Stability testing

almighty15 · 12 Jan 2014 at 01:08

Hey 8

I know you're not a firm believer in running Intel Burn Test and stuff for stability but is there a particular reason?

Surely if a chip can do infinite loops at stock it should be able to do them when overclocked too?

Asking because I can get my 2600k to pass Linx at 4.8Ghz at just over 1.4v

I can get into Windows and do pretty much every at 5Ghz with 1.45v but it will not pass any stress tests.

Rroff · 12 Jan 2014 at 05:14

A stress test typically only tests a CPU's ability to function in a specific way or a small range of tests and doesn't fully represent every day use.

However I find it a good way to save myself from potential data/OS corruption, etc. while finding the stable point that could otherwise result in spending a lot of time reinstalling the OS and/or recovering from backups/problems down the line. Once its given me a feel for the CPU's stability I can then narrow down where its likely to be fully application stable.

setter · 12 Jan 2014 at 06:40

Running endless loops of p95/ibt/LinX., only proves your system can run them. Fire up a game, bf3/4 or crysis 3 are good examples. Ive had them fail on the load screen but stress stable. The unrealistic heat they generate is another matter. For the cpu i find a couple of runs of asus real bench works well. Ok on ram too as ive found this week. Bumped my 2400mhz kit up to 2600mhz.

8 Pack · 12 Jan 2014 at 18:28

You need to stress the PC in the way you would normally use it. setter is along what I consider the correct lines. Game/encode/browse/unpack/ Move large files etc .. Normal stuff!!!

vonex · 15 Jan 2014 at 00:52

can you pass a run of real bench 2.0? i'm using this for quick test when testing out different clock speeds it will find an unstable clock pretty quick even in a quick test.... and then pretty much was setter has already said bf3/4 and crysis 3 will find an unstable clock after a few hours of playing.

setter · 15 Jan 2014 at 17:35

Ive been trying a few different settings for real bench.

4.5ghz cpu, 1.267v
4300mhz cache, 1.25v
2600mhz ram, 1.75v
11-12-12-32-193 2T
ioa, 1.25
iod, 1.25
sa, 1.25

The above completed 2 loops without issue but failed on super pi 32m with a rounding error, bsod on 0x50 when shutting the system down. Im gonna see if i can get the ram running 1T, same voltage but timings at 9-12-12-25-196

a66as · 16 Jan 2014 at 13:02

hi setter i noticed you have the same cpu + mobo as me, just a quick question, does your pc ever fail boot the first time and then immediately boot up again?

setter · 16 Jan 2014 at 13:30

Never had that on the current board. But did on an asus x58 p6t deluxe v2. Otherwise it ran fine, i7 920 @4.2ghz.