is possible pusinng it past the edge. what is the general consensus about what the Max safe voltage is for 24/7 use? I believe I am close to that max, so what does the NB GTL do? there are only 2 settings for NB GTL on this board (besides AUTO), could seting it to .61 help stability, or is that not the issue?

Sorry ti be such a NOOB, but I just don't want to burn anything up. It looks like 4gb at 1800mhz is quite a load on the NB, even at stock ram timings. I was hoping to attempt to tighten some of them, but that may not happen now.

[Infa]

I have also some gaming issues, first thought it was the RAM and Vdimm, but after playing some with vMCH and PL settings, it did affect 32m s-pi, that are a great tool to measure RAM stability. so PL and high ram freq must be pushy for the NB.

A question please if someone can aswer, been googling this an hour, and havent found it...

im Looking for a setting in BIOS that will change the TREF! i just cant find it, and Asus are known for renaming their things...please some one knows this?

regards.

[voodo]

Alternative to adding more voltage to the NB is not to overclock.

[nunor]

I tried to push my fsb to 425 using GTL 0.66/0.675, 1.24VTT, 1.64PLL, 1.34NB and 1,28vcore. After 59 minutes of prime blend test i got an error on the 2nd core. If i change the voltages a little bit i get errors in 5 minutes Seems that GTL Ref is the reason why i can't find the stability but i don't know how to work with it.

[soapman72]

I tried to push my fsb to 425 using GTL 0.66/0.675, 1.24VTT, 1.64PLL, 1.34NB and 1,28vcore. After 59 minutes of prime blend test i got an error on the 2nd core. If i change the voltages a little bit i get errors in 5 minutes Seems that GTL Ref is the reason why i can't find the stability but i don't know how to work with it.

If I read your specs right, and especially with the timings you list in your sig, you should probably be around 1.45v on the NB as a starting point. It also depends what PL you are running, but my board needed 1.45v (NB) to run 7-7-7-24 timings at 1600mhz with PL 7 and about 1.51 for PL 6.