X48 Rampage Formula Preview.

[GAR]

ive tried nb voltage of 1.65, its about as high as ive gone with air, ive had fsb up to 1.55 with no luck, ive ran my memory timings really high with trd of 10, even 11 with no luck, pretty much gone through the thread and tried a bit of everyones settings with this board, i was with a 790i before this board which was stable at 475fsb with no issues....this board is nice and stable at 400fsb, but i think it should be able to do 450 atleast, im missing something i know it, with a bit of help from people who have owned this board with similar hardware i think i can get somewhere.

[Zucker2k]

ive tried nb voltage of 1.65, its about as high as ive gone with air, ive had fsb up to 1.55 with no luck, ive ran my memory timings really high with trd of 10, even 11 with no luck, pretty much gone through the thread and tried a bit of everyones settings with this board, i was with a 790i before this board which was stable at 475fsb with no issues....this board is nice and stable at 400fsb, but i think it should be able to do 450 atleast, im missing something i know it, with a bit of help from people who have owned this board with similar hardware i think i can get somewhere.

Try this for starters, you can move on from there after you get stable:

Extreme Tweaker
Ai Overclock Tuner : Manual
OC From CPU Level Up : AUTO
CPU Ratio Control : Manual
- Ratio CMOS Setting : 8
FSB Frequency : 450
FSB Strap to North Bridge : 400
PCI-E Frequency: 100
DRAM Frequency: DDR2- 1080
DRAM Command Rate : 2T
DRAM CMD Skew on Channel A: AUTO
DRAM CMD Skew on Channel B: AUTO
DRAM CLK Skew on Channel A: 300
DRAM CLK Skew on Channel B: 300
DRAM Timing Control: AUTO (This is same as):

Primary: 5-5-5-15-3-AUTO (for 52)-6-3

Secondary: 8-3-5-4-6-4-6

Tertiary: 14-5-1-6-6

DRAM Static Read Control: Auto
Ai Clock Twister : Auto
Transaction Booster : Manual
CPL: 7

CPU Voltage :
CPU PLL Voltage : 1.50
North Bridge Voltage : 1.55v
DRAM Voltage : 2.10v
FSB Termination Voltage : 1.44v (You can go lower after you're stable, then test again)
South Bridge Voltage : 1.100v
Loadline Calibration : Disabled
CPU GTL Reference : AUTO
North Bridge GTL Reference : x67
DDR2 Channel A REF Voltage :
DDR2 Channel B REF Voltage :
DDR2 Controller REF Voltage :
SB 1.5V Voltage :

NB LED Selection : NB Volt
SB LED Selection : SB Volt
CPU LED Selection : CPU VOlt
Voltiminder LED :

CPU Spread Spectrum : Disabled
PCIE Spread Spectrum : Disabled

Goodluck,

Richard.

[GAR]

I tried your settings zucker, and still prime fails within seconds........it passes memtest in dos, but in windows it fails within seconds, im guessing it has something to do with the NB, maybe i need to remount it????, with 1.43v it idles at 50-52c loads at 58c..........this is very weird

[Zucker2k]

I tried your settings zucker, and still prime fails within seconds........it passes memtest in dos, but in windows it fails within seconds, im guessing it has something to do with the NB, maybe i need to remount it????, with 1.43v it idles at 50-52c loads at 58c..........this is very weird

No, don't touch anything yet, simply revert DRAM:FSB ratio to 1:1; set RAM frequency to 900. Don't touch anything else for now.

Edit: You'll probably have to up Performance Level to 8.

[RealTelstar]

No, don't touch anything yet, simply revert DRAM:FSB ratio to 1:1; set RAM frequency to 900. Don't touch anything else for now.

Or raise the nb a notch.

[hartigan]

Alright.

Well I'm still having cores crap out at around 30 minutes.

The temperatures seem a little high but I'm not sure. I topping off at around 65c on cores 2 and 3, which I guess is OK if it's only happening at peak usage circumstances.

Below is a screenie of my Voltages, temps and Memory timings so you can see what I'm running and deduce what I need to change.


FYI, since you can't see it . . . I'm running my Core and NB GTLs at 0.67x like you suggested Zucker2K. I also lowered the vFSB from the 1.520V to the 1.504.

I'm definitely open to any solutions.

Thank you!!

[Zucker2k]

Alright.

Well I'm still having cores crap out at around 30 minutes.

The temperatures seem a little high but I'm not sure. I topping off at around 65c on cores 2 and 3, which I guess is OK if it's only happening at peak usage circumstances.

Below is a screenie of my Voltages, temps and Memory timings so you can see what I'm running and deduce what I need to change.


FYI, since you can't see it . . . I'm running my Core and NB GTLs at 0.67x like you suggested Zucker2K. I also lowered the vFSB from the 1.520V to the 1.504.

I'm definitely open to any solutions.

Thank you!!

Leave all of your previous settings intact and try this instead:

Set both GTLs to x67
Reduce FSBT to 1.28

Also, if you're having core failures, then you want to play around with vcore a bit. Especially if you're running prime blend (which seems to require more vcore for 4x2Gb than small ffts).

[Vivi]

yea it gets unstable with the wrong balance of VTT.

lower your vFSB and it should fix the prob.

U Xeon ppl i can never undertand it

why u buy xeon? (as in what is your reason), just curious

[hartigan]

Leave all of your previous settings intact and try this instead:

Set both GTLs to x67
Reduce FSBT to 1.28

Also, if you're having core failures, then you want to play around with vcore a bit. Especially if you're running prime blend (which seems to require more vcore for 4x2Gb than small ffts).

Thank you very much for the consistent feedback!! I really appreciate it.

Play with as in increase? or do you mean go down and up?

yea it gets unstable with the wrong balance of VTT.

lower your vFSB and it should fix the prob.

U Xeon ppl i can never undertand it

why u buy xeon? (as in what is your reason), just curious

I got a Xeon because I was under the impression that a server chip is possibly built to handle more extreme environments (higher temps.) than the standard desktops. Also I thought perhaps a server chip could handle the extreme tuning for a longer period of time.

Thank you to you, as well, for the feedback!!

[Zucker2k]

Play with as in increase? or do you mean go down and up?

[doodah10]

yea it gets unstable with the wrong balance of VTT.

lower your vFSB and it should fix the prob.

U Xeon ppl i can never undertand it

why u buy xeon? (as in what is your reason), just curious

This is my first Xeon, there are several reasons. This is from my previous post on Pg#26 of this thread.

I built this machine with a purpose. I work where they use and teach CAD/CAM programs like Catia and Solidworks. I use a machining app called Mastercam X2 and they have a very good forum site where there is a thread on benchmarking one of the very large sample files that comes with MC. A year ago I held the record with a P5B-Dlx running a C2D 6420 OC'd to 3.5 GHz. Every so often someone would post a time 1 to 5 seconds faster than the last best time. I was in 5th place when someone posted a time 45 seconds faster going from 5 min. flat to 4 min. 15 sec. How could this be, a near 20% decrease overnight from a Dell T7400 with a slower CPU clock and RAM speed than my P5B?

Research showed me that in Mastercam the Xeon's always were better than equiv. C2D's and quads. The leader had 2 Xeon quads with a 1600Mb FSB and 12Mb cache. A friend told me how Xeon's use their cache differently. A benchmark run by a Xeon machine and similar C2Q machine with task manager running plainly showed a difference. So I built the rig in my sig. and retook the lead by 10 sec. and still have it. I have $1300 in my Rampage and the Dell guy is in second and only spent $6800 Canadian's. The real surprise was that I can run 2 instances of the benchmark simultaneously with NO DEFFERENCE in time.

Xeon's seem to run a little cooler, need a bit less voltage, cost a few dollars less, and are usually still available when the equivalent C2Q is out of stock.

I was told that all of a Xeon's cache is available for use by any one core and only 1/2 of a C2Q's cache is available to any one core. I have a screencapture of my X3350 and a Q6450 running the same benchmark and taskmanager. The Xeon has much more activity.

[GAR]

No, don't touch anything yet, simply revert DRAM:FSB ratio to 1:1; set RAM frequency to 900. Don't touch anything else for now.

Edit: You'll probably have to up Performance Level to 8.

Tried this as well, with no luck, this is very strange indeed......i have never had a board that was this bad at overclocking, and it cost $325 USD

Also what MHZ do you guys use for PCI-E??

[mikeyakame]

Thank you so much for providing that information!! Very useful.

I was wondering if you could post a screenshot of your Computer>Sensor screen in Everest. I want to compare my voltages being reported to what you have.

One thing that's really weird is my WRITE to READ Delay (S) is set to 11T in my BIOS but Everest is reporting it as 3T.


I was also wondering how you set your tREF in the BIOS. I can't find the setting anywhere.

Here's what I have running currently, trying to match your settings best I can:

AI Overclock Tuner.......................................   Manual
CPU Ratio Setting.........................................   8.0
FSB Strap to North Bridge..............................   333MHz
FSB Frequency............................................   475MHz
PCI-E Frequency..........................................   100MHz

DRAM Frequency..........................................   DDR2 - 951MHz
DRAM Command Rate....................................   2N
DRAM CMD Skew on Channel A........................   AUTO
DRAM CMD Skew on Channel B........................   AUTO
DRAM CLK Skew on Channel A.........................   Advance 300ps
DRAM CLK Skew on Channel B.........................   Advance 300ps
DRAM Timing Control.....................................   Manual

    CAS# Latency............................... 5 
    RAS# to CAS# Delay...................... 5 
    RAS# PRE Time............................. 5 
    RAS# ACT Time............................ 15
    RAS# to RAS# Delay..................... 3
    REF Cycle Time............................ 45
    WRITE Recovery Time................... 14
    READ to PRE Time........................ 5
    READ to WRITE Delay (S/D)........... 8
    WRITE to READ Delay (S).............. 11
    WRITE to READ Delay (D).............. 5
    READ to READ Delay (S)................ 4
    READ to READ Delay (D)................ 6
    WRITE to WRITE Delay (S)............ 4 
    WRITE to WRITE Delay (D)............ 6
    WRITE to PRE Delay..................... 14 
    READ to PRE Delay....................... 5
    PRE to PRE Delay......................... 1
    ALL PRE to ACT Delay................... 7 
    ALL PRE to REF Delay.................... 6

DRAM Static Read Control..............................   Disabled
AI Clock Twister..........................................    Stronger
AI Transaction Twister.................................    Manual

    Common Performance Level..............8
    Pull-Ins.........................................Disabled

CPU Voltage...............................................    1.41250V
CPU PLL Voltage..........................................    1.50V
Northbridge Voltage.....................................    1.65V
DRAM Voltage.............................................    2.10V
FSB Termination Voltage...............................    1.52V
South Bridge Voltage....................................    1.100V
SB 1.50V Voltage.........................................    1.60V
Loadline Calibration......................................     Disabled
CPU GTL Voltage REF...................................     AUTO
NB GTL Voltage REF.....................................     AUTO
DRAM Controller Voltage REF.........................     AUTO
DRAM Channel A Voltage REF........................     AUTO
DRAM Channel B Voltage REF........................     AUTO

CPU Spread Spectrum.................................     Disabled
PCI-E Spread Spectrum...............................     Disabled

I'm able to get Prime95 to run for 4 minutes until Core 4 fails.

If you have any suggestions as to what will help, I'm open to anything. I'm still unsure as to how to set the tREF.

Thanks again for sharing your settings with us!! It has been very useful.

Write to Read (Same Rank) in Everest/Memset, is tWR_RD(S) - tRD_WR(S/D) in bios. (11 - 8 = 3) . Try changing tRD_WR(S/D) and you'll see what I mean.

If you are still having stability problems at high fsb, try delaying clock skews on Channel A, Delay 50-100ps, and Advance on Channel B, 150-250ps. As you need to remember trace lengths on the board are always shorter to Channel A than B. So if you advance A & B equally then the data will arrive too early from Channel A to MCH and too late from Channel B to MCH, which will mean CPU gets data out of sync.
I've gone back to 400fsb at the moment till I find time to iron out memory read errors. It's breaking my ability to write dvd's successfully.

[mikeyakame]

Seems the P5E3 Premium bios has the option to adjust CPU CLK Skew / NB CLK Skew with 0405 bios, hopefully Asus implements this if it's possible in the Rampage bios too. I've played with it on the P45 boards, and it's definitely something that comes in handy when getting close to 500FSB. Would probably even make going higher on quad cores very possible and even easy in some cases. I know I personally need to skew NB & CPU clocks as I've nearly hit a wall which I'm stuck adjusting REF and PLL voltages to create -/+ skews on NB/CPU clocks, but this has detrimental effect on other voltages and isn't exactly foolproof.

Never increase FSB VTT unless you have no other option. As you increase the Termination voltage it has an effect on other voltages too. vNB, vCPU, and vNB/CPU GTL Ref. It should never be higher thans vCPU or vNB. Always keep it below those 2 voltages, higher VTT FSB will create problems with bringing down each clock to VSS (common ground) for falling side of clock signal. At least that's what I understand. FSB VTT needs to be at least 5-10% lower than vNB, and as vNB is raised it seems to need to have a exponentially larger gap. ie. for me 1.55v NB wants around 1.34-1.36v vFSB_VTT anymore and its unstable. Adjusting vFSB_VTT -/+ stable voltage is only useful to create delay/advance skews for NB/MCH clocks in relation to CPU clocks.

[3lfk1ng]

Did some quick overclocking today:
4.5ghz at 1.456v




Oh...and here is that desktop wallpaper for us Rampage/Maximus users I promised in the OG post.


The photo's credit goes out to my close friend Jeff.

[Dynacore]

Does anyone know if there are any big problems with using crucial ballistix tracer dimms on this board? I Just bought a 2GB kit of pc6400 from someone and they dont seem to handle the slightest oc even my crappy OCZ sticks could easily do. And i have the double sided version which should have the d9 chips.
Then i found this thread:

http://www.blazingpc.com/forum/showt...830/index.html

Do i put my old ones back in or can anyone clear this up for me?

Edit:
Not even stable at default clock on this board, seems like bad luck.

[TheVaLVe]

What is the highest voltage you can set for NB in this MB?
Im interested in buying one, but need to be sure if i can go as high as my p35 with this one. Thx

[Rickpatbrown]

Is anyone using the Kaspersky AV program bundled with this MB? Does it work with Vista 64 bit?

Dyancore - I've heard of people using them without a problem. I still have 8gigs of DDR2 800 Ballistix that I need to test out. I had memtest problems, but it was because of a USB hub. Have you manually set the timings and voltage in BIOS. You need to configure them with this board. Especially raise the voltage to 2.2v (that was Ballistix spec, check for tracer but they should be the same chips).

Valve- 1.85volts. Have fun cooking your NB

[Dynacore]

Tnx for the reply, I have indeed set the timings and voltage manually. Also put 2.2v on it without luck. Also tried different strap settings to see if it would run on the specs the guy before me had them on (1000Mhz 5-5-5-15/ 1066 5-5-5-18), but no luck. Now I put back in my old OCZ and they run on Command Rate 1N and higher speeds without problems. Really disapointed by this so if there is anything else I can try I hope someone can tell me.

[Rickpatbrown]

Can I hurt anything with GTL settings? Is lower x.67 . . . x.65 . . . x.60 more (if you know what I mean)?

I found that setting the NB GTL to .60x signifigantly increased the time that Orthos could run (from about 15 seconds to about 45 minutes). I didn't realize that I was changing NB GTL, I meant to change CPU GTL. Now I'm messing around with CPU GTL.

[Rickpatbrown]

Update your BIOS Dyancore.

What is your test for instability/stability? Can they pass memtest86+? Are they failing Orthos/Prime95 or other? Is your CPU def stable?

Post your BIOS settings and as much info as possible. Someone more knowledgable than me should be able to help you out!

[Dynacore]

My bios is updated recently I believe version 0401.
As for testing my memory I use Memtest for windows, similar to 86+. Haven't tested with other programs for this one always works best for me for quick testing. Also the CPU should still be stable at used speeds. Havent tested default CPU speeds with the new dimms yet, but it is pc6400 so mem was running on default speeds (also tried 400fsb 1:1 with no luck).
Also worth mentioning is I sometimes get this bug when I enter the bios it sais FSB Termination voltage is set at 2.0v (!!!) No idea where this comes from but happened a couple of times...

List of current stable settings with my OCZ ram:

AI Overclock Tuner....................................... Manual
CPU Ratio Setting......................................... 9.0
FSB Strap to North Bridge.............................. 400MHz
FSB Frequency......................................... ... 411MHz
PCI-E Frequency......................................... . 101MHz

DRAM Frequency......................................... . DDR2 - 822MHz
DRAM Command Rate.................................... 1N
DRAM CMD Skew on Channel A........................ AUTO
DRAM CMD Skew on Channel B........................ AUTO
DRAM CLK Skew on Channel A......................... AUTO
DRAM CLK Skew on Channel B......................... AUTO
DRAM Timing Control..................................... Manual

CAS# Latency............................... 4
RAS# to CAS# Delay...................... 4
RAS# PRE Time............................. 4
RAS# ACT Time............................ 12
Everything else..............................AUTO.

DRAM Static Read Control.............................. Disabled
AI Clock Twister.......................................... Stronger
AI Transaction Twister................................. Manual

Common Performance Level..............7
Pull-Ins.........................................Disabl ed

CPU Voltage........................................... .... 1.40V
CPU PLL Voltage.......................................... 1.60V
Northbridge Voltage..................................... 1.45V
DRAM Voltage........................................... .. 2.14V
FSB Termination Voltage............................... 1.38V
South Bridge Voltage.................................... 1.100V
SB 1.50V Voltage......................................... 1.60V
Loadline Calibration...................................... Enabled
CPU GTL Voltage REF................................... AUTO
NB GTL Voltage REF..................................... AUTO
DRAM Controller Voltage REF......................... DDR2_REF
DRAM Channel A Voltage REF........................ DDR2_REF
DRAM Channel B Voltage REF........................ DDR2_REF

CPU Spread Spectrum................................. Disabled
PCI-E Spread Spectrum............................... Disabled

It's hard for me to believe Micron D9 chips can't handle this...

[Rickpatbrown]

0403 is the most up to date BIOS you can get off of Asus's website.

Why do you think it's your memory, is it failing memtest? Get your memory stable at stock CPU speeds first. Than you can push them individually to find their limits. 1.38v is pretty low voltage for 3.7Ghz for a conroe, isn't it? Your CPU might need more than that.

Use 2N for Ballistix. Set Clock twister to normal. Lower NB to 1.4volts or even lower.

Download Memtest86+ v2.01 from http://www.memtest.org/ I found it to be a better program (it is updated).

[TheVaLVe]

Is anyone using the Kaspersky AV program bundled with this MB? Does it work with Vista 64 bit?

Dyancore - I've heard of people using them without a problem. I still have 8gigs of DDR2 800 Ballistix that I need to test out. I had memtest problems, but it was because of a USB hub. Have you manually set the timings and voltage in BIOS. You need to configure them with this board. Especially raise the voltage to 2.2v (that was Ballistix spec, check for tracer but they should be the same chips).

Valve- 1.85volts. Have fun cooking your NB

Cheers m8, ty.

[mikeyakame]

I've also found adjusting the NB GTL Ref Voltage helps stabilize the system at high fsb, usually 0.60 - 0.63x. This sets the NB core reference point. It can help with power/ground bounce from high frequency FSB clock signals near FSB holes. Also worked out DRAM Reference (not channel A / B) is the MCH/DRAM ref voltage, much like CPU GTL Ref is MCH/CPU, NB GTL Ref seems to be an intermediary reference.