Stress Test

[Pablo 54]

Using an Asus program but it gives me no data after a crash. Anyone have a good freebie they could recommend?
I have always just used what I had on hand and went with that but I need something a bit more robust.

[PP Mguire]

Overclock one piece at a time, stress test it, and if it crashes you know what the cause was.

[CDsDontBurn]

Overclock one piece at a time, stress test it, and if it crashes you know what the cause was.

How can you overclock a motherboard? :?:

[PP Mguire]

Eh, who said anything about a motherboard? I figure he meant GPU and CPU so just do one piece at a time so if your PC crashes you know which part it is.

But, if we want to be technical by raising the bclk on SB and IB you are overclocking the motherboard lol.

[CDsDontBurn]

Nobody said anything about the motherboard. But at the same time, nobody said anything about the motherboard :p. When you mentioned "Overclock one piece at a time, stress test it" I immediately thought of motherboard, hahaha.

And, I'm outta the loop on benching and stuff. What is the "bclk"?

[TehYoyo]

Nobody said anything about the motherboard. But at the same time, nobody said anything about the motherboard :p. When you mentioned "Overclock one piece at a time, stress test it" I immediately thought of motherboard, hahaha.

And, I'm outta the loop on benching and stuff. What is the "bclk"?

I'm guessing.... bridge clock? That would make the most sense to me, although it's not really adjusted in the processor...

[PP Mguire]

BCLK is base clock. Sandy and Ivy systems use a 100MHz BCLK and if you adjust it you are messing with your PCI clock and a few other things as well. That is why they have K CPUs with unlocked multipliers so you can overclock them, otherwise everybody would be stock. SB-E happens to have a BCLK ratio changer that some how manages to raise only the BCLK on the CPU to OC to an extent (say 3820) but you are very limited.

By one piece at a time I meant stock everything out, then set an overclock and test one piece of hardware at a time that you want to overclock. It's best that way so you aren't guessing what is unstable.

[werty316]

The question asked is too vague.

[khangdan1992]

So if u wanna know your crash issues, use some programs such as IBT ore OCCT, it mentions why it crashed

[CDsDontBurn]

OCCT explanation is very vague and doesn't realyy explain the issue. IMO, Prime95 does a better job on that.

[PP Mguire]

IMO the simplest answer has already been given. One piece at a time until stable, then move to another piece.

[khangdan1992]

OCCT explanation is very vague and doesn't realyy explain the issue. IMO, Prime95 does a better job on that.

yeah, but i like the diagram of OCCT, it's very scientific.

[CDsDontBurn]

yeah, but i like the diagram of OCCT, it's very scientific.

It also looks Communist Russian :p

[khangdan1992]

It also looks Communist Russian :p

nice point, but actually i don't wanna talk about Politics in a technology forums. I'm tired to hear about that everyday......:roll:

[CDsDontBurn]

nice point, but actually i don't wanna talk about Politics in a technology forums. I'm tired to hear about that everyday......:roll:

it was joke, not a political comment....

[TehYoyo]

Using an Asus program but it gives me no data after a crash. Anyone have a good freebie they could recommend?
I have always just used what I had on hand and went with that but I need something a bit more robust.

Do you desire a program? Because Prime95 is pretty much the defacto standard. (Link (http://files.extremeoverclocking.com/file.php?f=205))

Dunno if that's been posted already.

[CDsDontBurn]

Do you desire a program? Because Prime95 is pretty much the defacto standard. (Link (http://files.extremeoverclocking.com/file.php?f=205)

Dunno if that's been posted already.

didn't know there was a 64bit version of Prime95! :-D

[TehYoyo]

didn't know there was a 64bit version of Prime95! :-D

Hay hay! Learn something new everyday!

It makes sense I guess... most people OC'ing by now probably have 64-bit systems... although 64-bit for something as primitive as Prime95.... eh.

[That Guy]

Hay hay! Learn something new everyday!

It makes sense I guess... most people OC'ing by now probably have 64-bit systems... although 64-bit for something as primitive as Prime95.... eh.

Quoted b/c it rhymes :D

Personally, I'm using Prime95 32-bit (or at least, whatever was around a few years ago). I don't think it really matters. I'm not really benchmarking in Prime95.

[PP Mguire]

It doesn't matter. The point of the program for 99% of the people using it is stability. A 32bit program can address all the cores in todays modern systems when built specifically to address a load to each thread.

[CDsDontBurn]

It doesn't matter. The point of the program for 99% of the people using it is stability. A 32bit program can address all the cores in todays modern systems when built specifically to address a load to each thread.

But would a 32bit burn-in program be able to address 64bit extensions of a CPU even if the program is coded to use all the cores? Perhaps a system would be able to run perfectly in 32bit environment but there might be something wrong in one of the cores' 64bit extensions that the user would not know about.

[PP Mguire]

Most stability problems come from heat or lack of voltage. If it is stable in 32bit with those set right, usually the program will run perfectly fine in x64. I have never had specific crashing personally.
Considering 64bit programs usually are only made that way to utilize natively memory addresses beyond 32bit I don't see where it would make a difference.

[ryandigweed]

Is the OP only looking for memory based testing?

for graphic card stress testing, there is also "FurMark" which is a burn in test for graphic cards :)

[CDsDontBurn]

I find FurMark to be a controversial. Not sure if it's because of everything I've read or if it's because of how I do testing.

[ryandigweed]

I find FurMark to be a controversial. Not sure if it's because of everything I've read or if it's because of how I do testing.



What have you read about it?

[PP Mguire]

Furmark puts more of a load on graphics cards than a normal load. It can sometimes cause weak cards to go poof. The same can be said on some of these CPU stress programs though.

[CDsDontBurn]

Furmark puts more of a load on graphics cards than a normal load. It can sometimes cause weak cards to go poof. The same can be said on some of these CPU stress programs though.

And that's where I'm torn on this. I like to use FurMark because I want to ensure that the card is properly working and if it fails FurMark, then it won't continue any additional QA testing. But I've also read that FurMark will cause a good card's life span to be shortened as the stress on components caused by FurMark take their toll on the components and effectively reduces the card's lifespan. So, IDK. I'm personally torn on using FurMark because of this kind of stuff.

[PP Mguire]

For the card to have any damage to begin with it has to be defective or damaged really. The amount it may reduce lifespan is the same as regular gaming. These tests like Furmark are not meant to be left running in the least.

[ryandigweed]

Yeah any burn in test will naturally reduce the life of your hardware. Its like leaving your system on for a long period. Only that it pushes it to the limit within a small time :-)

[Pablo 54]

Well, thanks for the suggestions. I will use Prime95 64bit when I can wrestle the computer back from my wife, who does not think it is a problem to download every "free" coupon link on the d**n Internets. :givebeer: I now have four toolbars and some shopping popups. Once she is done, I am going to do a clean re-install and start over. On my F@H box, I found out that though the q9550 that I am using can take a lot of heat it can't run for more than 4 hours with even the slightest OC. The bump in PPD is not worth buying a water block etc.

[CDsDontBurn]

For the card to have any damage to begin with it has to be defective or damaged really. The amount it may reduce lifespan is the same as regular gaming. These tests like Furmark are not meant to be left running in the least.Yeah any burn in test will naturally reduce the life of your hardware. Its like leaving your system on for a long period. Only that it pushes it to the limit within a small time :-)

I fully understand what you guys mean. For me, it's more of a thing about using these tests here at work and ensuring a product is working well or possibly making it DOA when it arrives to the customer. That's what it comes down to for me, really.

[TehYoyo]

Well, thanks for the suggestions. I will use Prime95 64bit when I can wrestle the computer back from my wife, who does not think it is a problem to download every "free" coupon link on the d**n Internets. :givebeer: I now have four toolbars and some shopping popups. Once she is done, I am going to do a clean re-install and start over. On my F@H box, I found out that though the q9550 that I am using can take a lot of heat it can't run for more than 4 hours with even the slightest OC. The bump in PPD is not worth buying a water block etc.

Oooooh yeah. I love me some reinstalls.


I fully understand what you guys mean. For me, it's more of a thing about using these tests here at work and ensuring a product is working well or possibly making it DOA when it arrives to the customer. That's what it comes down to for me, really.

Where do you work?

[PP Mguire]

Well, thanks for the suggestions. I will use Prime95 64bit when I can wrestle the computer back from my wife, who does not think it is a problem to download every "free" coupon link on the d**n Internets. :givebeer: I now have four toolbars and some shopping popups. Once she is done, I am going to do a clean re-install and start over. On my F@H box, I found out that though the q9550 that I am using can take a lot of heat it can't run for more than 4 hours with even the slightest OC. The bump in PPD is not worth buying a water block etc.Lol looks like somebody needs a guest account. I wouldn't tolerate that crap on my PC as I don't use an AV. (Well I do as of right now, but under different circumstances)


I fully understand what you guys mean. For me, it's more of a thing about using these tests here at work and ensuring a product is working well or possibly making it DOA when it arrives to the customer. That's what it comes down to for me, really.Yea I can understand that, but then again doing this on a bone stock machine with good airflow wont damage the components. And like I said, if it does, it was probably damaged anyhow. Don't want to give a customer a bad machine. As for a personal rig, I think if you need an OC that stable for something that important you might want to invest in something better. I said that before, but just my personal opinion on that.

[Goliath182]

If you would normally have the component off then it will last longer since the components have a limited lifetime, but if you have the computer sitting there idling you aren't saving it from anything. As PP said if the temperatures are in check and nothing is getting higher voltages than it should then it has no to very little impact on the components life span. However I wouldn't personally run components under stress for a long time nor advise someone to do so.

[CDsDontBurn]

Where do you work?

I work at a company that tests products :-P

Yes, I know I'm being vague, but I don't want everybody knowing where I work at. Some people from here who are on my FB already know where I work at. And in another post in the "post yourself" thread, for about 30 minutes or so, there was a picture of me showing where I work. I know at least one person on here saw that picture.


If you would normally have the component off then it will last longer since the components have a limited lifetime, but if you have the computer sitting there idling you aren't saving it from anything. As PP said if the temperatures are in check and nothing is getting higher voltages than it should then it has no to very little impact on the components life span. However I wouldn't personally run components under stress for a long time nor advise someone to do so.

I totally agree with you and PP on this, but like I said, it's just a matter of testing something to make sure it works the way it should and not over-testing something to the point where it damages components and make the product DOA to the customer.

[najiro]

I know at least one person on here saw that picture.

I saw two monstrous HD 7990 in your hands!! hahahaha
well I don't really have any idea where would that be.. all I know is you're surrounded with s3xy PC hardware
LOL

[CDsDontBurn]

I saw two monstrous HD 7990 in your hands!! hahahaha
well I don't really have any idea where would that be.. all I know is you're surrounded with s3xy PC hardware
LOL

Yes. Yes I was.