4yr old system just failed IBT. Never had any problems...

Slackjaw190 · 24 Oct 2011 at 21:40

Using my old rig atm, e6600 @ 3 GHz

When I built it, it was 8hrs stable with prime 95 but with my new rig I've been using IBT for initial stability testing so I thought I would give it a go with my old rig out of interest.

It failed after 2 cycles... lol. Been using it for 4 yrs and I can’t think when it’s ever crashed. Played games for hours, video edited which maxed out cores for ages no problem.

Is IBT not designed for rigs this old, have I been running an unstable rig for years or is IBT not a true test to stability?

Just out of interest

Slack

weird_dave · 25 Oct 2011 at 09:00

IBT is a perfectly valid stability test for older systems. It pounds the CPU a lot harder than games and video editing, your system is borderline unstable in my opinion.

Zeed · 25 Oct 2011 at 15:59

In my opinion its stable. If it runs fine for over 4 years.....

ThomasTheTankEngine · 25 Oct 2011 at 16:56

Don't forget that IBT will heat your cpu higher than 99% of normal programs.

It's quite possible that your cpu is 100% stable at normal operating temperatures and only becomes unstable when it gets very hot.

There's also the fact to consider that a lot programs are much more forgiving than stress tests, and don't usually crash when they encounter the occasional error. Eg. you might get an error whilst playing a game, which causes a texture to get borked for a single frame, and then get fixed on the next one.

Slackjaw190 · 25 Oct 2011 at 18:37

Then why are we all obsessed with passing these stress tests if your pc is never going to get that stressed.

Also I dont think it was the temps, the highest the cores got was 52°C

Hotwired · 25 Oct 2011 at 18:43

Stress tests in all environments are significantly above what the expected load on the system will be.

You pressure test a pipe so if there's a chance the system goes above normal pressure there's a safety margin.

You stress test a computer you've modified the CPU settings on so that when you play the odd game which may well heavily stress it, it won't crash out on you.

If you're not going to play any kind of game or significant workload then don't bother overclocking.

weescott · 27 Oct 2011 at 00:13

Stress tests should only be relevent to the required level of stability. For example a PC that is Prime 95 stable can be made unstable with other tests.

CPU stable Vs system stable Vs required stable = you decide.

drunkenmaster · 27 Oct 2011 at 04:10

Slackjaw190 said:
Then why are we all obsessed with passing these stress tests if your pc is never going to get that stressed.

Also I dont think it was the temps, the highest the cores got was 52°C

I'm not, and peopel shouldn't be, its like freaking Furmark and people being obsessed with being stable in it. If you can run, every day, all day for 4 years with say an overclock of 950Mhz core/1300Mhz mem on your graphics card, no game crashes, some nice gpu accelerated encoding is fine, but Furmark is only stable at 900Mhz/1250mhz...... why the hell should you care.

Simple answer is, you shouldn't.

I overclock till my system is stable for anything I ACTUALLY USE, I won't install extra benchmarks/stress tests because the only thing IBT tells you, is if your system is stable IN IBT, it does NOT tell you if your system is stable in other software.

Would anyone not buy a car because some dude bought it and decided to test if an elephant can sit on it without collapsing the roof, and the roof failed? No, if the car works as you'd use it, but isn't a magic car and can't do something, you don't need it to do anyway, who cares.

For me the guys who boast about their overclocks being prime stable or something, are as bad as the guys who buy a 990x, as an upgrade to their 980x, and tell everyone all about it under the assumption everyone else thinks they are somehow special for being able to use a website to order a product online

ATIorNvidia · 27 Oct 2011 at 04:52

I've passed Furmark for 12 hours, only to crash in games within 5 minutes. I've also passed LinX/IBT and failed Prime95, and vice versa.

Stability in overclocking is subjective. People that say running Prime95/LinX/IBT for 24 hours guarantee's 100% stability are idiots. How do you know you wouldn't fail at the 25th or 26th hour?

If it's stable enough for your uses, then that's all that matters.

drunkenmaster · 27 Oct 2011 at 05:03

ATIorNvidia said:
I've passed Furmark for 12 hours, only to crash in games within 5 minutes. I've also passed LinX/IBT and failed Prime95, and vice versa.

Stability in overclocking is subjective. People that say running Prime95/LinX/IBT for 24 hours guarantee's 100% stability are idiots. How do you know you wouldn't fail at the 25th or 26th hour?

If it's stable enough for your uses, then that's all that matters.

Yup, i overclock, then I play or just use the computer, if it crashes, i back off, or generally tweak till I'm happy. I mean loads of people completely fail to realise that IBT/prime "stability" means nothing when you play a game, because you're putting completely different stresses on the system. On say a 480gtx you can go from idle in prime, to using another 200 watts in a game, and wonder why the system is unstable. Voltages drop slightly due to the higher load, higher chipset, cpu, memory temps, higher psu temps causing higher ripple and less stable voltages, etc, etc.

I have to say in general if I'm stable in games after I put a system together, its very rare to find any other real world situations a computer won't be stable in.

Zeed · 27 Oct 2011 at 12:44

I am not MAX STRESS MANIAC.
For me system needs 2 run prime for 1 hour and 3dmaks at max settings for 1 hour.
Its like with power line fails..... Your system can run 50 hours of stress tests and will still reboot if power goes down for 1 second iat your place.
So should all of us buy UPS ?????. This year in nottingham i had 3 power losses so far....