Lapping your CPU

[Deleted]

Guy here on the bit tech forums lapped his i7 2600k and managed to knock 5-9degrees off each core.

http://forums.bit-tech.net/showthrea...ghlight=lapped

[Deleted]

my point was merely that if you have a system where your temperatures are so high that they're at an unacceptable level, you failed somewhere in your cooling solution.
while lapping might (really, might, thermal paste is designed to do what you're trying to do, and unless you have clocked a lot of hours with grinding equipment you're not going to get a perfectly flat surface anyway, might seem like it to the naked eye tho) reduce your temperatures, it's not going to be a magical five degrees celsius, and as such yes, if you want a colder CPU by a couple of degrees, go for it, if your CPU is too warm for you right now, then you need to rethink your cooling solution overall or alternatively downclock it a bit.

[Adappy]

my point was merely that if you have a system where your temperatures are so high that they're at an unacceptable level, you failed somewhere in your cooling solution.
while lapping might (really, might, thermal paste is designed to do what you're trying to do, and unless you have clocked a lot of hours with grinding equipment you're not going to get a perfectly flat surface anyway, might seem like it to the naked eye tho) reduce your temperatures, it's not going to be a magical five degrees celsius, and as such yes, if you want a colder CPU by a couple of degrees, go for it, if your CPU is too warm for you right now, then you need to rethink your cooling solution overall or alternatively downclock it a bit.

It's not really about getting temperatures to more acceptable levels though, it's about getting every last ounce of cooling out of your system that you can so you can potentially overclock higher. Also, metal to metal contact conducts heat better than thermal paste (and paste is a much better conductor than air). If both the heatspreader of the CPU and the base of the heatsink were 100% smooth and flat, you wouldn't need any paste at all. (that's why they always say not to use a ton of thermal paste)

I've considered lapping my 2500k because its heatspreader is a somewhat concave in one axis... but I decided against since I don't want to shell out another $220 if I screw it up (and I'm fine with my clock speed and temperatures)

[Chronius]

It's not really about getting temperatures to more acceptable levels though, it's about getting every last ounce of cooling out of your system that you can so you can potentially overclock higher. Also, metal to metal contact conducts heat better than thermal paste (and paste is a much better conductor than air). If both the heatspreader of the CPU and the base of the heatsink were 100% smooth and flat, you wouldn't need any paste at all. (that's why they always say not to use a ton of thermal paste)

I've considered lapping my 2500k because its heatspreader is a somewhat concave in one axis... but I decided against since I don't want to shell out another $220 if I screw it up (and I'm fine with my clock speed and temperatures)

I'm curious are the newer batches of the 2500k and 2600k more willing to oc due to better cooling, than the first batch?
First batch all seemed to have a point at which no matter how much cooling you threw it, it wouldn't go higher. I believe the average was around 4.5-4.8Ghz.
I know my own 2600k hits its wall at 4.8Ghz, I had it at 5.3Ghz but that's only just "bootable" running any bench for more than 1min will result in a fail. Despite my WC keeping it around 60c

[dicertification]

I have a 920 C0 I lapped. The temps dropped a solid 5c and stabilized my overclocks allowing me to do 24 hour Prime runs @ 4.0Ghz which I couldn't do before. Fortunately I picked up a D0 when they finally came out and gave me the same results as the C0 without having to lap it (probably just luck of the draw), haven't lapped it yet as it's a time consuming process I'm not quite prepared to do again. I will also warn you that if it's your first time lapping. Be patient with it as it could take a lot longer than you planned. My first one took almost 6 hours of sanding. Ended up sanding my thumb down to the point where it was bleeding without even noticing it was happening lol. Gotta love that fine grit paper.

[Saithes]

I'm curious are the newer batches of the 2500k and 2600k more willing to oc due to better cooling, than the first batch?
First batch all seemed to have a point at which no matter how much cooling you threw it, it wouldn't go higher. I believe the average was around 4.5-4.8Ghz.
I know my own 2600k hits its wall at 4.8Ghz, I had it at 5.3Ghz but that's only just "bootable" running any bench for more than 1min will result in a fail. Despite my WC keeping it around 60c

The Limitation of Sandybridge's clock speeds is because of the poorly implemented PLL on the CPU.

[Auralian]

How to destroy your CPU by lapping it

http://hallicino.hubpages.com/hub/Ho...-by-lapping-it

I agree with this guy, pretty stupid and risky thing to do for something you pay a nice chunk of money for. I guess if you have money to burn go right ahead.

[shoebox]

I lol'ed, just read through the comments of that article and see how the author responds to people, just reading the article makes me doubt its credibility, and if that wasn't enough,i just needed to read the comments.

[Asmekiel]

I lol'ed, just read through the comments of that article and see how the author responds to people, just reading the article makes me doubt its credibility, and if that wasn't enough,i just needed to read the comments.

I rather believe the 4 guides I've read about lapping that are made my people who I know are credible, than 1 person I don't know who says otherwise.
Soapy water? 300 grit? Barrack Obama? 20C difference?

Where did he get that info... Yes, it's potentially dangerous and does void warranty, but so does opening an OEM Dell PC. Does that mean it'll instantly destroy it and drop my IQ to arctic regions -_-'

Having said that, I don't lap my CPU. I've never done it, don't have an old CPU to try it out, am not willing to sit down ~1.5-2 hours to do it and I also don't have €250 to spare in case I fuck it up. I also don't recommend it if you don't really want to get that last 100 extra MHz out of your CPU.
But in the end it's really up to yourself.

[Cilraaz]

I lol'ed, just read through the comments of that article and see how the author responds to people, just reading the article makes me doubt its credibility, and if that wasn't enough,i just needed to read the comments.

I also agree that the poster of that article was insanely biased without even attempting to hide it. However, note that most gains from lapping are 2-5°C. You can lower the temp of your CPU by 2-5°C with a good re-application of TIM. I can't say how much of the cooling reduction from lapping is due to the actual lapping and how much is just from a fresh TIM application.

[Icpie]

How to destroy your CPU by lapping it

http://hallicino.hubpages.com/hub/Ho...-by-lapping-it

I agree with this guy, pretty stupid and risky thing to do for something you pay a nice chunk of money for. I guess if you have money to burn go right ahead.

That guy sounds like what he's exactly what he described all lappers to be.

Anywho, I've got some 600, 1000 and 1500 grit sandpaper, a box of cookies, 3 hours to burn and 3 CPUs that need some grinding. (Found another CPU collecting dust! Some old single core athlon) . Good luck to anybody planning on doing this.

[Deleted]

I did it to my Q6600 and watercooling block, evened the difference between the cores.

*Watercooled 4.0GHz*
Before lapping
Idle - 32~36c
Load - 62~65c
Difference between cores 2~6c

After
Idle 28~33c
Load 58~62c
Difference between cores 1~3c

[FlawlessSoul]

I've lapped an old Core2, and while it only dropped temperatures under an H70 by about 4 degress, the most important thing the lapping did was flattened the IHS and gave me similar temperatures across all cores. Pre-lap it had its four cores up to 8 degrees apart under load at some points. Afterwards, it was at most 3 or 4 degrees apart.

I'm tempted to do the same to my i7 920 now that I've had it a while, and next time I pull my machine apart (in a couple of weeks) I may very well do it.

[shoebox]

Dont forget to dunk it in soapy water, then shove it in the microwave to dry it off.

[FlawlessSoul]

Dont forget to dunk it in soapy water, then shove it in the microwave to dry it off.

I just took an inch-wide grinder to the top, one of the ones used to sharpen axes. Once all that silver stuff on top was gone, and I had a direct contact with the circuits and junk inside, I just started up the auto-clave and tossed it in to clean it off.

After the steam clean, I hammered it back into the socket and bolted the heatsink back on. Job done.

[Afrospinach]

How to destroy your CPU by lapping it

http://hallicino.hubpages.com/hub/Ho...-by-lapping-it

I agree with this guy, pretty stupid and risky thing to do for something you pay a nice chunk of money for. I guess if you have money to burn go right ahead.

I read this article a couple of years ago, and was the first thing that popped into my head when i read the thread title. I still don't know how anyone can read it and not see that this guy is a complete tool. You can PROVE lapping works for a start, and saying that intels engineers spent hours designing it ignores the fact that they have more on their minds than perfect heat transfer. By that logic, why are stock coolers not superior to everything else. Hell, I have an intel performance cooler that was concave to the naked eye. It is a bit like aftermarket exhausts for cars, the engineers are looking at aspects other than performance when they design them, leaving a lot of improvement to be had for someone who does not mind noise or expense. One comment brings up a valid point, copper corrodes, that is probably the reason they have excess material on the cpu to begin with.

People are taking a risk as soon as they touch the clocks on their cpu and that is their business. I would say this guy cared about they greater good but clearly he is just a condescending douche who can't actually deal with anyone that does not think exactly as he does.

[chaud]

I lapped a heatsink more than a few years ago and didn't see a significant gain. I don't think it is worth the effort, but if you are bored and need something to do it isn't a big risk, provided you have the right heatsink for it. I wouldn't ever lap a CPU though.

[Cleetuss]

isnt that was the thermal compound is for? to make contact between the 2 surfaces?

[chaud]

Yes, but it isn't as efficient as two flat surfaces.

[Azoth]

Check this out also
http://www.l3p.nl/other-builds-mods/cpu-lapping/
Something slightly insane about lapping a i7 980x

Avast! AV threw up numerous warnings when I visited this site. I would recommend EVERYONE stay as far away as possible.

As far as lapping unless you are really trying to max out every single Mhz out of an overclocked CPU then maybe. For the average cpu tuner I wouldn't recommend.