4770k oc-stability

[GeorgeZoo]

Hello everyone, i would like to ask the best program-version to stress my i7 4770k.
I have tried aida64 but it raises the max TDP to 100w so that causes the temperatures to rise even at 1.25v, also prime does that... So which programm keeps TDP in normal levels?
Also which programm are they using at TPU reviews to mark their oc as stable?

 

[cadaveca]

There is no singular program to use. You need to try many applications, or run the software you plan to use, and see what ya get.

you can check here for my suggestions:


http://www.techpowerup.com/reviews/Intel/Haswell_OC_Guide/

 

[GeorgeZoo]

There is no singular program to use. You need to try many applications, or run the software you plan to use, and see what ya get.

you can check here for my suggestions:


http://www.techpowerup.com/reviews/Intel/Haswell_OC_Guide/

Thats a lot of hours testing... i only use it for gaming...
also i tried to test aida64 FPU and it raises my temperatures at 98c within seconds and the TDP at 118w. My voltage is 1.25v. am i doing somehting wrong or my cpu doesnt like voltage?

 

[cadaveca]

It's your cooling, but also, 4770K does run quite hot with FPU-only testing. IF you can pass that, that's good. Try ASUS RealBench, 3DMark, maybe a game or two (I prefer BF4), and if it all passes, then typically you are good to go.

And yeah, it takes time to test an OC for any type of load, for sure. Just do as much as you feel comfortable with.

 

[vega22]

aida for heat, handbrake for stability.

 

[cadaveca]

aida for heat, handbrake for stability.

I've had stuff pass handbrake, then crash 15 minutes into a round of BF4. Handbrake uses cores well, but not memory sub-system.

 

[Batou1986]

There is no singular program to use. You need to try many applications, or run the software you plan to use, and see what ya get.

you can check here for my suggestions:


http://www.techpowerup.com/reviews/Intel/Haswell_OC_Guide/

Is there a reason no one uses linpac based tests because linpac seems to generate more heat and find more instability than anything else i've tried

 

[cadaveca]

Is there a reason no one uses linpac based tests because linpac seems to generate more heat and find more instability than anything else i've tried

I have used Linpack, and have found other options to either do just as well, or equal. Also, it simulates a specific load, which many users will not run into.

right from the software's site:

LINPACK is a collection of Fortran subroutines that analyze and solve linear equations and linear least-squares problems. The package solves linear systems whose matrices are general, banded, symmetric indefinite, symmetric positive definite, triangular, and tridiagonal square. In addition, the package computes the QR and singular value decompositions of rectangular matrices and applies them to least-squares problems. LINPACK uses column-oriented algorithms to increase efficiency by preserving locality of reference.

LINPACK was designed for supercomputers in use in the 1970s and early 1980s. LINPACK has been largely superceded by LAPACK, which has been designed to run efficiently on shared-memory, vector supercomputers.

The original goal of the LAPACK project was to make the widely used EISPACK and LINPACK libraries run efficiently on shared-memory vector and parallel processors. On these machines, LINPACK and EISPACK are inefficient because their memory access patterns disregard the multi-layered memory hierarchies of the machines, thereby spending too much time moving data instead of doing useful floating-point operations. LAPACK addresses this problem by reorganizing the algorithms to use block matrix operations, such as matrix multiplication, in the innermost loops. These block operations can be optimized for each architecture to account for the memory hierarchy, and so provide a transportable way to achieve high efficiency on diverse modern machines. We use the term "transportable" instead of "portable" because, for fastest possible performance, LAPACK requires that highly optimized block matrix operations be already implemented on each machine.

W1zz taught me well: don't make statements without having an explanation ready.

 

[GeorgeZoo]

It's your cooling, but also, 4770K does run quite hot with FPU-only testing. IF you can pass that, that's good. Try ASUS RealBench, 3DMark, maybe a game or two (I prefer BF4), and if it all passes, then typically you are good to go.

And yeah, it takes time to test an OC for any type of load, for sure. Just do as much as you feel comfortable with.

I have watercooling, more specific raijitech triton, that had good results on 4670k @4.5 1.320v. I dont believe its the cooling though.

 

[GreiverBlade]

It's your cooling, but also.

i have the same AIO: Raijintek Triton with 2 Corsair SP120 on it and i rarely see more than 75° in heavy load, ok i have a Devil's Canyon (4690K) ... that might explain but! check the seating of the block or change the thermal past (you're still using the stock bundled with? )also room temp can affect ... i live in mountain.

 

[GeorgeZoo]

i have the same AIO: Raijintek Triton with 2 Corsair SP120 on it and i rarely see more than 75° in heavy load, ok i have a Devil's Canyon (4690K) ... that might explain but! check the seating of the block or change the thermal past (you're still using the stock bundled with? )also room temp can affect ... i live in mountain.

no i am using arctic silver 5. its hot in Greece now but the temps are pretty high..

 

[GreiverBlade]

no i am using arctic silver 5. its hot in Greece now but the temps are pretty high..

well AS5 being a good (for me it's a So-So past, i wouldn't use it, but i have my reason) it should not be a problem, since i get my temp using the stock past (changing for Thermal Grizzly Hydronaut the next time i do a rig cleaning)
check the block seating then (might be a screw or a shifted contact, i got that with a Phobya UC-2LT block once and that was the only time i saw my 4690K going higher than 87° )

 

[vega22]

I've had stuff pass handbrake, then crash 15 minutes into a round of BF4. Handbrake uses cores well, but not memory sub-system.

for sure, but i dont see any mention of ram overclocking so i aint sure of the relevance.

realbench is a good real world test, but i have had it stress test all night and still crashed in no time with handbrake.

since haswell it has become my default cpu test as it finds instability where others do not.

 

[cadaveca]

for sure, but i dont see any mention of ram overclocking so i aint sure of the relevance.

realbench is a good real world test, but i have had it stress test all night and still crashed in no time with handbrake.

since haswell it has become my default cpu test as it finds instability where others do not.

If it works for you, that's great. It isn't good enough for me, personally.

That's the fun part of OC, we all use our OCs differently, so we all need to test differently. There is no single app that is good enough for everyone.

I have watercooling, more specific raijitech triton, that had good results on 4670k @4.5 1.320v. I dont believe its the cooling though.

OH, but it is though. No AIO is capable of deal with AVX2 workloads. NO water loop, either. You need dry ice and LN2. 4770K gets a voltage boost when specific workloads are running, and this can overwhelm most coolers since the TIM used under the heatspreader is a bit too thick. This was done intentionally so that such workloads would lead to throttle well before damaging the CPU. You can de-lid and replace with your own TIM and see a drastic drop in temps, or ignore those workloads.

 

[vega22]

not sure i follow you now dave :s

having a stable oc isn't good enough for you, personally?

interesting concept, i imagine unstable oc are much easier to achieve xD

while that statement makes 0 sense to me i think i get the point you're making.

if all you do is game then gaming is a good enough stability test. if you encode, render and compile you need to be able to do those things so they work as a stress test too.

i aint sure i subscribe to that line of thought, as if anything can crash your system it aint stable imo. but yea, horses for courses.

 

[cadaveca]

Oh, I agree with your sentiments, but I can fire up BF4 and crash after running Linpack for 24 hours. So to me... it's useless. It tests specific workloads that my other testing covers already. I use Handbrake for encoding, since that's what I use to encode.

You can suggest that my way of testing is inferior to yours, if it makes ya feel good. No sweat off my back! ROFL.

 

[Schmuckley]

Hello everyone, i would like to ask the best program-version to stress my i7 4770k.
I have tried aida64 but it raises the max TDP to 100w so that causes the temperatures to rise even at 1.25v, also prime does that... So which programm keeps TDP in normal levels?
Also which programm are they using at TPU reviews to mark their oc as stable?

For a quick test,use Cinebench 11.5 or 15.That will get you in the ballpark.
http://www.maxon.net/products/cinebench/overview.html

Then for real stability use Hyper-Pi 32m.It stresses the CPU and RAM really hard for about 8 minutes.If it passes 2 rounds,chances are slim you'll ever get a BSOD out of nowhere.(Unless it's sleep-related)
http://files.extremeoverclocking.com/file.php?f=211