Excellent link - I hadn't seen this yet - thanks much!Quote:
Originally Posted by unclewebb
Printable View
Excellent link - I hadn't seen this yet - thanks much!Quote:
Originally Posted by unclewebb
Quote:
Originally Posted by XtremeTiramisu
4 ghz quad core.. :) now that's an OC.. :D congratz..
that is some nice prime setup.. :)
Quote:
Originally Posted by Spanki
SO HOW IN THE H ARE WE SUPPOSE TO KNOW THE TEMPS?????!!!!!!
Amazing... just amazing!! Great clocks yet AGAIN with the "Dream Quad". I tell you what, I will trade you straight up for my 6600. It clocks just a bit higher than yours....
... <Hoping you don't notice mine has an E in front of 6600 and not Q ;)>
So I guess I am the only person concerned about the fact that there is no way to tell what temps you are getting with the quads?!
Temps are meaningless. Keep track of how far you are away from the throttling point using CoreTemp 0.95 and you'll never have a problem.Quote:
Originally Posted by SLINROB
The only Intel documented temperature info for a dual or quad core is provided by the on chip digital sensors which CoreTemp reports as Delta to Tjunction. Keep that above zero and as long as your computer is stable you'll never have a temperature related issue to worry about.
Quote:
Originally Posted by unclewebb
So, to put this to bed for us still a little confuse UncleWebb, Just use Coretemp 95 track your delta to Tjunction and check for stability. The larger the delta, the more headroom you may have with overclocking?
I just checked my Coretemp 95 and is showing 33C/34C and my delta is 52/51C to Tjuntion 85C for my E6600. Is this correct. Disregard Speedfan, Bios, and Ntune? :shrug:
Yes near enough, except its really delta to the TCC throttling point. They've just called it Tjunction. Don't forget though the throttling point is set for a chip at stock speeds. No guarantee an overclock is still stable even below the throttling point, just means you're below temperatures the Engineers believe won't harm the chip.Quote:
Originally Posted by Sniper.nkc
http://img124.imageshack.us/img124/1342/image1ca4.jpg
I'd like to show you my next screen shot then...;)Quote:
Originally Posted by Eightballrj
Quote:
Originally Posted by unclewebb
Thanks again for the explanation Unclewebb & Fornowagain. You guys really know your stuff :)Quote:
Originally Posted by fornowagain
Im confused, this is the first time I have ever heard of this temp issue.
That's because G0 is reporting temps that some of us find hard to believe, so this is a fairly reliable way (at the moment) to evaluate your temps.
A New High Record! ~ 3800MHz Small FFT Prime Stable Preliminary Results ~2hrs.
Notice it's FSB 475 Priming stable approx 2hrs and counting using x8 multiplier :cool:
Core temps are still awsome consider it's air cooled. Of course this is with a little help of lower ambient temp today at 21c.
Good news watercoolers: Consider what I've acheived on air, 3800mhz on water is def. feasible if you've a good batch ;)
http://img505.imageshack.us/img505/3...establesq6.jpg
Notice: I'm currently transfering all the previous screen shot database to the new server. Dead pics will be fixed soon. Thanks guys :)
3.8 GHz on air is ABSOLUTELY INCREDIBLE for Q6600 G0. Great Job!
All I want is 3.6 and I will be happy like a pig on mud. Are you planning to do the 1333 MHz mod on your Q6600 G0 to see if you can go a little farther?
How about sharing your settings for those of us using the same motherboard!
:D
:clap:
Of course orion23 :)Quote:
Originally Posted by orion23
There's no point to get shy and keep them to myself right ;)
I'm still busy updating the dead pix in the meatime, I'll use a quotation to show my detail BIOS configurations later in the day. You're welcome:)
I only ask because no matter how many times I have replaced my motherboard, CPU and RAM, I can never get my PC to pass Prime95 or Orthos.Quote:
Originally Posted by XtremeTiramisu
Never had a problem doing anything else from gaming to video encoding though.
Same thing with my new P5K Deluxe and this Q6600 G0. I've been playing all day long with it, @ 3.7ghz, using 1.5V (1.47 in Windows).
Prime95 fails on @ least 1 core. Temperatures go all the way to 60-65c.
I tried playing with some extra options in the bios such as the transaction booster and something else related to that, and my PC wouldn't even boot.
Upping the NB or SB voltage has very little effect.
I guess sometimes it could just be due to many USB or HD + other devices in use? :shrug:
Lower your OC.Quote:
Originally Posted by orion23
That's exactly what I found when playing around with my E6400. As the MHz went up I needed to leave more temperature head room to maintain long term, 100% Orthos stability.Quote:
Originally Posted by Sniper.nkc
Here are my approximate findings:
3200 MHz ~ DTS=5
3400 MHz ~ DTS=15
3600 MHz ~ DTS=25
Intel uses readings from the DTS to control the thermal throttling point as well as the THERMTRIP# processor shut down point. Stability when overclocking is also directly effected by how much head room the DTS is reporting so it's a good idea to forget about absolute temperatures and concentrate on how much head room you need to remain stable.
If you're not stable then you need to back off your overclock and that's about all there is to it. The engineers have built in a huge amount of temperature head room into the Core based processors and these processors are better equipped than any before to take care of themselves if they get too hot.
I have thoroughly tested CoreTemp 0.95 and it reads and reports the DTS 100% accurately in real time and is the only program I know of that reports the DTS value directly. SpeedFan 4.32 also reads the on chip DTS correctly but with the Q6600 it seems to be using a TjMax=85C instead of TjMax=100C so its absolute temperature readings need to be corrected by 15C as I previously posted in #269.
I wouldn't trust any other temperature information or software unless someone has taken the time and thoroughly tested it by reading process register 0x19C using CrystalCPUID or similar utility. CoreTemp passes this test. Many other programs, including TAT, do not.
This post explains how to test temperature monitoring software running on a Core based processor.
my mobo has serious vdroop. at 1.356 in bios it idles at 1.31 and under load goes down to 1.26.
time to start compensating a bit mroe for that.
seems my mobo droops more than most on here.
Good info, as usual (note the link in my sig is to your info thread over on {H} forums).Quote:
Originally Posted by unclewebb
I wonder how long it'll take for that paradigm shift to happen? :). People are obsessed with absolute temperatures (it's all we've had to go on until recently), but with all the various confusion and changes Intel makes (often on a per-chip or per-stepping basis), absolute temperatures just don't mean what they used to.
For example, I don't care if my cpu is reporting 120C (or 1000C, for that matter) - if it can run stable at that temperature (and the voltage is not dangerously, life-shorteningly high).
I think it's probably going to at least take some HSF review sites that list 'DTS' as thier primary measurement before enough people start thinking in terms of DTS overhead instead of absolute temps.
Gotta disagree with this if the tjunction is an unknown variable which is seems to be for these G0 chips... someone posted an intel document stating something about an 11 °C difference from the B3 chips... so that's either 111 °C or 89 °C -- either way it's potentially not what the software is reporting.Quote:
Originally Posted by Sniper.nkc
Gotta disagree! The only Intel documented temperature information for a Core based processor is all based on what the on chip digital thermal sensor (DTS) says. For all cpus in the Core series, thermal throttling doesn't begin until DTS=0.Quote:
Originally Posted by graysky
Assuming a maximum junction temperature of 100C may or may not be right but you can never argue with what the DTS is telling you.
The 11C change from B3 to G0 is a change in the maximum recommended Tcase temperature at TDP and is not documented by Intel as a change in the maximum junction temperature between the two revisions. TjMax may be different between B3 and G0 but Intel doesn't document this.
hey!...heat is just one issue of many...other variables involved are: cpu stepping, memory, board, cpu heatsink, quality PSU and how good of a case you have to dissipate or get the heat outside of it!
@unclewebb: I dunno man... point I'm making is that the temps on these g0's seem really low compared to the b3's. Yeah, they can changed the tdw by 10 W but something is up with these temps.
Quote:
Originally Posted by graysky
why be concerned over the temps being lower if you can get a higher overclock out of it???