Asus Rampage II Gene at MADSHRIMPS
 

Board just arrived this afternoon, haven't had the time to make any pictures yet (but still to follow), because I wanted to quick-test the board for it's overclockability.

200/2000 doesn't work, as with most motherboards, straight out-of-the-box so more tweaking is necessary.

NOTE: Vqpi swings between 1.4 and 1.45 when set at 1.4v. Bios problem or board problem or maybe something else?

Vqpi isn't bios-related, so probably board-related, which isn't good.

I'm in need of 1.5v (bios) to run Uncore at 4GHz 32M stable, which is about 0.1v more than I need on the DFI. Quite a significant difference!

Triple channel 1GHz CL7 with 1.65v seems véry unstable: 2M runs fluently, but crashes already at 4M loop1. In single/dual channel, I don't notice too much issue :)

upped IOH voltages ? seems to do the trick on the RE II

Back to I7 roundup and looking after some double performanc on the Asrock lol.... need to get at least 6500 points for 3dmark at 280FSB I'm maxing at 4.3K lol (but shite 4:3 divider)

Yes, briefly, but no improvement on the stability.

I was able to run 32M with triple channel just a few minutes ago, but after reboot, same 4M issue :-/

After some tweaking, I seem to have gotten the same level of stability again: 32M ran perfectly. For future reference, these are the settings I used and that were stable:

Ai Overclock Tuner : Manual
Oc From Cpu Levet Up : Auto
Oc From Memory Level Up : Auto
CPU ratio setting : 20
CPU turbo power power limit: disabled

CPU configuration :
--------------------------- :
C1E support : disabled
Hardware prefetcher: enabled
Adjacent Cache Line Prefetch: enabled
Intel virtualization tech: disabled
Cpu TM function : disabled
Execute disable bit : disabled
intel HT technolgy: disabled
Active provessor cores: all
A20M: disabled
Intel C-state tech: disabled

bclk frequency: 200
pcie frequency: 100
dram frequency: DDR3-2005MHz
uclk frequency: 4009MHz
qpi link data rate: 7218MT/s

DRAM Timing Control

1ST INFORMATION
CAS# latency : : 7
RAS# to CAS# delay : : 8
RAS# PRE time : 7
RAS# ACT time : AUTO
RAS# to RAS# delay : AUTO
REF cycle time : AUTO
WRITE recovery time : AUTO
READ to PRE time : AUTO
FOUR ACT WIN time: AUTO
Back-to-back CAS delay: AUTO

2ND INFORMATION
DRAM Timing mode: 1N
DRAM Round Trip Latency on CHA: Delay 1 clock
DRAM Round Trip Latency on CHB: Delay 1 clock
DRAM Round Trip Latency on CHC: Delay 1 clock

3RD INFORMATION
WRITE to READ delay(DD) : AUTO
WRITE to READ delay(DR) : AUTO
WRITE to READ delay(SR) : AUTO
READ to WRITE delay(DD) : AUTO
READ to WRITE delay(DR) : AUTO
READ to WRITE delay(SR) : AUTO
READ to READ delay(DD) : AUTO
READ to READ delay(DR) : AUTO
READ to READ delay(SR) : AUTO
WRITE to WRITE delay(DD) : AUTO
WRITE to WRITE delay(DR) : AUTO
WRITE to WRITE delay(SR) : AUTO

EPU II phase control : full phase
CPU Load-line calibration : Enabled
CPU differential amplitude: AUTO
NB OCP: enabled
DRAM OCP: enabled

Extreme OV: disabled

CPU Voltage : 1.38125v
CPU PLL voltage : AUTO
QPI/DRAM Core voltage: 1.5v
IOH Voltage: AUTO
IOH PCIe Voltage: AUTO
ICH Voltage: AUTO
ICH PCIe Voltage: AUTO

DRAM bus voltage : 1.65681v

DRAM Ref Voltages
Data Ref voltage on CHA: auto
Ctrl Ref voltage on CHA: auto
Data Ref voltage on CHB: auto
Ctrl Ref voltage on CHB: auto
Data Ref voltage on CHC: +75.0mV
Ctrl Ref voltage on CHC: +62.5mV

CPU spread spectrum: disabled
PCIE spread spectrum: disabled
CPU Clock Skew: Delay 200ps
IOH Clock Skew: Delay 100ps

So messing with the ref volts on Channel C did the trick ? Never messed with those before, it seems you are never too old to learn :p

Are those settings Hyperpi 32M stable ?

I don't know what fixes the issue, because after a reboot, the 32M stable settings are again 4M unstable. I presume it's some kind of timing I set to AUTO that has to be configured right

Found the setting that's making my system stable:

Back-to-back CAS delay: 12

All other tested settings are back to auto :)

looks like quality and easy mounting system for the HSF:)

This board also has s775 mounting holes, by the way :-)

Since increasing the back-to-back cas delay timing is a real performance killer AND it's apparently possible to get this timing to run at AUTO on other asus motherboards, I asked Asus if it's fixable.

By the way: the Vqpi/dram isn't very stable on this motherboard. When 1.5v is set in the bios, the bios measurement tool gives me values in between 1.488v and 1.541v, which is WAY too much.

Update: I contactad Asus for more information on the timing and Vqpi issue.

Any updates Massie on the matter... will start the tests on thursday (woman permitting :p)

Nope, haven't had any feedback on this topic.

New Bios 0809

1. Support OC station
2. Improve memory compatibility
3. Supported EIST function when CPU ratio is in manual mode
4. Fix Vista loading slowly if use over 3GB RAM

Maybe that fixes it; I need my blades to test 2000mhz C7 :p

Always free to pick 'em up :-)

Having a hard time getting 200Bclock 2000mhz ram Hyperpi stable 32M ( HT on) I seem to cap at 1995 Ram and 3980 Uncore with 167Bclock...

QPI has to be at 1.5 region for 4000mhz uncore to be stable... gonna try latest bios to see if resolves anything

Max stable Bclock is 226, 227 boots, no prob but can't pass Pi 32M

12gb up next ( looking for trouble :p )

How much tweaking involved?

not much, TRFC 100 and some looser sub timings.

For 226MHz bclk?

TRCF > 100 is ... well :-/

I was talking about the 12Gb runs, 90 TRFC was an instant freeze (mind you it's the Blades together with the Corsair dominators) Back to back is at 10 for all runs, 8 is doable but is still not 100% stable.

Also I'm still on the H50 unit which is quite good for normal ops and OC's. If I hook it up to the high end watercooling kit or phase I can get higher for sure... but for a mATX mobo not bad at all... it just needs some ram fixes

For the 226Bclock I need to be over 110 PCI-E otherwise Pi freezes on me

Will game some on it this evening to see if remains stable with 12Gb...

Tested Max ram speed with the Blades (at 1.65Vdimm) Trfc 90, B2B 12

A whopping 2200mhz :)

Good cpu!

Now you can write the text too :p

working on it

Don't be silly that's my job lol

These findings shouldn't be in the review itself, but in a seperate article :)