Rattler unboxing - XS Exclusive

[PiLsY]

Teaser @ 1.376v load

[Hondacity]

1.3x volts for 5.x ghz

sweeeet

[PiLsY]

Youre telling me . Same voltage on P8P67 Pro does about 4700mhz.

Temps are incredible too - I know these Sandy's run cool but still, its maxed at just under mid 70's running through the 8k tests (hottest temp tests) on a budget cpu only loop.

This board sure does bring back the love .

[Mescalamba]

Thats pretty amazing.. QF seems to know how to do proper OC mobo.

[PiLsY]

Id have to agree....

[chawks2]

Very nice Pilsy. Looking forward to sandy with Foxconn.

[BeepBeep2]

So you measured with a DMM right?

[Hondacity]

measuring points ..i bet he used them...

[PiLsY]

Measuring point for vcore is out of whack when using dynamic voltage. Its reading 1.6v+, and theres no way my cooling can hold those temps at that speed and voltage. I'm trying to find a measure point now but I cant get to the back of the socket so its prooving difficult.

Temperature is scaling as if it were around 1.38v though. If I switch to manual voltage 1.38v and 4.7ghz it drops around 12c off load temps (about what id expect for purely a 400mhz decrease). I would put it at 1.4v max. I've seen higher temps than this on the P8P67 Pro @ 4.95ghz 1.43v on exactly the same cooling . I was even lazy and didnt bother replacing the paste when I swapped the Rattler back in .

Cooling is an xspc rasa, EK DCP 4.0 pump @ 70% and Aquacomputer XT 360 with 3 cheap fans. Built from my spares box .

[Unseen]

Another Rattler owner here. Still getting used to it. I will upload some pics+results in a few days

[PiLsY]

Just urging a little caution if anyone uses the dynamic option. Not sure how the cpu will hold up long term. It would appear that the voltage really flies around depending on load - im seeing it get as high as 1.596dmm measured at the pads which would appear to translate to around 1.52v real voltage (finally managed to match manual voltage clock with dynamic temperature wise). Dynamic is definitely a more stable setting in terms of cpu clocking, but should really only be considered for benchmarking.

I've decided i'm going to be the guinea pig this time and am settling on a 5ghz daily clock at high volts (but good temps). I don't think i'll have any issues with cpu degradation or death and im prepared to risk my cpu on the hunch . Its a totally average cpu with the exception that itll boot at high multi's (though stability has no chance - good for suicide cpuz shots but im not interested in that at all).

[Mescalamba]

It will degrade in year or so, but probably wont die.

[PiLsY]

TBH Mesca I've been told that about every cpu i own. My 950 has been at 1.52vcore now for almost 3 years. It even actually clocks a little better now - managed to drop to 1.50 and keep stability .

My oldest cpu is an E8500 that still clocks like the day it came out the box. Thats been at 1.6v with no speedstep or c-states for what must be 4 years now.

My intention is to run the maximum vid table value - 1.52v and see what, if anything happens. Over 1.55vcore temps increase exponentially - when you get to 1.6v it starts leaking like crazy. My theory is if I stay under the tipping point and nice and cool i'll be ok. Time will tell!

[PiLsY]

New bios P04 is an improvement over the previous development bios. Improvements to memory clocking are the order of the day this time. No extra timings or features, all bug fixes and behind the scenes tweaks.

AIDA64 needs updating to read the cpu mhz properly (then itll calculate bclk and memory speed properly), but bandwidth and timings are right .

This is 50 x 100 @ 2133mhz 1.598vddr bios (1.627dmm) with Patriot hypers @ 9-8-8-27-8-74-6-6-6-24-1T. Finally P67 matching hyper clocking with P55 but with the extra bandwidth of Sandybridge .

[Unseen]

I can confirm the Blue screens when opening cpuz with mems at high frequence.
So far I am @ 52x100 , 1.47-148(dynamic +180) 32m stable on aircooling. I know i know... superpi is not the way to test if your system is stable.

[PiLsY]

For testing stability you have to use a combination of 3 things - Linx with AVX support set to all mem for 20+ loops, custom 3200mb prime blend for at least 3 full passes (about 20 hours) then custom 3200mb small fft for around 4-8 hours.

I had Linx 25 loop stable and blend 24 hour stable and it dropped in an hour of doing small fft's. Switched to testing with small FFTs and got it stable then tried blend and it failed after 7 hours. Unfortunately you need to use all 3 this time .

Havent tried superpi 32m yet but I know that doesnt push the cpu hard .

CPUz bug is annoying!

[saaya]

only happens in cpuz?
pilsy, can you get in touch with franck and help him to debug?
hes always great at getting stuff fixed quickly

sounds really good, im glad the voltage override function works fine...
if only BR had had this feature...
then again if, intel had ever provided proper VID tables and wouldnt have been so secretive and sneaky about dynamic turbo vids and all... and would allow mainboards to override those settings or ignore them... then thinks would have been fine :P

anyways, great results pilsy!

[D0m|nat0R]

Of a few days ago I own this MB have the latest BIOS and still not corrected the bug with memory
I have a 2133 CL8 Kingston 1t and it can not set it to a stable does everything I tried uncore to 1.45
PiLsY which bios do you have and what are your voltages for the memory settings
Ageis panel does not work in Windows 7 64bit Sp1 CPU Z and of course the real problem which is already said PiLsY

sorry for my english using google translate

[PiLsY]

Currently using P04 from the foxconn site.

Dont use uncore over 1.2v - you dont need it. I dont need to set over 1.109v (1.13v actual)

Yeah Aegis is 32bit windows only at the moment. They are working on fixing it though.

CPUz bug is also a priority fix.

Can you get stable at 1866 or 1600? May be better to try lower speed and reduce timings for now until the extra memory timings are available. After that higher speed / increased timings should be available.

Hope that translates ok .

[D0m|nat0R]

with uncore under 1.3V a i can't boot
What happens witn speed strap i set to disable but remains enable
what your PCH core voltage?

[webking]

I've received this mobo too. Amazing. Reached 5,2 GHz on Air, at 4,9 GHz I was on stock voltages, didn't touch any config on CPU power voltage.

See here: http://www.hardwareforyou.it/compone...a?Itemid=0#613



Test without any software tweaking, I'll benchmarking again with a fresh install!

Now I'm with the P04 std, something will be better with future releases!

With water-cooling is possible to reach better results. I'll try asap!

[PiLsY]

Yeah its a very capable board. If you get the right memory in it its a beast. If anyone has any Micron D9 GTS/GTR I'd suggest giving it a go.

The board as it is is set up for incredibly tight memory latency - this is whats causing the issues at 2133 and even 1866 in some cases. At 2133mhz its 2000mb/s faster than the P8P67 Pro. Same everything.


Dominator you have an issue if you cant boot with uncore below 1.3v. Thats well into danger levels. I'll say again - 1.2v max, but you wont need that. I've not needed over 1.13v yet on either the Asus or Foxconn.

I have all voltages at default except vccsa (0.943v), vcore and uncore (1.109v). Uncore I just leave at 1.109v in bios (1.13v actual) permanently.

[D0m|nat0R]

My mistake they want to boot I dropped voltage and stabilized memory in 1866 Article 7
I will test memory still in UD4 and pure black that I see as a problem in it or the Rattler

I too didn't touch any config on CPU power voltage

[PiLsY]

Make sure you compare AIDA64 and Maxxxmem numbers when you do as you will see the Rattler runs very high bandwidth. Too high! This is where the 2133mhz stability problems are coming from.

[CoolZone]

Hey PiLsY,

Sorry for disturbing, but I also have a Rattler and I have a really big problem with it (VCore Related). When I am trying to set any value to the VCore, it stays fixed to a value. In Dynamic Mode, the voltage jumps from 1.01V to 1.238V in full load (and this voltage does not raise no matter what I set in the BIOS (+xxmV)). Also, when I set a fixed voltage (I have tried, 1.15V, 1.20V, 1.25V, 1.35V and more values between them), all I get is a steady 1.1V. I have also tried to modify it through the Windows software application, AEGIS, and even if I perform the modification, the voltage remains the same, 1.1V). I have done measurements at the read points too with a multimeter, and the voltage reported by AEGIS and in the BIOS is correct.

Tried 4 BIOS versions so far and the board behaves exactly the same way; however, all other voltages can be set without problems at all.

If you have an idea, please let me know!