Man of Honour
- Joined
- 12 Jul 2005
- Posts
- 21,460
- Location
- Aberlour, NE Scotland
What pump is that then? It must be quiet. My Aquaextreme 50z buzzes away so i would know if that stopped. Looking for a quieter pump actually.
Overclocking - what`is maximum temperature i should really stop overclocking at???
- Thread starter Sylver123
- Start date
What pump is that then? It must be quiet. My Aquaextreme 50z buzzes away so i would know if that stopped. Looking for a quieter pump actually.
D-Tek DB-1. Only bought it 'cos it was cheap, but seems perfectly suitable for a simple CPU loop like mine.
http://www.overclockers.co.uk/showproduct.php?prodid=WC-002-DT&tool=3
Audible but only if I stick my head up against it.
Associate
- Joined
- 18 Mar 2009
- Posts
- 876
- Location
- Indiana, United States.
Mine max out at 73c. I don't think 67c is really saving you that much. I would say you are okay so long as you don't hit 80c. That's when some cpus will start to throttle.
and 1.4 vcore is the max I would ever be interested in.
Also I too have throttled my pc, but it bsod, literally seconds after it started to throttle. That was at 4.2ghz on air, and temps could just not cope.
and 1.4 vcore is the max I would ever be interested in.
Also I too have throttled my pc, but it bsod, literally seconds after it started to throttle. That was at 4.2ghz on air, and temps could just not cope.
Soldato
- Joined
- 5 Feb 2009
- Posts
- 3,962
I don't like to go above mid-60s Prime and 70C IBT.
You can get away with a few degrees give and take. There's no exact 1-degree cut-off line that delimits what CPUs can safely handle - but anything past 70C makes me too uncomfortable.
In the end once you get past 70 or so it comes down to the Dirty Harry test... are you feeling lucky?
Well, are ya, punk?
You can get away with a few degrees give and take. There's no exact 1-degree cut-off line that delimits what CPUs can safely handle - but anything past 70C makes me too uncomfortable.
In the end once you get past 70 or so it comes down to the Dirty Harry test... are you feeling lucky?
Well, are ya, punk?
When checking cpu temps should the overall temp be the governing guide or the hottest core value (ie I7 temps on the individual cores are always higher on my setup than the overall cpu temp reading - checked with everest) in fact further to that.....why do the individual core temps vary?? i sometimes see 5 -8c difference....just interested really 
Man of Honour
- Joined
- 12 Jul 2005
- Posts
- 21,460
- Location
- Aberlour, NE Scotland
I try to keep IBT testing under 70C and havent pushed Vcore past 1.4V yet
If it get's too hot it will throttle before eventually shutting down. If you give it too many volts you will kill it. Voltage causes the most damage. You have a 45nm core do not go past 1.45v. You are already at the ideal maximum. Just because you go with watercooling does'nt mean you can increase the voltas.
You're running that big an overclock on the p5q deluxe auto settings?
I'm not sure if the board or the chip will burn out first, but it's not going to take long.
The p5q is known for putting mental voltages through components on auto. I recommend putting everything to auto, then setting northbridge, fsb term, cpu pll to one above zero and vcore to your vid. Loadline calibration is one to research and possibly argue over. Then slowly increase in the tried and tested style, keep fsb term and northbridge below 1.4V, preferably far below.Cpu pll should be fine at 1.52.
Higher voltage will kill things just as easily as temperatures
High fsb term will (probably) kill your chip
High northbridge will definitely kill your motherboard
High vcore will both damage and overheat your cpu
Auto is a bad choice
I'm not sure if the board or the chip will burn out first, but it's not going to take long.
The p5q is known for putting mental voltages through components on auto. I recommend putting everything to auto, then setting northbridge, fsb term, cpu pll to one above zero and vcore to your vid. Loadline calibration is one to research and possibly argue over. Then slowly increase in the tried and tested style, keep fsb term and northbridge below 1.4V, preferably far below.Cpu pll should be fine at 1.52.
Higher voltage will kill things just as easily as temperatures
High fsb term will (probably) kill your chip
High northbridge will definitely kill your motherboard
High vcore will both damage and overheat your cpu
Auto is a bad choice
There are three sets of temps you need to consider
- Real World Max Temps (from gaming, encoding, stuff you do etc)
- Synthetic Real World Max Temps (Prime 95 Smalls, OCCT etc)
- Synthetic Worst Case Scenario Max Temps (IBT)
but in the Real World I cannot find anything to put this much pressure on a chip . . . but regardless its possible and my systems can take it should it happen,Lol ain't that the truth!
These dudes must have been reading some ancient Conroe/Kentsfield overclocking guides, 45nm is a different beast from 65nm, I'm getting the bulk of extra MHz from these Wolfdales at around 1.35vCore with severely diminishing returns much over that.
I guess it depends what kinda overclock you are talking about but I've personally found all three of those voltages needs to be raised to some extent to get a good stable result, vNB and vTT need to be adjusted for sure somewhere between 400MHZ-FSB-500MHz-FSB and the CPU PLL helps a lot once you start pushing big MHz from your chip. a 1.52 PLL will result in my machine rebooting during a Prime Small FFTs or 10-100 IBT loops but 1.56-1.58 PLL sorts out the problem. No amount of vCore can replace the effects of increased PLL. If in doubt as what PLL to set leave it on [Auto] until you have reached a stable clock then turn PLL back to 1.52v and see if the system is still stable under heavy load! . . . if it is then great but if not increase it step by step until the system is stable.
As above, I've had two chips that need a PLL boost once the overclock gets big!
The above ^ shows that I wasn't terribly clear, I apologise
I meant put fsb term and nb at one notch above the minimum while at stock, and then increase these as required while overclocking up to about 1.4. It probably isn't helpful to refer to minimum as zero either.
The reason for one notch up from minimum is that the p5q is often reported to treat minimum the same as auto, and hence overvolt while set to minimum.
Did you need higher pll with dual core chips? My e8400 went up to 3.8 or so with minimum pll so I assumed it was only quads that needed more, one chip is hardly a large sample though.
I meant put fsb term and nb at one notch above the minimum while at stock, and then increase these as required while overclocking up to about 1.4. It probably isn't helpful to refer to minimum as zero either.
The reason for one notch up from minimum is that the p5q is often reported to treat minimum the same as auto, and hence overvolt while set to minimum.
Did you need higher pll with dual core chips? My e8400 went up to 3.8 or so with minimum pll so I assumed it was only quads that needed more, one chip is hardly a large sample though.
What CPU PLL Voltage and NB Voltage would one recommend me trying on my dual core E7500 overclocked to 3.8GHz to see if i can decrease my VCORE Voltage and making my processor a little cooler???
I`ve came across some sites with guides on what`s recomended setting try these at but unforchenitly only for Quad Core chips, and I haven`t came across one for dual core chips.
I`ve came across some sites with guides on what`s recomended setting try these at but unforchenitly only for Quad Core chips, and I haven`t came across one for dual core chips.
Yeah, I did the same with my Q6600. Was running a suicide run at 3.8ghz, when it just blacked out. I don't think it slowed down though, just everything went off. I thought I'd killed it.
A got reason not to disable CPU TM function in the bios, as a lot of people still say to turn it of when over clocking
Intel have programmed the CPU to shut down to protect itself at a pre-determined core temperature. That temperature varies with the CPU, but it usually lies somewhere between 95C and 115C. It does this whether the BIOS Thermal Management functions are enabled or not. That's all proveable fact.
Statements like "I wouldn't like see anything over 70C for extended use" are subjective and if you did this in your day-to-day life you would stop your car to let it cool down on a hot day, or turn the fridge off because the compressor was running flat out. A non-overclocked Q6600 with the standard Intel cooler will idle in the high '50's and load in the mid-'80s in a Dell Optiplex desktop in a warm office and no-one gives a hoot because they are blissfully unaware that CoreTemp even exists.
The ultimate bottom line is that if you can't afford to replace the components in your PC if you blow them up, you shouldn't be overclocking.
You always overclock at your own risk.
And I would echo what has been said above about voltages rather than temperatures being the problem with reducing the life of CPUs. JonJ678's original post about the dangers of the [AUTO] setting on the P5Q being especially valid in my opinion.
Statements like "I wouldn't like see anything over 70C for extended use" are subjective and if you did this in your day-to-day life you would stop your car to let it cool down on a hot day, or turn the fridge off because the compressor was running flat out. A non-overclocked Q6600 with the standard Intel cooler will idle in the high '50's and load in the mid-'80s in a Dell Optiplex desktop in a warm office and no-one gives a hoot because they are blissfully unaware that CoreTemp even exists.
The ultimate bottom line is that if you can't afford to replace the components in your PC if you blow them up, you shouldn't be overclocking.
You always overclock at your own risk.
And I would echo what has been said above about voltages rather than temperatures being the problem with reducing the life of CPUs. JonJ678's original post about the dangers of the [AUTO] setting on the P5Q being especially valid in my opinion.
As above. I like to run mine nice and cool @ no more than 50-60 while priming.
Intel chips will throttle @ their given TJ MAx as stated above. High PLL and VTT are easy chip killers and TBH if your having issues while stability i would look at your NB first before using insane vcore @ PLL.
Last edited: