Is it worth delidding the 4790k?

Is the TIM good enough on the 4790k that it doesn't need to be delidded? I built a new system and plan on overclocking it. But wanted to know if you guys think it's worth delidding?

depends on expectations

for the best ocing results on ambient I think the EK naked kit is top shelf

http://www.ekwb.com/shop/ek-supremacy-precisemount-add-on-naked-ivy.html

Unequivocally, YES.

The TIM used is horrible. Idling during even the cooler months was horrible temps. After delid and Liquid Pro + PK-3, and I've yet to see more then 62C peak on any core so far. I'm sure come summer I will see over 70C, but nothing like before where idling was close to 50C, and 90C when stressed was as easy as converting audio files for my phone.

I don't think that delidding is worth for most users. Even if temps are higher then what does it change ? Overclocking ? Like it makes any difference in daily work if you make 100MHz more or less.
CPU won't die because of 5-10*C more. In most cases you barely reach that 10*C difference.
These CPUs have throttling point above 90*C. Shutdown temp at about 100*C+.
If you are using air cooling then delidding can help to make couple of MHz more. If you are using good water cooling then you barely see difference other than couple of degrees less in monitoring software. If you are using sub 0 cooling then you won't see any difference.

Simply all depends what you do and what you need. However if you are playing games and use other more standard programs then I see no point.

it is only needed for those that are badly temp limited, that is if you are hitting 90-100c at 4.6 or so, still it's pretty much only for benchers.

Woomack said:

I don't think that delidding is worth for most users. Even if temps are higher then what does it change ? Overclocking ? Like it makes any difference in daily work if you make 100MHz more or less.
CPU won't die because of 5-10*C more. In most cases you barely reach that 10*C difference.
These CPUs have throttling point above 90*C. Shutdown temp at about 100*C+.
If you are using air cooling then delidding can help to make couple of MHz more. If you are using good water cooling then you barely see difference other than couple of degrees less in monitoring software. If you are using sub 0 cooling then you won't see any difference.

Simply all depends what you do and what you need. However if you are playing games and use other more standard programs then I see no point.

Click to expand...

Exactly...

If you are temperature and NOT voltage limited, go for it. If your voltage is high too, there isn't a point.

I don't understand how you can get such a temp difference Mpegger.

My (NOT delidded) 4790k @ 1.150v @ 4400ghz (c-state enabled so it throttle down the mhz) "idle" at about 2-5*c higher than the water temps of mye H100 depending on the cores. So currently, with room temp at ~21*c, H100 temp @ 26*c and cpu temp (while posting this) 31-30-29-27 (HWiNFO64)

Because the thermal paste is inferior and there is a larger than needs to be gap between the IHS and core. That is why temperatures can drop so dramatically.

As far as your temps... idle temps are not really relevant. Its the load test and temps and each CPU will vary (for many many reasons).

EarthDog, my question was not based on that I don't know how the CPU i built, and the reason why temps are different from each CPU.

I've been trough 4 different LGA1150 cpu's now, and none of then have been "horrible" regarding temps either at idle or load. Infact the G3258 is the worst one, of two 4690k and a 4790k. But Idle @ close to 50*c sound like either insane high ambient temps, the cooler is not mounted right, or the airflow is "non existing".

Even the original Intel cooler is not even close to that bad with any of the mentioned CPU's I've had.

I would say that while pushing the limits both voltage and mhz vice, it is worth delidding the cpu to gain the extra temp difference.

I would say that while pushing the limits both voltage and mhz vice, it is worth delidding the cpu to gain the extra temp difference.

Click to expand...

That depends though. To me, so long as the CPU is running under its temperature limit, there isn't a point to delid. Temps dropping 10C or so isn't worth it (again, to me), to risk borking your CPU or at best voiding the warranty.

Only in a temperature limited but there still exists voltage headroom should this be done (to me). I suppose I am risk averse when there really isn't anything to gain.

EarthDog said:

That depends though. To me, so long as the CPU is running under its temperature limit, there isn't a point to delid. Temps dropping 10C or so isn't worth it (again, to me), to risk borking your CPU or at best voiding the warranty.

Only in a temperature limited but there still exists voltage headroom should this be done (to me). I suppose I am risk averse when there really isn't anything to gain.

Click to expand...

I agree with you. What I tried to say is that, if you are pushing everything to its absolute limit! Then yes! Otherwise, don't bother.
For daily use, get a better cooling system, or just stop with the "OHMYGASHLOOKATMYINSANEEXTRA100MHZOVERCLOCK@ROFLMAOPIMPVOLTAGE" thing going around the world.

If you run 4.4ghz, or 4.8ghz, you won't notice in daily use anyways.

I delidded and saw a few c's drop, that's it. Nothing huge, nothing spectacular, nothing OMFG. It was more of.. Oh, that's meh. And I did it because I wanted to, not because I needed to.

I still have delidded 3770K , temps dropped by about 10*C but OC is without changes regardless of used cooling.

Hard to say, I have heard everything form 3C to 20C changes in load temps

If you are one of the guys on the high end of this obviously it is worth it, on the low end not so much

Ultimately this is simply comes down to if you want to do it because you can ?

The drop in temps you see will depend on power tested, since the temp gradient through the chip obviously depends on power tested. Intel and other chip manufactures have stated in past in papers that 1/3 gradient is across tim junction (referring to paste tim). Also there are some that removed intels polymer tim and replaced with paste tim, and should see much less improvement. Those that use liquid metal for tim1 would see much lower temps.

AT idle 2-3W, you should see less than 1-2C temp improvement with delidding, because gradient through chip is only few C. (gradient through chip is 5C at idle, then 5/3 = 1.7C max temp drop.
Before and after at 130W with aida64 at 4.7ghz, 1.29v, mine was 12C lower temps, from 72C to 60C. (gradient through chip is 72c-25C or 47c, ~1/3 is tim or 47C/3 = 15.7C max temp drop).
Before and after at 185W with prime v 28.5 small ffts at 4.7ghz, 1.29v, mine was 22C lower, temps from 98C to 76C max, with clp. (gradient through chip is 98C - 25C (ambient) and across tim ~73C/3 = 24C).
Before and after at 185W with prime v28.5 small ffts at 4.7ghz, 1.29v, mine was 10C lower, temps went from 98C to 88C, when I tried pk1 for tim1.

Those that saw 18-20+C improvement on 4790k were using prime 95 v28.5 or linx at higher vcore, and many posted that typical temp drop with prime 28.5 small/linx in OCN delidded thread.
Those using aida64, x264 encoding, etc, with much lower wattage, most of those posted 8-12C temp improvements.

But all depends on wattage tested. I would bet if all tested at similar watts, the range would be much tighter among various reported temp drops.

Rge, good post.

To keep this basic and to the point, of all The chips Ive de_lidded, Amd or Intel air or liquid cooling... AN average of 10c drop in temps upon load.

Amd seems to really do well without an IHS plate all together, but have done a couple where the IHS plate is reused and still had better temps strictly speaking because solder is horrible at transferring heat.

With Intel, the gains could be minimal, and worthless as ambient temps play a big, no huge role in cooling as its the most important part of the mathematics. Today if your room temp is 3c cooler than yesterday, you'll notice that it effects temps on the processor.

Thickness of the IHS plate also plays a part in cooling. With many de_lids, i've replaced the copper with a lapped flat morgan silver dollar as silver conducts heat a little better, we are talking a gain of maybe 1-2c drop in temps if the room temp happens to be the same before and after a delid.

As far as clock speed goes.... You'll be lucky to see 100mhz increase, and lets face it.... Thats not enough to bring better daily performance to the table.

With Intel, the gains could be minimal, and worthless as ambient temps play a big, no huge role in cooling as its the most important part of the mathematics. Today if your room temp is 3c cooler than yesterday, you'll notice that it effects temps on the processor.

Click to expand...

Intel or AMD this is true. It effects them both the same really. If not, why would that be different? I am not sure I understand that passage as written?

I dont understand it either.... Lol.

Want lower cpu temps, lower ambient temps?

In retrospect, the benefit of delidding mine was 1) 3D benching/ie firestrike etc at 5ghz (vs 4.9ghz without delid) and 2) educational purposes.

I delidded so I could run prime 28.5, but now realize prime 28.5 (omitting small ffts in blend) or using prime 27.9/Occt works just as well and avoids delidding. My 24/7 settings 4.7 or 4.8, requires same vcore for stability before and after delid. And 24/7 use, cpu max wattage never above 100W even gaming, max in 80's, hence temp decrease much smaller for 24/7 use, ie gaming avg before delid 50's C, after delid high 40's C, neither is an issue. So delidding for 24/7 use was irrelevant for temps and Mhz increase. But it definitely helped 3D bench at 5.0 ghz stable (and with lower vcore), as opposed to 4.9 ghz 3d max.

Which makes me wonder why 99% people (even 99% of people here) even bother...

<- Risk averse.