IntelBurnTest - The new stress-testing program

[tekjunkie]

This is my QX9650 @ 4GHZ. My residue is all over the place and I do get occasional BSODs, but I am pretty happy with my stability.

[Marvin_The_Martian]

Space reserved for results ( when my normal cooling get's sorted out again, not the gheto fix ).

I will say I like this tool allot, but the output reading could have been easier. It took me to look at the actual numbers from Loonym to see what was the diffrence with a stable results ( numbers stay equal ).

[Loser777]

Meh, I don't see the need for such a program... any DOS shell with a GUI will do in reading results.

[Kilyin]

This is my QX9650 @ 4GHZ. My residue is all over the place and I do get occasional BSODs, but I am pretty happy with my stability.

BSODs can cause data corruption, and are a great indicator that you don't have stability. But I guess if it doesn't bother you...

[Der_KHAN]

I used Prime95 on my Q6600 with Small FFTs, [...]

use large FFT. problem solved.

[Cy4n1d3]

This test is hardcore. Under load vcore drops to 1.184v.
Had to raise vcore to 1.23 idle - before I ran this system with 1.184v idle, 1.168v load which was primestable 12 hours. Simply insane ^^

Also noticed that 10 runs with 100 mb ram were stable at old settings, 10 runs with 2 gigs were not. Guess this test is really extreme

[RejZoR]

If you want to check actual stability you have to dedicate as much memory as possible. The less you dedicate it, the less it'll be stressed.

Also now i know my system is stable. 3,3GHz stable after 5 passes. 2GB of memory used.

[jcool]

Well I was a big fan of linpack initially, but went back to prime simply because linpack will run on my Q6600 even when prime fails.
It's much like Furmark vs. ATITool with GFX cards, FUR heats the card up all right but ATITool is the one to detect the errors.

[STEvil]

Technically none of you guys are doing this right.

There are two ways to do this.

The first is to run with the smallest amount of ram possible. This places as much data into L2 as possible keeping the CPU working at as close to maximum as possible (small FFT's).

The second is to run with as much ram used as possible to create as much stress as possible on the memory controller, chipset, and memory.

Testing both ways is preferred.

[Cy4n1d3]

If you want to check actual stability you have to dedicate as much memory as possible. The less you dedicate it, the less it'll be stressed.

Also now i know my system is stable. 3,3GHz stable after 5 passes. 2GB of memory used.

Ye sure.
Used 3.5 out of my 4 gb

[Kougar]

Nice tool AgentGOD! Going to be having some fun with this one I think and putting it through its paces. Best program I've seen for heating up the CPU despite my WC loop, that is for sure.

Small bug report, just nitpicking: It reads my CPU as 3.6GHz when it really is 3.2GHz, it doesn't notice the lowered multiplier.

[RejZoR]

It would be nice if there was option that would automatically dedicate all available memory to the program to test. Right now users have to find out on their own how much they can use. Two options would be ideal. Minimal possible for maximal CPU stress and maximal possible for maximal overal system stress.

[JohnZS]

Good Tool Mr AgentGod,
Just a quick question, what is deemed a "safe" (or thereabouts) maximum temperature for 45nm Yorkfield CPU?'s
IntelBurnTest got one of my cores up to a scorching 67C
All numbers were the same at the end though.
Just worried seeing such a high temperature on Core2. Cores 1 3 and 4 reached 63C (which is still hot!!)
John

[btdvox]

Havent downloaded this yet, Ive found to be Prime too unpredicatable.

Ive had times where stability is rock solid during the day (12 hour passes starting at 6 am-6pm.
And then Ive had other times when i start at 6pm till 6 am (showing that Temp doesnt spike because of night..) And itll cause errors,

But Can we get like a post stating a good way to use this program, I see that we should be doing one with barely any memory and one with all the memory.

BSODS are a bad sign and Abnormal residual numbers mean unstable data but is ok to have a couple scrambled numbers...??

[RejZoR]

@JohnZS
As long as temps are below 70°C you really don't have to worry... Also note that actual games and programs never heat up the CPU as much so worry even less.
You can check that with some graph monitoring tool or by RealTemp's Max Temp values. Just reset them before playing some game or using intensive app.

[RejZoR]

Hm, is it normal for this test to take longer than on lower clocked CPU ? With 3,3GHz it took 120 seconds for each loop. With 3,5GHz it takes 155 seconds for each. End results were consistent between all 5 loops. Is this normal? Oh, i used 2048MB of memory for each run.

My bad, forgot multi on 6x instead 9x after BIOS upgrade...

[Datektrader]

theres mine with 6 runs, its a little scary though, the CPU hit 88c

[Leeghoofd]

Hm, is it normal for this test to take longer than on lower clocked CPU ? With 3,3GHz it took 120 seconds for each loop. With 3,5GHz it takes 155 seconds for each. End results were consistent between all 5 loops. Is this normal? Oh, i used 2048MB of memory for each run.

With all other values the same more cpu speed should give a faster run

Here are my results first at 3.2 then 4ghz with the quad : so 400FSB X 8 , then 10X... ram and co at 1066Mhz 5-5-5-5, 2048MB tested on WinXP64 Pro

[majestik]

I love it. Well... actually I love LINPACK, whatever way you run it.
It's fast, brutal and cosistent - I don't need any other stress-testing tool, with this thing I can fine-tune the FSB within 1mhz

[RealTelstar]

The data is easy to look at, because you can scroll through it extremely fast, and if any line is different, you will notice very easily because it sticks out like a sore thumb. But still, it would be nice to have it report if there is an error.

Also, I believe that even if they all are the same, but the number is simply too large, then that also means it's not stable. I'm not sure what the cutoff for stable vs not stable is though.

I still dont get how you see if the test failed. Can someone post an example screenshot of succeed/fail, highlighting the error line?
I have never used linpak...

[panheimer]

I still dont get how you see if the test failed. Can someone post an example screenshot of succeed/fail, highlighting the error line?
I have never used linpak...

At page 1 o this thread loonym posted a screen of failed test. If you look on the last column on the right (residual norm), the values were incoherent each other: this is a failed test.

But no one else but me experienced weird system issues runnign this test?

[Leeghoofd]

I still dont get how you see if the test failed. Can someone post an example screenshot of succeed/fail, highlighting the error line?
I have never used linpak...

Post 23 clarifies clearly mate... all values should match... any discrepancy is not good... though hard to tell if it's ram or cpu that craps out... think iwll use this as a last test

One question how much rma do you guys put in ? I took the free ram of physical Memory available in the windows Task Manager... about 3Gb here...

What do you guys set ?

[krille]

Wow, AgentGOD, didn't know you were an XS'er! Need to OC that quad much to debug BW?

[panheimer]

Well, i've got another rampage and a q6600 on my 2nd rig, so i tried this tool to see if it gives me strange behaviour too. So i've run a 10 cycle test and it runs very smooth without an issue!
Now i'm sure it's not a software fault, but something in my first rig that doesn't work. But the fact that even with stock settings i've got this strange behaviour worries me a bit...don't want to rma this board!
I'll track down every component of this rig to search for the guilty one...any suggestion?

[RejZoR]

@Leeghoofd

I screwed it up there. I forgot to set multi from 6 to 9 when updating BIOS. Thats why it took longer and it appeard to be more stable. But it was in fact running at 2,3GHz...
My error...