• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Retail 4770k from microcenter OC results

Womper said:
It doesn't sound like you're putting a heavy load on the CPU. Prime95 small-fft should do the trick.
Click to expand...

especially the more recent versions of prime95 that use avx. older versions that don't use avx won't stress the cpu as hard.

linx with the latest linpack binaries seems to be even more demanding than prime95. my system that was running prime95 stable for hours blue screened within five minutes running linx.

I can play games all day @ 4.8GHz with voltage peaking at 1.3v and temps at 65C. But turn on Prime95, and my voltage blasts off to 1.5v.
Click to expand...

are you using adaptive voltage?
 
Default 10-passes of IBT. Quick and simple test of stability.
Follow up with a 100-pass run. If it survives (and doesn't literally go nuclear), slap on long (12/24/++ hours) runs of Prime95, 3DMark or what have you.
 
livefastdieyoung said:
especially the more recent versions of prime95 that use avx. older versions that don't use avx won't stress the cpu as hard.

linx with the latest linpack binaries seems to be even more demanding than prime95. my system that was running prime95 stable for hours blue screened within five minutes running linx.



are you using adaptive voltage?
Click to expand...

Ah that's very true. It's those AVX versions of Prime95 and linpack that really crank the heat up.

And yeah, adaptive voltage for sure. It keeps temps and voltages as low as possible. The FIVR does a great job of keeping the chip stable while moving around different voltage and frequency steps.
 
Is anyone bumping up thier cachce or just leaving it at 39? I see all kinds of OC's but very little info is given about the cache settings everyone is using.. and the cahce voltage used to get stability.
 
Copyright said:
Is anyone bumping up thier cachce or just leaving it at 39? I see all kinds of OC's but very little info is given about the cache settings everyone is using.. and the cahce voltage used to get stability.
Click to expand...

I've tried, but I must be an idiot. I changed Uncore multiplier in BIOS to 43x but NB frequency still shows as 3900 MHz in CPU-Z and Aida64.
 
Copyright said:
Is anyone bumping up thier cachce or just leaving it at 39? I see all kinds of OC's but very little info is given about the cache settings everyone is using.. and the cahce voltage used to get stability.
Click to expand...
I tried bumping mine up to 43 x 101, but it didn't seem to make a discernible difference on quick CineBunch runs (compared to leaving it at 39 x 101).
 
anyone see a before/after from delidding? Is it pretty much the same results as with IB?
 
My CPU is not from MC but I run 1.25v adaptive voltage @ 4.5 with 4300 uncore but I needed 1.15 CPU ring voltage to get it stable. Also running 4 sticks of ram at 1866 9-9-9-18- 1T
 
Rattle said:
My CPU is not from MC but I run 1.25v adaptive voltage @ 4.5 with 4300 uncore but I needed 1.15 CPU ring voltage to get it stable. Also running 4 sticks of ram at 1866 9-9-9-18- 1T
Click to expand...

Sounds like mine and yours is real close... 1.25V adaptive is A LOT if stressed wtih Aida64 tho. I honestly did not find much of a temp difference using adaptive over manual. I really expected Adaptive to drop my temps more than it did and I only ran 1.22V. I need 1.255 manual for stability. You are probably getting 1.34V when stressing using aida64. Mine instantly starts to throttle at that voltage. Maybe intel is using crappy paste again between the core and IHS. I may de-lid mine and try for better temps. I could probably hit 4.6 or more if I could control the heat. Heat got me even at 1.275V and these should be able to run 1.3V without issue.
 
I am unhappy with my 4770K on Asus Z87 PRO, at STOCK my TEMPS are 59c LOAD. My Volt is auto around 1.120 . My 3770k overclocked much better and remained cooler.
 
Got my 4770K from Microcenter. Installed it and set the multiplier to 46, everythign else on auto, litterally a 30 second overclock. Using an Asus P87 Pro board.

Seems stable. I've ran LinX for a couple hours with no problems. Max temp I've seen is 82c using my Swiftech H220. My CPU automatically switches between 800MHz 0.75v and 4.6GHz 1.3v and everywhere in between.

I'm really happy with Haswell. Im seeing performance about equal to a 5.1 GHz Ivy bridge, like a 10.06 score for Cinebench 11.5.
 
MCRAYGSX said:
I am unhappy with my 4770K on Asus Z87 PRO, at STOCK my TEMPS are 59c LOAD. My Volt is auto around 1.120 . My 3770k overclocked much better and remained cooler.
Click to expand...

Just like many others here you haven't stated your ambient Temps.

It makes a HUGE difference if you are sitting in 33C room over a 15C room.

For all we know the last time you took any notice of your 3370K it was winter and now its summer!,
 
Copyright said:
I think I am going to try this!


Only TIM I have is Tuniq TX-4. It may be a little to thick to spread right under the IHS? Anyone know which TIM is best or if TX-4 is good enough?
Click to expand...

Use this between IHS and core: http://www.frozencpu.com/products/3...ry_Liquid_PRO_Thermal_Interface_Material.html

Use any decent tim between core and heatsink.

my delid. Cant test bc all my component wont be in until next week.

1000
 
inglewood78 said:
Use this between IHS and core: http://www.frozencpu.com/products/3...ry_Liquid_PRO_Thermal_Interface_Material.html

Use any decent tim between core and heatsink.

my delid. Cant test bc all my component wont be in until next week.

1000
Click to expand...

What method did you use? Why not use Ultra? I'm still trying to find it but I heard ultra was easier to spread.

MCRAYGSX: What method of TIM application did you use? My temps were poor but that was due to bad airflow and improper TIM application. Remember, don't spread the TIM, make it into a pea/corn kernel shape and let the heat sink spread it out.
 
I can't even find the Ultra in stock anywhere and from what I have been reading the difference between the 2 are almost nothing. It's funny you linked that .. I just ordered some! lol
 
This heat sucks. I can boot and run cinebench at 4.6HT with 1.25V, but when it comes to prime small ffts, 4.3HT 1.15V hits 90C. 21-22C ambient.


Delidding tomorrow.

I wouldn't even try to use CLP or CLU with Haswell. Any touches the VRMs when the IHS slides around when clamping, and your CPU will be toast.
 
Cecil said:
This heat sucks. I can boot and run cinebench at 4.6HT with 1.25V, but when it comes to prime small ffts, 4.3HT 1.15V hits 90C. 21-22C ambient.


Delidding tomorrow.

I wouldn't even try to use CLP or CLU with Haswell. Any touches the VRMs when the IHS slides around when clamping, and your CPU will be toast.
Click to expand...

Yeah, I was kinda wondering that also. I know it gets good results tho. I don't have any other paste that would work well. I got some TX-4 and some Prolimatech paste that came with a heatsink I got years back. Not even sure what to use.
 
Cecil said:
This heat sucks. I can boot and run cinebench at 4.6HT with 1.25V, but when it comes to prime small ffts, 4.3HT 1.15V hits 90C. 21-22C ambient.


Delidding tomorrow.

I wouldn't even try to use CLP or CLU with Haswell. Any touches the VRMs when the IHS slides around when clamping, and your CPU will be toast.
Click to expand...

Just coat the VRMs with nail polish cover. You can always remove it later with a little acetone.
 
SkeL said:
What method did you use? Why not use Ultra? I'm still trying to find it but I heard ultra was easier to spread.

MCRAYGSX: What method of TIM application did you use? My temps were poor but that was due to bad airflow and improper TIM application. Remember, don't spread the TIM, make it into a pea/corn kernel shape and let the heat sink spread it out.
Click to expand...

Vice and hammer. I wasn't sure what to expect but I had to hit it pretty hard with the hammer :eek:. This method is not for the faint of heart, neither is the razor method.

I feel pro and ultra are interchangeable even though the ultra is suppose to the the successor to the pro. The pro has a needle for easier application but the ultra has a brush for easier spreading. There's more experience with the pro so I just went with that. Either would work just fine. Just don't use regular TIM (Mx-4, AS, etc)

Mobo and my custom water loop will be in later this week and hopefully I can finish this is a few weeks. Will post up results. I'm shooting for 5.0 and planning on pushing around 1.4v to get there max.
 
SirMaster said:
Just coat the VRMs with nail polish cover. You can always remove it later with a little acetone.
Click to expand...

You can also just use some non conductive TIM (MX-2/4) and spread it over the VRMs just in case the coollabs application gets messy. The Coollabs pro and ultra provide significant improvement in temps. Based on what I've read, Using regular tim between core and IHS provides around 5c improvement in temps, while pros and ultra provide 10-20c drops (depending on application). Risk/reward.
 
Last edited:
If you apply it right, I got a 12C drop on 3570k with MX-2.
 
inglewood78 said:
Vice and hammer. I wasn't sure what to expect but I had to hit it pretty hard with the hammer :eek:. This method is not for the faint of heart, neither is the razor method.

I feel pro and ultra are interchangeable even though the ultra is suppose to the the successor to the pro. The pro has a needle for easier application but the ultra has a brush for easier spreading. There's more experience with the pro so I just went with that. Either would work just fine. Just don't use regular TIM (Mx-4, AS, etc)

Mobo and my custom water loop will be in later this week and hopefully I can finish this is a few weeks. Will post up results. I'm shooting for 5.0 and planning on pushing around 1.4v to get there max.
Click to expand...

Surprised you say that. There are videos of guys getting 20C drops in temps using mx-2 between the core and IHS.
 
Cecil said:
Delidding with MX-2 dropped 14C at 4.3 with HT 1.15V.
Click to expand...

Hmm...interesting. Most of the other delids using reg tim didn't do much. I guess maybe they applied it wrong? Anyways, I'll be pushing some good volts so I need all the temps I can get.
 
Couldn't run small ffts for more then 2-3 seconds before delidding with 1.22v at 4.5 with HT on (would hit 95C then BSOD or I manually stopped it). Now ran it for 10 minutes with 1.225v HT on and only got to 83C.
 
I'm at 4.4Ghz with 1.170v stable now, I think thats not too bad :)

What I noticed in regards to Adaptive vs. Manual voltage mode is that my idle temperatures are pretty much the same. Manual would be 1.170v for me, idle voltage with Adaptive mode 1.017. Cores are hovering around 28°C with either. Maybe I will just stick to Manual mode unless there are any problems running 1.170v 24/7.
 
Copyright said:
Sounds like mine and yours is real close... 1.25V adaptive is A LOT if stressed wtih Aida64 tho. I honestly did not find much of a temp difference using adaptive over manual. I really expected Adaptive to drop my temps more than it did and I only ran 1.22V. I need 1.255 manual for stability. You are probably getting 1.34V when stressing using aida64. Mine instantly starts to throttle at that voltage. Maybe intel is using crappy paste again between the core and IHS. I may de-lid mine and try for better temps. I could probably hit 4.6 or more if I could control the heat. Heat got me even at 1.275V and these should be able to run 1.3V without issue.
Click to expand...

Well I knew about the additional .1v adaptive adds while stressing with avx instructions.
I found my 4.5 sweet spot with adaptive on and minus .1v so 1.15 would give me 1.25 while running prime. The I went back and set it to 1.25 and that's the max voltage for everything besides prime 95 and anything else which would use avx. Which I have no idea what would use those Lol.

I've thrown everything at it including 30 minute handbrake runs at 4.5 @ 1.25v without issue.
Handbrake got it to 75c easy. That's the most intense thing my 4770k will do besides BF3 which hits 60c with H100 on low.
 
I have some Prolima tech paste I was thinking of using.. came with my heatsink a few years back. I assume its PK-1 paste but it does not say it is. From what I have read PK-1 is great thermal paste. Anyone know if they had used something previous to PK-1 or was it always PK-1 , PK-2 etc.. ?
 
livefastdieyoung said:
at idle, cpu drops into c1e (enhanced halt state), so temps and power consumption aren't affected much by voltage.
Click to expand...

That makes sense.. I noticed the same thing he did. For me knowing that I don't think I will even risk adaptive. Manual works like a champ for me.
 
4770K HT off 4.5GHz 1.225v vs 3570K 4.5Ghz 1.225v, both delidded with MX-2, Haswell runs 2C warmer above ambient (just my sample case obviously).
 
So far, I lowered the cache speed (but not the voltage). There was a tiny performance drop (close to the margin of error), and no temp drop.

4.5 no HT 45x cache
Cinebench 7.77
Spi 1m 8.047s
Spi 8m 1m 29.468s

4.5 no HT 39x cache
Cinebench 7.75
Spi 1m 8.078s
Spi 8m 1m 30.157s


Going to try 39x with lower cache voltage to see if it effects temps.
 
Cecil said:
So far, I lowered the cache speed (but not the voltage). There was a tiny performance drop (close to the margin of error), and no temp drop.

4.5 no HT 45x cache
Cinebench 7.77
Spi 1m 8.047s
Spi 8m 1m 29.468s

4.5 no HT 39x cache
Cinebench 7.75
Spi 1m 8.078s
Spi 8m 1m 30.157s


Going to try 39x with lower cache voltage to see if it effects temps.
Click to expand...

Why no HT?
 
You must log in or register to reply here.
Back
Top