AMD and Intel

U

UK_Widowmaker

Guest
Love the new intel chip i just bought...but I do wish that intel heatsinks were made like the AMD ones, when it comes to installing!

I always try to use AMD for workstations when I'm building them, as the heatsink has the side clip...So easy!

Intel on the other hand, are a bit of a pain...and upgrading the CPU involves removing the entire motherboard to re-seat the new heatsink.

They sure missed a trick there!
 
Wotcha got, wotcha got?! Huh?! Huh?! :icon_lol:

I take it you bought a new heatsink which came with a different mobo-mounting to the old one?
 
Uhh, negative, mate :) Removing the heatsink can be done without removing the motherboard. (And I should know, I've just done two in the past week). I will grant that the Intel stock heatsink/fan is a major pain in the butt - seems everyone knows that. But, what I've found is that - if it's done properly - it does work.

To be specific, the compression pins used on the Intel stock heat sinks are specifically designed so that you *don't* need access to the back side of the board to install or remove it.

If you've purchased an after-market heatsink fan that uses 'through-the-board' fasteners, then yes, you have to remove the mainboard to get to the fasteners. But that's the reason Intel doesn't use this type of fasteners.

Not to be blunt, but what I find is most people just aren't doing it properly. (And even when you do, it is admittely a pain - but it works).

HTH
 
Hi Siggi

I went with the E8500 in the end, with an Asus P5Q with P45 chipset (with a view to a bit of OC(ing)
But...I forgot to get a better Heatsink (excitement got to me!...lol) which means I will have to run it as per normal for now...and that's why I think it will be a pain...cos when I get the new heatsink, I will have to take the mobo out to fix the lugs of the new one into the mobo.

Silly me!
 
cliff527 said:
Uhh, negative, mate :) Removing the heatsink can be done without removing the motherboard. (And I should know, I've just done two in the past week). I will grant that the Intel stock heatsink/fan is a major pain in the butt - seems everyone knows that. But, what I've found is that - if it's done properly - it does work.

To be specific, the compression pins used on the Intel stock heat sinks are specifically designed so that you *don't* need access to the back side of the board to install or remove it.

If you've purchased an after-market heatsink fan that uses 'through-the-board' fasteners, then yes, you have to remove the mainboard to get to the fasteners. But that's the reason Intel doesn't use this type of fasteners.

Not to be blunt, but what I find is most people just aren't doing it properly. (And even when you do, it is admittely a pain - but it works).

HTH
Click to expand...

Oh right...I have'nt looked at it yet, so perhaps I shall get away with that...thanks for the headsup
 
UK_Widowmaker said:
Hi Siggi

I went with the E8500 in the end, with an Asus P5Q with P45 chipset (with a view to a bit of OC(ing)
But...I forgot to get a better Heatsink (excitement got to me!...lol) which means I will have to run it as per normal for now...and that's why I think it will be a pain...cos when I get the new heatsink, I will have to take the mobo out to fix the lugs of the new one into the mobo.

Silly me!
Click to expand...

Aaah...4ghz shall be yours! :jump:
 
Siggi said:
Aaah...4ghz shall be yours! :jump:
Click to expand...

I may be asking for a bit of help soon Siggi...regarding Overclocking (as I am a frightened OC virgin!..who doesn't want to melt his rig!)

Hope you wont mind?:kiss:
 
UK_Widowmaker said:
I may be asking for a bit of help soon Siggi...regarding Overclocking (as I am a frightened OC virgin!..who doesn't want to melt his rig!)

Hope you wont mind?:kiss:
Click to expand...

Not at all. :)

In preparation it would be a good idea to download Prime95 and the free trial version of Everest.

http://files.extremeoverclocking.com/file.php?f=103

http://www.lavalys.com/products.php?lang=en&page=11

And grab cpuid while you're at it:

http://www.cpuid.com/cpuz.php

For future reference:

22135613223l.jpg
 
If I can toss in a crumb here: The BIOS screen shows memory and CPU voltages being increased, which is typical in OCing. However, the "NB" (North Bridge) voltage is left at "auto". This also commonly requires bumping up a bit as well.

I say this because my first few attempts to OC failed...until I finally learned about the NB voltage :) I'm almost certain you'll see this in the overclocking documentation you'll find.

Good luck.
 
cliff527 said:
If I can toss in a crumb here: The BIOS screen shows memory and CPU voltages being increased, which is typical in OCing. However, the "NB" (North Bridge) voltage is left at "auto". This also commonly requires bumping up a bit as well.

I say this because my first few attempts to OC failed...until I finally learned about the NB voltage :) I'm almost certain you'll see this in the overclocking documentation you'll find.

Good luck.
Click to expand...

I've posted the BIOS pic only for future reference, so I can point out what WM will possibly need to adjust (it's his board's BIOS). Every OC is different. You start off at stock everything, bump the CPU and run Prime95. If it fails you bump the CPU voltage and run Prime95 again. Once you've reached the highest CPU voltage you're comfortable with, but Prime95 still fails, you bump all the other volts one notch each. Eventually you get Prime95 stable. Then you can start bringing the CPU voltage back until Prime is unstable again. Put the CPU back up one notch, then start bringing the other volts back one notch.

Which is a laborious load of bollocks and I personally can't be arsed with it. I creep the CPU up to reasonably short of it's safe max, RAM at full spec, and all the other volts up to max blue (but only if Prime failed with CPU and RAM adjusted).

DON'T use Asus auto-tweak or auto anything, it sets the CPU volts WAY too high. It actually set my CPU too high at STOCK! 1.45v at 3ghz, when I have it now at 1.28v for 4ghz. Way to go Asus, maybe they're in cahoots with Intel. :icon_lol:
 
Siggi said:
DON'T use Asus auto-tweak or auto anything, it sets the CPU volts WAY too high. It actually set my CPU too high at STOCK! 1.45v at 3ghz, when I have it now at 1.28v for 4ghz. Way to go Asus, maybe they're in cahoots with Intel. :icon_lol:
Click to expand...

Hi Siggi

Ah...glad you told me that!..cos i installed it!..hahaha (havent used it yet)
(downloaded the stuff you've advised above)

Ok, the new board and case are sorted (didnt even need to repair/format WindowsXP as it was Asus 2 Asus which was a nice surprise!

I have a 750W Gigabyte PSU....but as I only have the generic heatsink/fan..can I assume that it's too risky to continue at this stage?
 
UK_Widowmaker said:
Hi Siggi

Ah...glad you told me that!..cos i installed it!..hahaha (havent used it yet)
(downloaded the stuff you've advised above)

Ok, the new board and case are sorted (didnt even need to repair/format WindowsXP as it was Asus 2 Asus which was a nice surprise!

I have a 750W Gigabyte PSU....but as I only have the generic heatsink/fan..can I assume that it's too risky to continue at this stage?
Click to expand...

It's best not to chance it mate. People do OC that chip with the stock cooler, but only through ignorance or lack of funds to buy a decent one.

Make sure you get some quality thermal-paste with the new cooler, Arctic Silver is good stuff.

Do you have the whole lot up and running? Check in cpuid and see what the CPU's v-core (voltage) is at.
 
Siggi,

No offense intended, man. I was simply pointing out that (in *my* experience, as I said) the NB voltage has always required some adjustment. And (as I said) I'm sure the references one can find online will corroborate that.

As for overclocking on a stock fan (you said "only through ignorance or lack of funds..."): Well, to a point, I agree. But then, I've successfully OC'd 8400s on a stock fan and kept them within the Intel thermal spec (72-74C, depending on the revision).

If it'll run full bore (100% load, using Prime95) and not exceed termal spec, then OK - but this *clearly* would depend on a number of factors not even mentioned here. How much CPU overvoltage, case cooling, the actual FSB used, will all affect the temps - and therefore what cooling you need. I agree that the stock Intel HSF leaves much to be desired - but it's not true to say it won't work. It simply depends.

Obviously, cooler is better, but maybe someone doesn't intend to jack the OCing to the roof, and therefore doesn't require very aggressive (and perhaps very costly) after-market cooling solutions. Yes, most anyone will tell you that if you want to OC, the stock fan is out. But I say use the tools available (Prime95, Coretemp, etc.) to make your own determination.

I definitely did *not* say to rely on Asus (or any other) mfr "automatic" OCing. You should always learn how things work, and use measurement tools. "Automatic" seldom applies; and *never* in the case of OCing a PC.

Finally, the CPU voltage: I've run them, quite successfully, at or near 1.5v, without problems. Depends on who you listen to, what risks you want to take v. how fast you want to go. Again you have to take into account that higher voltages mean higher temps - and it goes without saying where voltages are concerned - the less you can use and still be stable, the better.

You said "Once you've reached the highest CPU voltage you're comfortable with..." well, let me ask you: You're already going outside spec...so, especially assuming someone who's never done this, exactly how would they know what to be 'comfortable' with? My point is that you're already inferring some level of experience/level of comfort - which no one really has if they haven't done this before.

Again, what matters is measurement using tools available, and learning about what you're doing. And, in every case, it's a question of how much you want to risk, to get that next percentage of improvement. I DO NOT advocate first timers try it - but people have gotten 50% overclocks with 8400 CPUs - they simply learn what to watch for, use the tools at their disposal, and decide how to manage or accept the inherent risks.

As always, everything I post is strictly my opinion and/or experience, and I usually say so. I also usually say "Do your own math", "Your experience may vary" or some such.
 
UK_Widowmaker said:
1.128 V at idle
Click to expand...

Sweet. :)

Cliff, no offence taken mate. Did you think there was? The spec for the E8 series is reportedly as low as 1.25v to 1.36v. "Comfortably" by my definition is not to go above 1.36v then, because both those specs are from Intel and clearly the lower is intended to give them a huge margin for error. As you say, people go up to 1.50v and the chip still works. But for how long. There are two types of degradation, a simple fry-up due to heat and some kind of electron-decay shennanigans going on inside the tiny little pathways (millions of them) that route the electricity through the chip.

As I said in the earlier post, I'm aware some people OC on stock coolers. And I'm sure it can be done safely by going slowly and monitoring all the parameters (voltage vs heat etc). But bollocks to that kind of on-the-edge ducking and diving. If you're going to race your car beef up ALL the gubbins, not just the pistons. Or be prepared to see the lot come through the bonnet when you accidentally miss a gear in 2nd while the pedal's to the metal. :icon_lol:

but people have gotten 50% overclocks with 8400 CPUs - they simply learn what to watch for, use the tools at their disposal, and decide how to manage or accept the inherent risks.
Click to expand...

Usually people who are OC'ing for the sake of getting the highest possible numbers, even if only for long enough to take a screenie before the chip blows or the OS shuts down. What we're looking for here is 24/7 useable game-playing stability. :)
 
You must log in or register to reply here.
Back
Top