![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
IntelBurnIn question
- Thread starter groebuck
- Start date
How much Ram do you have and are you on a 32-bit OS? I had same problem in a recent build with 4gb or RAM and Vista 32. Prime and OCCT tested fine, but Intel Burn Test would get errors, although for me it was instant.
For me, it was when running the test you pick the stress level on the RAM - if you pick maximum IBT tries to use 4gb but the OS only recognizes 3gb due to the 32-bit OS limits and Burn Test gets errors. At least thats my theory, reducing the level or ram to stress to 50% so the test only tried to use half the available memory resulted in burntest passing.
If you are prime stable, I wouldn't be worried.
For me, it was when running the test you pick the stress level on the RAM - if you pick maximum IBT tries to use 4gb but the OS only recognizes 3gb due to the 32-bit OS limits and Burn Test gets errors. At least thats my theory, reducing the level or ram to stress to 50% so the test only tried to use half the available memory resulted in burntest passing.
If you are prime stable, I wouldn't be worried.
Operaghost
[H]ard|Gawd
- Joined
- Jun 4, 2004
- Messages
- 1,315
I personally give two shits whether or not my OC is "stable" in Intel Burn or even Prime.
If it never crashes or hangs during your normal uses, then who really gives a fuck?
Unless you're trying to break some kind of OC record.
If it gets the job done it gets the job done. Those stress tests put your system through a load that you will most likely never put it through. Only if you're doing heavy animation/video/graphics editing will you even come close.
If it never crashes or hangs during your normal uses, then who really gives a fuck?
Unless you're trying to break some kind of OC record.
If it gets the job done it gets the job done. Those stress tests put your system through a load that you will most likely never put it through. Only if you're doing heavy animation/video/graphics editing will you even come close.
If you don't care about stability, then that's your prerogative. But don't pretend that your PC is actually stable if it doesn't crash during general usage but fails Prime and/or IBT. Some people actually do care about having a fully stable PC.
If your machine is unstable enough to fail a calculation error in Prime or IBT it will most likely fail encoding video, compressing files, etc.
nissanztt90
2[H]4U
- Joined
- Apr 13, 2007
- Messages
- 3,858
When i was messing around with OCing my Q9650, i had changed a BIOS voltage setting and forgot about it. Ran all my games fine for plenty long enough, except for GTA4. Couldnt figure out why i kept getting a crash. Ran IBT...failed it. Checked my BIOS and realized i had the wrong setting...changed it back...GTA4 ran fine after that, as well as IBT passing.
Joe Average
Ad Blocker - Banned
- Joined
- Apr 6, 2008
- Messages
- 15,459
Did anyone ever stop to think that perhaps IBT itself is flawed, badly coded, and that might account for everything else maxed out and stable without issues for hours on end but that one single particular utility craps out?
Funny, nobody ever does... I don't trust any single piece of software to tell me a system is stable. Daily regular usage of the whole box and every piece of software that I normally use tells me that when it doesn't freeze, overheat, or randomly reboot.
Why do people make this so fuckin' tough on themselves?
One app is not an indicator of stability - using the PC without issues is.
nissanztt90
2[H]4U
- Joined
- Apr 13, 2007
- Messages
- 3,858
Are you serious? These torture tests force the PC to do calculations over and over again. If it isnt returning the correct result...there is an issue. If the PC randomly reboots when under this load, there is an issue. Those programs cant exactly be coded badly...they are designed to put an immense load on the CPU, and thats what they do. While i see what your saying...Prime and IBT are very good indicators, and will let you know fairly quickly if something is up...and not when your a few hours into a render or video encoding or huge RTS game.
I totally trust IBT to tell me my PC is stable. Just because your software doesnt hang or crash or reboot...doesnt mean its stable at all.
Operaghost
[H]ard|Gawd
- Joined
- Jun 4, 2004
- Messages
- 1,315
QFT
Honestly, who cares if a calculation stress test crashes/hangs your machine?
If you can't crash it doing anything you use your computer for, then wouldn't you consider it "stable"?
I mean, unless you're folding or something of the sort, it really is completely irrelevant. You can call my machine unstable all you want, but until it crashes from me using it AS MY COMPUTER, I'll call you a liar.
OMGZ you can't run P95 for 72 hours straight without a hang!? Thats totally WTF unstablz!
Gimee a fuckin break and get real.
Real life usage is all that matters unless you built your PC just to run stress tests 24/7.
Intel Burn Test is based on LINPACK which is developed by Intel, so I really don't think it's flawed or badly coded. If your machine can do five passes of IBT at stock speeds but fails on overclock, the problem is not IBT, it's your overclock.
It's really up to you as to when you want to call a system stable. I'm personally not comfortable calling a system stable unless it can do five passes of IBT, two hours of OCCT, and at least eight hours of Prime95.
10 IBT's, 4hrs OCCT large data set, and 24hrs of Prime95 here.
In all seriousness, no one wants to hear about your system issues or your all mighty overclock if you haven't tested for stability - save for running everything at stock speeds.
It's an insurance card for your system that what you've done to it is safe enough for everything to work properly together. Its also a formality to others on these here overclocking sites when you ask for help, you've at least tried to test for stability.
And the kicker is that video encoding and file compression don't go back and check their results. That's why IBT, P95, ORTHOS etc... actually FIND errors. There's more to instability than BSODs or CTDs. Calculating 2+2=4.1 and not detecting the error is a problem too.
I wonder how many people with technically unstable OCs find glitches in their video, or pops and clicks in their MP3s, and blame the encoders....
Operaghost
[H]ard|Gawd
- Joined
- Jun 4, 2004
- Messages
- 1,315
I don't believe this thread was created to gloat about his overclock or complain about his stability problems. Lets quote the OP:
So no, stability really doesn't seem to be an issue, unless there is something NOTICEABLE.
Yes there might be an error when encoding etc, but unless it gets noticed, does it matter?
Not in my mind.
This has turned into a battle between the uber overclockers, and the people who just OC to get some performance increases. The latter, really doesn't give a fuck whether thier system is "stable" by the former's standards or not. We just want our shit to work and if it is, without any noticeable problems, then thats good enough.
Considering IBT is just a wrapper for Linpack which was written by Intel to test their CPUs, it's quite unlikely that it's a flawed, badly coded program. I agree with the rest of your post though, in the sense that it's necessary to run a multitude of tests to actually determine true stability.
No, I wouldn't consider it to be stable, because if it fails even a single test, then it quite simply isn't stable. That isn't an opinion, it's a matter of fact. Whether or not an unstable machine is stable enough for you is a personal choice, but stability itself is not debatable. Either it is, or it isn't.
That's the thing though. There are problems, and apps like P95 and IBT are the things that tell you about them. If you don't stress-test, you won't see the problems, but that doesn't mean they don't exist. And for all you know, your system could actually be affected, you just haven't noticed it. For example, you could be doing some calculations in an Excel spreadsheet one day, and one of them will return wrong, and you'd never realize it.
heh that was kinda my point - 
so far I have seen zero impact on not being able to pass ibi (and I want tm on that calling it IBBY )
two good points though
"if it doesn't affect anything you do why do you care"
"it might affect everything you do and you don't notice it - yet"
ima tweak it a bit for my money anything over 3.2 means I got a 1000 dollar cpu for 229
so far I have seen zero impact on not being able to pass ibi (and I want tm on that calling it IBBY
)two good points though
"if it doesn't affect anything you do why do you care"
"it might affect everything you do and you don't notice it - yet"
ima tweak it a bit for my money anything over 3.2 means I got a 1000 dollar cpu for 229
- Joined
- Feb 3, 2004
- Messages
- 25,321
I agree with this 100%
I run my Q9650 at 4.0....3 weeks straight not a single problem, I ran about 20 3DM06 runs testing out different stuff on my video card, I can run FAH on the CPU and video card 24/7, play any game I want, never a problem..
I ran IBT and it crashed in 20 minutes.....I ran my volts up to 1.375 and it would run IBT for hours....but who cares...I put it back down to 1.300 and left it there......am I supposed to keep the extra volts, heat and the stress that goes with it when I obviously don't need it for what I do?..
I think not.
I don't believe the thread was created to gloat about an overclock either. Hence the fact I never claimed that. It was an OR statement created in response to people like you who try to persuade people not to test because it don't matter - when it in fact does.
Personally, I'd be pissed if that save game didn't work because of a simple rounding error.
That said, IBT shouldn't be the end all. On all cpu's it pushes your cpu higher than any other stability testing software. Upwards of 20°C on some. A simple fact is, Prime95 doesn't push my cpu over 60°C but IBT hits 70°C. Do any of my games hit 70°C? Not at all, and many never hit 60. Case and point, if you can prime for a while and occt for a while, your oc is fine.
vanilla_guerilla
Supreme [H]ardness
- Joined
- Oct 3, 2005
- Messages
- 5,895
i'm with this guy. if my system stays up during 4 and 5 hour game sessions, its stable enough. for me. when somebody is paying me to have a pc that can run prime95 or linpack or whatev for 24 hours, then i will worry about it.
Operaghost
[H]ard|Gawd
- Joined
- Jun 4, 2004
- Messages
- 1,315
You all keep making these scenarious where you'd hate if this happened or you don't want that to happen, but the bottom line is, until it DOES happen, there isn't anything to worry about.
We're primarily talking about everyday people who mainly game/surf/watch movies/rip cds/dvds. Not people working on government missile coding or some crazy NSA anti-hacker software. Thus, you really don't need proactively determine that absolutely NOTHING is causing any errors. Again, this is all personal preference, which is what the OP was asking. He wasn't asking if his system was "omfg without a doubt 100% flawless". He simply asked if he should be worried/concerned. In most normal PC enthusiasts situations, no he really shouldn't be. YMMV
lens pirate
Weaksauce
- Joined
- Jan 1, 2009
- Messages
- 107
Outstanding! You made a meaningless small reduction in O/C based on testing. Perfect. That is why we test. Now you have a great O/C.
.5ghz over Intel's fastest 1000 dollar part and you KNOW it is stable.
Why any one objects to this kind of validation much less engages in pointless arguments about why its not a good idea is beyond me.
Cost of O/C = Nothing
Cost of Validation = Nothing
Safe stable machine that wont crash or lose data = Priceless.
.5ghz over Intel's fastest 1000 dollar part and you KNOW it is stable.
Why any one objects to this kind of validation much less engages in pointless arguments about why its not a good idea is beyond me.
Cost of O/C = Nothing
Cost of Validation = Nothing
Safe stable machine that wont crash or lose data = Priceless.
nissanztt90
2[H]4U
- Joined
- Apr 13, 2007
- Messages
- 3,858
Why would someone object?
Id say they would object because they cant get there semi-uber-mediocre OC on there whatever CPU stable with P95 for more than an hour (much less IBT), BUT it doesnt crash or BSOD or restart while listening to music and surfing the web, but most importantly it doesnt crash while playing CSS...so they come on here and preach about how true stability testing means nothing and its stupid.
vanilla_guerilla
Supreme [H]ardness
- Joined
- Oct 3, 2005
- Messages
- 5,895
on that note, why does it bother you that my definition of stability is different than yours?
nissanztt90
2[H]4U
- Joined
- Apr 13, 2007
- Messages
- 3,858
Probably because if your definition of stable isnt 15-20 runs IBT stable...than your definition of stable isnt very good?
Thats like defining an automobile as well built cause its nice and smooth at 60mph...but it will shake and shimmy doing 80mph+ on a high speed interstate.
Does it bother me? No. But dont complain when you crash. Pun intended.
lens pirate
Weaksauce
- Joined
- Jan 1, 2009
- Messages
- 107
Because your definition as a Anti-tester, is not based on fact or actual results but on assumptions and emotion.
That's fine for you and your PC. But to come here and tell folks that "gee it seems like it works fine, so it must be OK", is an adequate measure, when simple, free, easy to do tests can actually ensure stability and data integrity, is little more than self aggrandizing INTERNET babble.
At worst it could lead someone to loose valuable data or miss opportunities. Straight question, straight answer.
Here is a question for you. Do you think your machine wont pass?
Do you doubt your ability to fix it?
nissanztt90
2[H]4U
- Joined
- Apr 13, 2007
- Messages
- 3,858
Ive had enough of your well thought out logic lens pirate. Im going to stress test right now.
I only listed one scenario this entire thread. Yes, that one was made up at the time, but I was attempting to strike a chord with some people. Failed I see...
Truthfully speaking, instability failures have happened TWICE to me personally, which resulted in corrupt data. Live and learn...
The first time, I lost the entire OS, wouldn't boot. The second time, checkdisk (thankfully) was able to recover 48 corrupt files (all services) after the OS crashed due to "an attempt to write to read-only memory." Both times due to a new overclock that hadn't been tested for stability. Take my experience with a grain of salt if you want, but my experience should prove that at least some testing should be done.
groebuck: What chip are you running that has a 19 multiplier? LOL Nice OC either way, just curious.
Wow, what are you even doing here at [H]?
Anyways if you don't really care about system stability enjoy your various gpu driver corruption / installation failures and random BSODs.
Jesus, another one. And you guys wonder why your nvidia/ati drivers fuck up, crysis bugs happen and gta 4 crashes.
Just FYI I've tested my Q6600 at 3.861 for 60+ IBT runs stable and 32 hrs of P95 stable (before I shut it off). And you know what, I've installed Crysis on my pc, didn't even bother patching it to version 1.1 or 1.2 and played through the whole game without any bugs, glitches or any problems of that sort that multiple users reported about the game. Hmmm...now that all of these statements spilled out I wonder why. Same goes for GTA IV; even during prolonged gaming sessions did I experience crashes. Ofcourse you guys would just bitch and blame the game, "Crysis is a shit coded game waaaaa", "GTA IV is a shit coded game waaahhhh" etc. Funny shouldn't the real world experience tell you something about the instability of your overclocks?!
Stop being a bunch of lazy girly men; completely test out your oc stability then enjoy gaming without crashes / bsods etc.
/rant mode off.
Seems like there's a lot of noise over a matter of semantics: the definition of stability.
I've a simple, hard metric. I consider an overclock stable when the computational reliablity of the system is equivalent to that at default clockspeed. Under this metric, if at default speeds, IBT runs 20 times without issue but fails at run 19 when overclocked, the system is unstable.
To lower the bar is to delude yourself. Sure, the system may be sufficiently issue-free for one's day-to-day purposes but the fact remains, it simply is not as reliable as it was in its stock state.
I've a simple, hard metric. I consider an overclock stable when the computational reliablity of the system is equivalent to that at default clockspeed. Under this metric, if at default speeds, IBT runs 20 times without issue but fails at run 19 when overclocked, the system is unstable.
To lower the bar is to delude yourself. Sure, the system may be sufficiently issue-free for one's day-to-day purposes but the fact remains, it simply is not as reliable as it was in its stock state.
BillParrish
Supreme [H]ardness
- Joined
- Aug 25, 2006
- Messages
- 7,519
/Yawn - Definition of "What is Stable" beat to death in many many other threads - same arguments different day.
Sufficient views of both sides have been given - If you do wish to continue this, do so with the forum rules foremost in your mind.
Sufficient views of both sides have been given - If you do wish to continue this, do so with the forum rules foremost in your mind.
danielkeyes
n00b
- Joined
- Jan 28, 2009
- Messages
- 53
okay discussion, the real point is, it's up to you. If your happy with the stability you do have, leave it. If the stability you have now isn't enough, change it. It's your machine, whatever works for you.
Operaghost
[H]ard|Gawd
- Joined
- Jun 4, 2004
- Messages
- 1,315
Precisely DK!
heh wow I started a crap storm with a lil question
in the end - for me - 3.8 is just as good as 4.0 with the piece of mind that It passes prime and intel burn in. I don't really notice a difference between the two in day to day use and if I have an effeup mp4 or mp3 I know it's because I am a moron and screwed something up and not because I am a moron and my overclock screwed someting up
in the end - for me - 3.8 is just as good as 4.0 with the piece of mind that It passes prime and intel burn in. I don't really notice a difference between the two in day to day use and if I have an effeup mp4 or mp3 I know it's because I am a moron and screwed something up and not because I am a moron and my overclock screwed someting up
vanilla_guerilla
Supreme [H]ardness
- Joined
- Oct 3, 2005
- Messages
- 5,895
Not exactly. That's all well and good as long as you're not posting on a public forum trying to mislead people into believing that your PC is completely stable when there are tools available that are very easy to use that will show otherwise.
To make a statement that "my PC is stable enough for what I use it for", is one thing.
To make a statement that "my PC is completely stable", is another.
Sounds similar but entirely different meanings. The fact is some people are more concerned with bragging rights than they are telling the truth.
BababooeyHTJ
Supreme [H]ardness
- Joined
- Jan 21, 2009
- Messages
- 6,951
I got my q9650 prime stable for 8 hours where I stopped the test and in one minute of IBT I BSODed. IBT draws way more an if you don't have loadline or some other vdroop mod enabled/installed your VCC will dip lower than in prime causing instability.
vanilla_guerilla
Supreme [H]ardness
- Joined
- Oct 3, 2005
- Messages
- 5,895
why would you care if someone types that their pc is completely stable, but it's not really stable?