or Connect
AVS › AVS Forum › Video Components › Home Theater Computers › Intel Haswell LGA 1150 (Socket H3) Stock Cooler Noise
New Posts  All Forums:Forum Nav:

Intel Haswell LGA 1150 (Socket H3) Stock Cooler Noise - Page 3

post #61 of 177
I have not tested over night... But I did bench it for an hour and so far... so good.

Temps get HOT eek.gif
post #62 of 177
5.0GHz is the impressive number. Chips that can easily hit 50x are the golden ones. The chips that don't hit 4.6-4.8GHz easily are duds. I've got a 3770k here that's a dud...
post #63 of 177
5.0ghz = crash. mad.gif
post #64 of 177
This thing cruises around at 800mhz or 1200mhz if I change the power options to balanced from high performance in windows. lol.

Temps in the 20's too.
post #65 of 177
Quote:
Originally Posted by Mfusick View Post

5.0ghz = crash. mad.gif

Yeah it's pretty rare to hit 5.0 without some tweaking.
post #66 of 177
What were your temps exactly after an hour of stress testing?

4.8Ghz on Haswell with acceptable vcores is a marvelous feat. Very few chips will reach that frequency (probably the top 10%) without melting the whole system. Congrats on receiving such a gem of a chip.
post #67 of 177
I think it was 85 or so... It seemed ok. (I am not sure what "good" temps are honestly)

I used to thermal throttle my E8500 mercilessly. 100c seemed ok back then lol. My Sandy with the same cooler ran cooler (70c)

But that is after an hour of prime95.

Normal use is around 40c overclocked and 30c stock clock.
post #68 of 177
Well I'd personally say anything under 90*C on the hottest core under Prime95 (27.7 or later) or IntelBurnTest/Linx (with AVX) is ok.
If you see that kind of temps under those kinds of loads you'll rarely see those temps under normal loads, I'd like to say never though. P95 and Linpack are just plain brutal.

P.S. My 3770K could be stable under 5.0Ghz. Haven't actually tried any higher multiplier to see if it can boot those.
I've done some benching at 5Ghz but stable, nah, not with my cooling. I'd need to get some better cooling.. True Spirit and a lonely GT AP15 isn't going to cut it.
I use my 3770K for encoding and such and so my storage is the bottleneck even at just 4.3Ghz. I currently have it at 4.5Ghz just because it's a nicer number than 4.3...biggrin.gif
post #69 of 177
4.5 is a good daily driver biggrin.gif
post #70 of 177


Not even bad temps under 4.5, I'm currently using offset and haven't bothered messing around much with it. With manual voltage it's stable with 1.21v but right now with offset it's getting 1.23-1.248v depending on load. Couldn't care less though..biggrin.gif
With IBT it maxes 77 on the hottest core while running it at very high load.
During encoding I see it hovering in the mid/high 60's. GT 1850 hovers between 1100-1200rpm.. Can't even hear it unless it speeds up to ~1400.. Case fans (Slip Stream PWM 120mm x2 + 140mm TR X-Silent) ~600-700rpm 24/7.
post #71 of 177
Quote:
Originally Posted by renethx View Post

IVB suffers from a thermal interface material (TIM) problem (google it if you haven't heard of it).

Actually, the TIM is okay (the TIM is better than the best aftermarket material). The main issue is that the gap between the die and the heatspreader is relatively large on many (most?) units, and that results in a thicker layer of TIM and lower thermal conductivity. Some units are lucky and the gap is smaller (probably the tolerance on the IHS "legs" is such that you get a few with short legs which makes the gap smaller).

(google if you haven't seen the report from a guy who carefully tested this)
post #72 of 177
Yeah jim2100. Exactly.
Here's the thread by Idontcare
http://forums.anandtech.com/showthread.php?t=2261855
And here's the gap post
http://forums.anandtech.com/showpost.php?p=34053183&postcount=570
It's as scientific as it can get but seriously though, when's the last time you've seen 10*C+ differences between different TIM if the application is good on both? It's like if you didn't tighten your cpu coolers screws enough and see bad temps, tighten it a bit and boom temps drop.

By the way, I think Mfusick's chip is one of those where the gap isn't big enough to cause issues when you increase the clock speed.(and hence increase the thermal load)
post #73 of 177
Quote:
Originally Posted by jim2100 View Post

Actually, the TIM is okay (the TIM is better than the best aftermarket material). The main issue is that the gap between the die and the heatspreader is relatively large on many (most?) units, and that results in a thicker layer of TIM and lower thermal conductivity. Some units are lucky and the gap is smaller (probably the tolerance on the IHS "legs" is such that you get a few with short legs which makes the gap smaller).

(google if you haven't seen the report from a guy who carefully tested this)

Quote:
Originally Posted by lagittaja View Post

Yeah jim2100. Exactly.
Here's the thread by Idontcare
http://forums.anandtech.com/showthread.php?t=2261855
And here's the gap post
http://forums.anandtech.com/showpost.php?p=34053183&postcount=570
It's as scientific as it can get but seriously though, when's the last time you've seen 10*C+ differences between different TIM if the application is good on both? It's like if you didn't tighten your cpu coolers screws enough and see bad temps, tighten it a bit and boom temps drop.

By the way, I think Mfusick's chip is one of those where the gap isn't big enough to cause issues when you increase the clock speed.(and hence increase the thermal load)

Regardless of whether it's a TIM quality issue or a TIM thickness issue it's still technically a TIM issue that requires removing the IHS and reapplying. When I first de-lidded my 3770K I used the Intel TIM that comes from vendors for CPU repair. Supposedly the same stuff used during manufacturing when the soldering method isn't used. My temps dropped ~3-5 degrees. Tried 3 times with the same results. When I used Coollaboratory Liquid Ultra 100% my temps dropped well over 20 degrees. I was running in the high 90's at 5.0GHz ~1.575vcore and after replacing with Liquid Ultra temps were mid to upper 60's.

Point is... it's a BS issue (regardless of the exact cause) that Intel should have addressed and it sucks that the enthusiast customer needs to take such risks just to get decent temps.
post #74 of 177
Quote:
Originally Posted by itznfb View Post


Regardless of whether it's a TIM quality issue or a TIM thickness issue it's still technically a TIM issue that requires removing the IHS and reapplying.

No, it is not a TIM issue. It is the heatspreader (IHS) height tolerance, as I already explained.

There is little reason for Intel to spend extra money to make the tolerance better since the issue only matters to a small fraction of consumers.
post #75 of 177
Quote:
Originally Posted by jim2100 View Post

No, it is not a TIM issue. It is the heatspreader (IHS) height tolerance, as I already explained.

There is little reason for Intel to spend extra money to make the tolerance better since the issue only matters to a small fraction of consumers.

You can sit there and circle jerk your semantics all you want. Bottom line is that the Intel TIM needs to be replaced in order to get normal temperatures. Reducing the gap and still using Intel's TIM doesn't solve the issue.
post #76 of 177
Quote:
Originally Posted by itznfb View Post

Bottom line is that the Intel TIM needs to be replaced in order to get normal temperatures. Reducing the gap and still using Intel's TIM doesn't solve the issue.

Wrong again. As I already explained, Intel's TIM is better than the best aftermarket material. Using the same gap with aftermarket thermal paste will not improve the thermal conductivity of the assembly. The key is to reduce the gap.
post #77 of 177
But you still end up changing the TIM..
So whatever the issue is, the issue is there. No point in arguing.
post #78 of 177
It kinda defeats purpose of a great cooler
post #79 of 177
Quote:
Originally Posted by jim2100 View Post

Wrong again. As I already explained, Intel's TIM is better than the best aftermarket material.

That's complete BS. Intel uses the cheapest crap they can find at any given moment. I just checked your post history and you do nothing but troll so you're going on my ignore list so don't be surprised when I don't respond to your crap ever again. rolleyes.gif
post #80 of 177
Quote:
Originally Posted by Mfusick View Post

It kinda defeats purpose of a great cooler

Thoughts?
post #81 of 177
Quote:
Originally Posted by Mfusick View Post

Thoughts?

On whether you should pop the top or not? Depends on your comfort level and if you're willing to take the hit on a lost chip.
post #82 of 177
If I were him, I wouldn't bother. If he's got like 4.6Ghz stable with nice volts and temps then I don't really see the need to go and pop it open and loose the warranty. Not even mentioning the possibility of damaging the chip. At least in x264 a ~4.4Ghz Haswell is about on par with 4.8Ghz Ivy..
post #83 of 177
Yeah.. I was considering chasing clock speed with a better cooler but since it might not help I think I will give up before I start. Makes sense. I am happy and don't need more speed anyways.

This thing is fast even stock.

I do a little encoding but nothing serious enough to justify overclocking seriously. I don't mind clicking a couple settings in bios but removing my CPU lid just seems crazy.
post #84 of 177
Quote:
Originally Posted by itznfb View Post

Intel uses the cheapest crap they can find at any given moment.

I have no idea how expensive the TIM is that Intel uses -- that is another issue.

But in the link already given in this thread, equal thickness of Intel's TIM and an aftermarket thermal paste were compared, and the Intel TIM resulted in a lower temperature.

It is odd that you blame the TIM for the issue. How do you account for the people who have a unit that has better thermal conductivity? Did Intel just randomly insert some alternate TIM with higher thermal conductivity in some of the packages?

It actually makes a lot more sense that it is the height tolerance of the heatspreader, since then it is easy to see that some fraction of the parts (in the tail of the Gaussian distribution) will have a shorter heatspreader which results in a thinner TIM thickness.
post #85 of 177
Quote:
Originally Posted by Mfusick View Post

Yeah.. I was considering chasing clock speed with a better cooler but since it might not help I think I will give up before I start. Makes sense. I am happy and don't need more speed anyways.

This thing is fast even stock.

I do a little encoding but nothing serious enough to justify overclocking seriously. I don't mind clicking a couple settings in bios but removing my CPU lid just seems crazy.

I've seen temp drops of up 30 degrees on Ivy by popping the top. The OC gain from those drops are minimal at best. Most of the time popping the top won't get you a better OC. I would only recommend it if you want significant temp decrease at current speed regardless of OC.

Sent from my Nexus 10 using Tapatalk HD
post #86 of 177
I have the same fan revving issue with an i5-2500K machine I recently built. When starting up or coming out of sleep mode, the stock CPU fan will rev loud for several seconds, same as in the posted video, before it settles down to a quite rpm. I have looked online to find a way to avoid this loud revving, but most persons reply it is normal. No, it is not normal. None of my other machines does it. Not my i5-2500 non K machine with stock fan, and not my i3-2100 machine with stock fan. Wondering if anyone has a solution to the CPU fan revving?
post #87 of 177
All of my rigs do that. It's normal..

Sent from my GNote
post #88 of 177
Quote:
Originally Posted by itznfb View Post

Quote:
Originally Posted by Mfusick View Post

Yeah.. I was considering chasing clock speed with a better cooler but since it might not help I think I will give up before I start. Makes sense. I am happy and don't need more speed anyways.

This thing is fast even stock.

I do a little encoding but nothing serious enough to justify overclocking seriously. I don't mind clicking a couple settings in bios but removing my CPU lid just seems crazy.

I've seen temp drops of up 30 degrees on Ivy by popping the top. The OC gain from those drops are minimal at best. Most of the time popping the top won't get you a better OC. I would only recommend it if you want significant temp decrease at current speed regardless of OC.

Sent from my Nexus 10 using Tapatalk HD

at 4800mhz I could not use my current cooler/set up without thermal throttling and crashes. I could set it at 4800mhz and cruise around AVS, but trying to run a stress or bench results in a BSOD mad.gif

That said, I'm happy at 4600mhz being the best. Seems like this is better than many are seeing with a 4770k. 4400mhz is my "safe zone" as temps seem to remain pretty reasonable for me.
post #89 of 177
I remember when a 1000+ MHz overclock was the holy grail. Now it's a default starting point...
post #90 of 177
Thread Starter 
Alright, my experiment with the stock cooler is over. While temps weren't all that terrible overall for a stock cooler (average in the mid to upper 50C range for light work/idle, close to 80C at load), the fan constantly kicking up and down in order to keep the temps under control really irks me. It's time for this build to run under the cooler that I bought just for this build, the Noctua NH-L9i. I will also be replacing the stock 80mm case fan with a Noctua 80mm fan. I was going to cut a hole in the back of the case in order to fit a 92mm Noctua fan, but, it just didn't work out in the measuring process. Oh well, cooler, quieter days ahead. biggrin.gif
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Home Theater Computers
AVS › AVS Forum › Video Components › Home Theater Computers › Intel Haswell LGA 1150 (Socket H3) Stock Cooler Noise