Another Q6600 issue

[boglwe]

Here is a pic of my current settings Screeny. The pic you see is while its running prime95 torture test.

Issue 1: The temps are kind of static whether I am running at 8x400 or 9x333. I have been trying to lower the voltage and it seems to stay the same. I have a Scythe katana 2. I like this cooler because the fan does not cover the memory as seen here and here like my old Scythe Ninja did. I would like to see lower CPU voltage and I will keep trying as well. Issue is. With the side panel off of my case the temps run much cooler at around 34c. With the side panel on they jump to 45c. Here is a pic of my case. Is this an issue? I mean can these temps sit at 45c idle and 71c full load?

Issue 2: My gigabyte board bit the dust recently. I did not like it that much cuase you need so much power to run the chip. So, I got an MSI P45 NEO. The memory needs a lot of juice to run at 402.2x2 as seen here. Does anyone have a nice modestOC of about 3.0 and the correct mem settings in bios?

As all you wonderful peeps figure this one out I am going to go make some Jiffy Pop and Anchovies, as seen here.

Thanks in advance for any suggestions.

EDIT: This pic was taken 1 hour later with a bit more juice. I can live with this.

 

[JonW]

Your last screenshot looks about right with 3.2Ghz@1.29v. BTW, if you want to post screenshot just press the "print screen button on your keyboard" open paint-select (edit)-paste, then just save your image in jpeg/jpg format so that it has a small footprint.

Try 4-4-4-12-6-48-3-3-3 timing on your ram and make sure it's running at 1:1 and 1.9v-2.1v.

Looks like your preparing some tasty treat there boglwe

 

[boglwe]

The memory will not even start up at anything less then 2.2v. Could you explain your 4-4-4-12 (this part I know) 6-48-3-3-3 (this part i do not know). 1.1 is my dream, and I think that is what its at now. 8x400 mem at 1:1 5-5-5-15 is current at CPU at 1.34v idle at 40c.

The treat was amazing. Popcorn, Salt and anchovies.

 

[JonW]

Glad you enjoyed your treat Boglwe Btw, Intel just released their tjmax for our Q6600. The tjmax is 90C, You can monitor your temps via everest ultimate edition or real temp. You just have to set the proper tjmax setting you want to use so it would read the right temp. So atm, you are actually running 59C with your Scythe Ninja.
Heres the link to the tjmax:
http://www.xtremesystems.org/f...d.php?t=179044&page=95

Ok for the ram timing:
4-4-4-18 (Tras)-6 (TWR)-60 (TRFC)-TREF/Refresh Period (Set this to max the bios allows) @ 2.2v (This is the rated voltage for your ram right?)
We'll deal with the other sub timing later (boost read performance). First we make sure you can boot to this setting successfully.

Goodluck,

 

[boglwe]

Yes, thats the rated voltage of my ram. I am going to go reboot right now and add these in. As for the TJmax part of your post, I do not really understand all that. What is the 59c? How did you get to that number and why?

Thanks for all your help so far, I think we can get this puppy up to some nice temps and speed.

 

[JonW]

Originally posted by: boglwe
As for the TJmax part of your post, I do not really understand all that. What is the 59c? How did you get to that number and why?

Thanks for all your help so far, I think we can get this puppy up to some nice temps and speed.

I think this thread can explain it better than I do. Long thread
http://www.xtremesystems.org/f...howthread.php?t=179044

So how did those settings worked out for your setup?

 

[Drsignguy]

A few things to consider, I would not use "speed fan" for your temp readings. Use CoreTemp or RealTemp, a better program for Quads. You can adjust the Tj max on each.

As for the Tj max, it is the max safe temp for your chip. I would not get it anywhere near that if all possible. The lower the better.

As for your case, you seem to have a airflow problem and you need to take a look at that. By removing your case cover and changing the results drastically means your front fans ( if any ) or your rear fans are not circulating the air properly and removing it.

Try keeping your memory at 1:1 ratio and 5-5-5-15 timings.

 

[james1701]

So if the tjmax is 90 does that mean i need to offset it in coretemp to -10, so it reads tjmax 90C?

 

[Drsignguy]

Originally posted by: james1701
So if the tjmax is 90 does that mean i need to offset it in coretemp to -10, so it reads tjmax 90C?

That would be correct.

 

[james1701]

That is awesome news. All this time I was running 67C on my hottest core, and now I am at 57. This rocks. I am going back to 3.5ghz now with this new information.

 

[Drsignguy]

Just so you realize, the Tjmax for a Q6600 IS 95c....So, you would make a -5c adjustment from 100c that Realtemp/coretemp shows.

 

[james1701]

http://www.tomshardware.co.uk/...-specs,news-29460.html

This was just released a couple of days ago. Its 90C for a Q6600 GO.

 

[Drsignguy]

Originally posted by: james1701
http://www.tomshardware.co.uk/...-specs,news-29460.html

This was just released a couple of days ago. Its 90C for a Q6600 GO.

Well I'll be damned, I read the chart wrong. This is very good news.

 

[Liet]

My question regarding the release of the TJMax info:

As I understand it, for a Q6600 G0, load temps over 65c are considering 'pushing it', and can burn out or reduce the life of the chip.

Now that everyone's been updating their heat monitoring programs with the 90c TJMax, and are excited that their CPUs are running 10c cooler than previously thought... does this also mean the new 'danger' threshold is 55c?

How did we arrive at the 65c danger mark? If it was trial-and-error, and people were using temp monitoring progs with an incorrect TJMax, doesn't that mean the chips were burning out at 55c?

Basically, how do we know what the safe temps are?

 

[VirtualLarry]

I don't know why people think 65C is the max for a Q6600. I know someone that overclocked and pushed his to the point of thermal throttling, and there was no lasting damage.

 

[myocardia]

Originally posted by: Liet
How did we arrive at the 65c danger mark?

Where did you hear that 65C was the danger point? Now, it's true that the lower the temp, the higher the overclock (assuming the RAM, motherboard, and that particular CPU can handle higher), and also that lower temps equal increased CPU life. Those are the reasons most of us do what we can to keep our temps low.

 

[Idontcare]

Originally posted by: Liet
Basically, how do we know what the safe temps are?

Intel already answered that question for you.

The chips will thermal throttle before they are allowed to operate at an unsafe temperature.

The problem is voltage. As the chip's operating temperature increases the required voltage for operating stable at that temperature continues to rise as well.

This is why overclockers spend so much effort to keep the temps low, it allows us to keep the voltages low(er) for a given clockspeed.

Since voltage kills, we'd like to get as much clockspeed as possible for the voltage we send to the chip, and that means we need to keep the temps as low as possible as well (not for CPU lifetime, but for clockspeed).

 

[Liet]

myo: I don't know, it's just something I've had stuck in my head over the years, reinforced somewhat by the general >65c load temps people keep quoting for their Q6600 G0 overclocks on various forums. The Q6600 overclock sticky says "The G0 stepping chip tolerates 71 °C, so you're probably safe a few degrees above that. You can decide on your own "red line" if you disagree with my admittedly conservative numbers." I guess I decided my redline was 65 with the lowest voltage possible, since my goal is a moderate overclock without significantly reducing CPU life.

Idontcare: Thanks for the very concise and understandable explanation! A lot simpler than the enitre heat/voltage section of the Overclock sticky.

Thanks for answering, guys.

 

[Tempered81]

Originally posted by: Liet
Basically, how do we know what the safe temps are?

check out these two consecutive pages of this thread (pages 97-98):

http://www.xtremesystems.org/f...d.php?t=179044&page=97

I think a q6600 would be safe at 63C or less (90C minus 27C = 63C). If you think your gradient is nowhere near that, then sure - go higher.

Technically when you hit around ~90C (ie DTS = 0 & Tjunction = 90C) Tcasemax will be reached as well. Then your precious Q6600 will shut down, or throttle the vcore, multiplier & FSB so no permanent damage will result. This may result in degradation/electromigration.

People have run c2d's at 90C for months and had no issues. OTOH, some people have reported degradation at less than 1.4vcore.

 

[Tempered81]

71C is the max tcase temperature for g0 q6600's dictated by Intel in their "Thermal Specification". This is not the same as what coretemp/realtemp output - which is the reading from DTS which measures tjunction.

Tcase = temperature measured at the top side center of the integrated heat spreader (aka the silvery square clump of copper on top of the cpu die that is connected to the pcb & soldered to the die.) No sensor measures Tcase, so rge tested this by drilling holes into his chip and placing a digital thermocouple there, seen here:
http://www.xtremesystems.org/f...469&stc=1&d=1224802563.
And here is a video of him measuring the temperatures and comparing them to the Tjunction (coretemp/realtemp/speedfan/everest) core temps: http://www.youtube.com/watch?v=q7ua2FFByfI&fmt=18

Tjunction = temperature inside of the cpu's cores measured by DTS.

the relationship between tjunction & tcase is kinda/sorta unknown or up in the air at the moment but it is estimated to be no greater than 27C. But hey, it could be just a couple degrees celcius

 

[Tempered81]

Originally posted by: Idontcare
As the chip's operating temperature increases the required voltage for operating stable at that temperature continues to rise as well.

Very important point here. You want low temps. Ever notice how water cooling rigs usually take less vcore to keep stability in Prime?

 

[Fistandantilis]

OK so I am a bit confused now, I was/am one of those people that thought 65c and greater was a bad thing for a processor, is this untrue?
right now my processor is running ~47c-50c using some version of coretemp and that is at stock clock frequency on a Q6600 G0, under load it usually gets to be about 55-60c and I was thinking n big deal cuz it was under 65c. I do want to OC, but I am afraid of the big, bad 65c temp limit.
So I ask, what is the temp limit for the Q6600? what is the "holy shat, thats too hot" spot for this processor?
Thanks,
Joshua