E6750 Retail - Some results Air Cooled

[rob[GL]]

433FSB @ 1.360vcore seems stable thus far... 5 hours in orthos... Will let it run all night while I sleep... If it's still good when I awake, I'll start raising FSB some more.

[alpha0ne]

3.70GHz BSOD at these settings



Settled for 3.40GHz with 1.45Vcore set in bios with only 2.0Vdimm

[MacClipper]

Got a fellow forummer apparently with the same E6750 batch running his at >4GHz on water and 1.6V Vcore on P35-DQ6 (imo, strictly a no droop mobo). Guess it's benchable for his but not 24/7 practical at such high Vcore.

Ah, some screenies just posted here, 4.1GHz with water -
http://forums.vr-zone.com/showpost.p...7&postcount=71

[rob[GL]]

Woke up this morning... 433 @ 1.360v stable... 15 hours of orthos, temps at 40c now under load... Will get home and bump FSB up to 466 for more testing...

-rob

[MikalCarbine]

Ah, some screenies just posted here, 4.1GHz with water -
http://forums.vr-zone.com/showpost.p...7&postcount=71

I think us IP35 users have to come up with a vdroop mod

[rob[GL]]

Same for us P35-DS3R guys...

They have one, but it doesn't actually work, so...

[eklipze]

alpha0ne ouch. that E6750 must have failed your hopes. sell it NOW!. and another one hahah. thats one of the worst new G0 steppings i've seen so far. some are even worse. @ 1.6vcore and still can't get over 4ghz. (e6750)

i guess its all about luck. with these things

[michaLcoughliN]

hmmf. still not impresive enough to purchase. my E6300 makes 3,3ghz with ease. the only other chip ide purchase would be E6850 as long as it hits 3.5+ preferably making the 3.8 to 4ghz wonder clock.

Or if they seem to gone for shat ill be getting a quad for the sake of the xtra cores

[alpha0ne]

alpha0ne ouch. that E6750 must have failed your hopes. sell it NOW!. and another one hahah. thats one of the worst new G0 steppings i've seen so far. some are even worse. @ 1.6vcore and still can't get over 4ghz. (e6750)

i guess its all about luck. with these things

Its GONE

[rob[GL]]

Once I hit 466FSB mine starts taking a dump... Pumping 1.55v and higher, still can't make the computer stable for more than ~2 hours...

Think I might go crazy tonight, and pump 1.75vcore just to rule out the board being the limitation...

[DEVIL K-ce]

Once I hit 466FSB mine starts taking a dump... Pumping 1.55v and higher, still can't make the computer stable for more than ~2 hours...

Think I might go crazy tonight, and pump 1.75vcore just to rule out the board being the limitation...

Ok stable for 2h Orthos ...

Try 06, 05, 01 3D ...

You are use 24/7 dual core ? Im say no So 450/460FSB will be ok for 24/7 comp


Martin

[rob[GL]]

Problem is... Orthos does not show any error... Computer just shuts off.
That's not stable enough for me... Have to figure out what's causing instability... Maybe I just need to back off my O/C a bit, be happy with 3.6GHz...

[DEVIL K-ce]

Hmmm

PSU You got nice

Vcore More bump ; probably do nothing

Maybe temp reported is wrong : actual temp is higher than reporting and cpu shut down.


Martin

[PJD]

Um, sure it is... At stock settings the cpu drops to 6x multiplier (6x333=2000) instead of 8x333 @ 2666MHz, and not only that - the Vcore drops from 1.35vcore to a bout 1.1v

On top of that, I have a laser temperature gun, I'm not using CoreTemp. I'm using Intel's Thermal Analysis Tool to report temperatures. At first I did not believe reporting 15celcius either... But first I touched the heatsink... It was COLD! Then I used the laser temperature gun to test the temperature of the heatsink, it was also below ambient.

If your bedroom is 25celcius, and you pickup say a metal screwdriver, it will be colder than ambient... Where maybe a spindle of cd's or a stack of paper will be the same warm as the rest of the room.

ANYWAYS... I have some early results with my new system, just started out very easy... Went to bios, disable C1E @ Speedstep (EIST) and then changed my VDIMM from stock 1.8v to 2.2v which is what my new Patriot memories claim to need. I left Vcore @ stock 1.35v (reports 1.32-1.33 in windows) and raised to FSB up to 400MHz.

Seems easily stable at this speed, and temperatures are still amazing me after HOURS of orthos, not even breaking a sweat.

::cough:: I want what you are smoking!! Take your analogy. A screwdriver isn't connected to a power source capable of producing temps of 700F. No way a cpu can be colder than the ambient temps without phase cooling/TEC. I've done watercooling for 10 years and even water running through the circuit without the computer turned on is +4C higher than ambient. My water block is cold also, but I still see temps +8C over ambient.

Also, if by chance Intel has been salvaging that UFO at Area 51 and you are really having temps < ambient, you'd also have issues with condensation. Condensation = mobo death, fried components, bonfire in the closet, etc.

[PJD]

::cough:: I want what you are smoking!! Take your analogy. A screwdriver isn't connected to a power source capable of producing temps of 700F. No way a cpu can be colder than the ambient temps without phase cooling/TEC. I've done watercooling for 10 years and even water running through the circuit without the computer turned on is +4C higher than ambient. My water block is cold also, but I still see temps +8C over ambient.

Also, if by chance Intel has been salvaging that UFO at Area 51 and you are really having temps < ambient, you'd also have issues with condensation. Condensation = mobo death, fried components, bonfire in the closet, etc.

Sorry for posting something old too!! I didn't want to wait for all of those images to render

Anyway, it appears those cpus are cooler which means for some watercooling may not be needed.

[adamsleath]

re temperatures: the only temp reading that seems to be realistic for me is the BIOS hardware monitor.
ambient: 17C; e4300@ 1.43v
ALL idle temps:
coretemp 0.95/0.94: 18C
speedfan: 17C
TAT: 35C
BIOStemp: 27C
ASUS probe: 12C

e6850@ 1.5v temp readings in SIG.

the readings are all over the place.

The only readings that seem consistent are BIOStemps.

[rob[GL]]

Guess so... But if cpu was overheating... I would be able to feel that heat passed onto the heatsink... Yet heatsink remains cool to touch.


Anyways, I tried more Vcore, and did not work at all... I tried 1.8 in the bios (which I presume would have been more like 1.75 real...) and it got into windows and then it locked up... computer started rebooting over and over and over and over...


hehe

had to clear bios.

i tried 1.6v in bios... and it would lock up in windows.

i'm running 1.575v in bios now... prime95 (version 25.3) and it has been running without error for 3 hours. i'll keep testing at this voltage... see how far i can get

[rob[GL]]

Prime95 ran fine all night... Yet orthos would crash (and system just reboots) within 3 hours.

I dropped my multiplier from 8x back to 7x and now 466FSB is totally stable.

I guess my particular E6750 does not like 3700MHz on air... Strange that it runs orthos fine for 3 hours without error, then the PC just reboots.


This board must not be able to provide enough volts to sustain a clock speed that high.


I will be happy if I can get 3.6GHz stable.

[Spanki]

Rob, you really need to think hard about your temperatures. The temps being reported are not correct. At Idle or at load. The sooner you accept this fact, the sooner things will start making sense.

Do you realize just how much heat you're generating by running at 1.5+ vCore(1.8?? are you nuts?) ? The problem is not that:

This board must not be able to provide enough volts to sustain a clock speed that high.

...the problem is that you're generating too much heat, so your system shuts down and/or reboots.

Here's the deal (or at least parts of it)... CoreTemp, TAT, SpeedFan and other apps determine a 'psuedo-core-temperature' by reading the on-die DTS and subtracting that value from the TCaseMax (or Tjunction) value. The DTS register 'counts down' as temperatures get higher.. when it hits zero (or maybe a few C before then), the cpu will throttle itself.

You can read the specifics at this link. But note the following:

Intel defines a certain Tjunction temperature for the processor. In the case of Yonah it is 85C&#176; or 100C&#176;. First of all the program reads from a Model Specific Register (or MSR), and detects the Tjunction temperature. A different MSR contains the temperature data, this data is represented as Delta in C&#176; between current temperature and Tjunction.

...I underlined the word 'detects', because as far as I know, this is not some value that can be read directly from the cpu registers, so 'detects' is somewhat of a misnomer. I think 'determine' or 'takes a wild guess at' would be better . What it really does is look at the cpu ID and compares that to a look-up table of some hard-coded Tjunction values (maybe only 85C or 100C, and it might default to one or the other if it doesn't recognize that cpu).

So, let's make up some numbers for example purposes to demonstrate how this works...

Tjunction = 85C (<-- 'common opinion' of proper value for all earlier desktop C2D cpus)
DTS = 40C
psuedo-core-temp = 85C - 40C = 45C, reported as a Core Temp

Tjunction = 100C (<-- 'common opinion' of proper value for all earlier mobile and quad desktop cpus)
DTS = 40C
psuedo-core-temp = 100C - 40C = 60C, reported as a Core Temp

Tjunction = 111C (new value for G0 stepping Q6600??)
DTS = 40C
psuedo-core-temp = 111C - 40C = 71C, reported as a Core Temp

...the problem is that Intel changed the Tjunction value for these new G0 stepping cpus, but noone really knows yet exactly what that means. If you look at this link, you'll see that for the new G0 stepping Q6600 at least, they:

Tcase for the Intel&#174; Core™2 Quad processor Q6600 and Intel&#174; Xeon&#174; processors X3220 and X3210 on G-0 stepping has been increased by 11 oC. Tcontrol offset will remain the same relative to increase in Tcase which will help reduce acoustics

...which probably means - at the very least - that they increased the Tjunction value by 11C for the quad. If the new E6x50 C2Ds actually use a 100C Tjunction, then your temps would be 15C higher than being reported now, but we don't yet know exactly what value they use (they might use 111C or any other value, up to 127C - who knows?).

Anyway, eventually Intel will spill the beans and let folks know how to determine Core Temps for these new stepping cpus, but all the current apps are mis-reporting them. In the meantime, there is one way to still get some 'usefull' information, using CoreTemp 0.95 and that is to enable the option circled in this image:



...doing that will show you the value of the DTS as it counts down to zero. If you do that, you'll quickly see why your computer is rebooting itself with those outrageous (on air, at least) vCore settings .

Cheers.

[msgclb]

...the problem is that Intel changed the Tjunction value for these new G0 stepping cpus, but no one really knows yet exactly what that means. If you look at this link, you'll see that for the new G0 stepping Q6600 at least, they:

...which probably means - at the very least - that they increased the Tjunction value by 11C for the quad. If the new E6x50 C2Ds actually use a 100C Tjunction, then your temps would be 15C higher than being reported now, but we don't yet know exactly what value they use (they might use 111C or any other value, up to 127C - who knows?).

Anyway, eventually Intel will spill the beans and let folks know how to determine Core Temps for these new stepping cpus, but all the current apps are mis-reporting them. In the meantime, there is one way to still get some 'usefull' information, using CoreTemp 0.95 and that is to enable the option circled in this image:



...doing that will show you the value of the DTS as it counts down to zero. If you do that, you'll quickly see why your computer is rebooting itself with those outrageous (on air, at least) vCore settings .

Cheers.

Now that I've got my E6750 up and running I believe that the BIOS CPU temp might be the only correct temperature being reported.

Idle
----
BIOS CPU Temp: 26c
AI Suite CPU Temp: 17c
SpeedFan CPU Temp: 17c
SpeedFan Core 0: 22c
SpeedFan Core 1: 24c
CoreTemp 0.95 Tjunction: 85c
CoreTemp 0.95 Core 0: 22c
CoreTemp 0.95 Core 1: 24c
CoreTemp 0.95 Tjunction (Delta): 85c
CoreTemp 0.95 Core 0 (Delta): 63c to Tjunction Remaining
CoreTemp 0.95 Core 1 (Delta): 61c to Tjunction Remaining
TAT Core 0: 22c
TAT Core 1: 25c

I have my suspicion that all of the CPU temperatures are at least ~9c low except maybe my BIOS reading. I'll be keeping an eye on my CPU temperature when I put a load on it.

I would like to know what the BIOS knows about the CPU temperature that the Windows based programs don't know!

[Volcom]

Hey I am planning on getting myself a new toy probably the E6750 as it seems like they overclock more and the temps are way lower then the Q6600. My question is if I want to hit 3.6Ghz + what ram should I get will 800Mhz ddrII do?

[Lome]

Guess so... But if cpu was overheating... I would be able to feel that heat passed onto the heatsink... Yet heatsink remains cool to touch.


Anyways, I tried more Vcore, and did not work at all... I tried 1.8 in the bios (which I presume would have been more like 1.75 real...) and it got into windows and then it locked up... computer started rebooting over and over and over and over...


hehe

had to clear bios.

i tried 1.6v in bios... and it would lock up in windows.

i'm running 1.575v in bios now... prime95 (version 25.3) and it has been running without error for 3 hours. i'll keep testing at this voltage... see how far i can get

1.6 isthe highest working setting on this board (p35-ds3r)

[tylerw13]

i just got my order comfirmation from ncix and its a

S/N: BX80562Q6600SLACR


nicely done

[MikalCarbine]

My E6750 stepping is an L719A747

Currently priming at 3.4GHz stock volts, more results and screenies to come!

Edit:

3.52Ghz with 1.38v in BIOS, 1.33v from uguru in windows

[alpha0ne]

My E6750 stepping is an L719A747

Currently priming at 3.4GHz stock volts, more results and screenies to come!

Edit:

3.52Ghz with 1.38v in BIOS, 1.33v from uguru in windows

My POS L719A823 e6750 would'nt do that at even 1.59Vcore