• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

i7 and Intel Burn Test

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

conk2003

New Member
Joined
Apr 21, 2009
Hello. I am trying to overclock my i7 920 to 4ghz but before I started to OC I used Intel Burn Test with everything at stock. I can pass the test when it is set at standard but when i use a higher level it fails on the second test. Does the test when set to high or very high use more RAM or does it just stress the CPU more?

Has anyone any ideas as to why it wont pass even at stock? The system runs fine and at good temperatures. I have a Corsair TX850 so should have enough power. Is the Intel test compatible with the i7's? I have run Prime95 for a few hours and had no issues. Is there any other programs that can tell you if your OC is stable in 10 minutes before finally doing 24 hours on Prime95?

I have searched for a few days for an answer but couldnt find one anywhere. Anyway any help would be much appreciated. Thanks.
 
Last edited:
Thanks for your help, but doesnt the i7 have 4 physical and 4 logical cores? So shouldnt I set it at 8 threads?
 
IntelBurnTest uses Linpack doesn't it? I heard it gets the results P95 gets in 24 hours in less than 1 hour.

When I use IBT i do 15 tests on the maximum stress.
 
Thanks for your help, but doesnt the i7 have 4 physical and 4 logical cores? So shouldnt I set it at 8 threads?

If you enable HT, you'll get 8 cores. As most of the time we won't be utilizing HT, it's good to turn it off.

IntelBurnTest uses Linpack doesn't it? I heard it gets the results P95 gets in 24 hours in less than 1 hour.

When I use IBT i do 15 tests on the maximum stress.

Prime95 is still a good old school method to test stability. I always go for that instead of whatever Linpack or burn tests.
 
If you enable HT, you'll get 8 cores. As most of the time we won't be utilizing HT, it's good to turn it off.



Prime95 is still a good old school method to test stability. I always go for that instead of whatever Linpack or burn tests.

+1 LinX is completely unreliable for me....Prime has never let me down. I have never had an unstable OC after testing with Prime...LinX lies to me:eek:
 
Yes it use's Linpack.

Personally I think Prime lies to me, well not lies but being able to run it 24/7 has never been a necessity here I guess, I've ran computers for years that have been very stable for years gaming etc and weren't fully stable in all the prime tests, and appears to be with the Intel one.

Still have to push it on a newer game I guess, have something interesting it look came from people I used to play with. I used to play 18 hours at a time occasionally in games on a computer that wouldn't run for over an hour on some prime tests that were loaded up on graphics intensive.

I'm getting a bit long in the tooth to do that these days, but I'm feeling in a gaming mood soon. I hadn't been serious about one in about 6 months when Warhammer Online looked like it was fudging out and no one could do city raids because the server always ate it.

Depends on what your doing I guess.

:beer:
 
Last edited:
If you notice carefully, the situation whereby your computer uses 100% utilization and produces so much of heat on normal daily usage is almost minimal. When you play games, it doesn't even run full utilization all the time. What Prime95 does is stressing your CPU/memory to it's limit and that itself clears the possibility of instability if in the event your daily usage hits a full 100%.

Sometimes you can fail Prime and still enjoy games (like Warcraft III) the whole day :) It really depends what you're doing.
 
If you don't pass Prime and play games like Crysis then you'll get some nice blue screens or random reboots I'm sure :D
 
I'd have to try it out but I'm sure I could :)

Never tried Crysis out I guess.

I've been gaming since before the first EQ beta about 12 years ago or so, I forget, was in that and played in Mechwarrior ladders before the web got popular.

And many since.

:beer:

Gonna give this a go soon I think.


http://www.aiononline.com/us/

I am always admitedly trying to get my prime95 stable as possible and always work on that one also, tweaking that in is the final say in stability.

Is what I've always used, but the others are pretty reliable to get in the ballpark, and usually stable enough for most things.
 
Last edited:
Yes it use's Linpack.

Personally I think Prime lies to me I've ran computers for years that have been very stable for years gaming etc and were'nt fully stable in all the prime tests, and appears to be with the Intel one.

Still have to push it on a newer game I guess, have something interesting it look came from people I used to play with. I used to play 18 hours at a time occasionally in games on a computer that wouldn't run for over an hour on some prime tests that were loaded up on graphics intensive.

I'm getting a bit long in the tooth to do that these days, but I'm feeling in a gaming mood soon. I hadn't been serious about one in about 6 months when Warhammer Online looked like it was fudging out and no one could do city raids because the server always ate it.

Depends on what your doing I guess.

:beer:

This really shows off my point pretty good. Any failure, from prime or LinX is a 100% gurantee that your system is not stable. What I am referring to is when these programs find no errors and you are still not entirely stable. LinX can run forever and not error but Prime will catch the problems. A bad OC is a bad OC and will bite you when you least expect it.
 
I've always been from the school of thought where no single test is the end-all-be-all test. For me it takes a battery of tests and lots of real-world testing (just using the PC as I normally would over time) before I will definitively call it 24/7 stable.

I've had instances where the system would pass small-fft for 20+ hrs, but then BSOD while just surfing the web.

I run all my tests. If I pass them, then I start using the PC normally. If I start running into problems w/ normal usage I make small adjustments until I find the culprit.



If your system won't pass IBT when the CPU is stock I would look to the RAM. Make sure you're giving it it's spec'd voltage. Also try lowering the RAM speed (below stock if you have to) and/or loosening the timings. If you have 6 sticks of RAM, try running just 3. Mobos do a good job of running CPUs when set to stock/auto, but RAM is a different animal. Even if you're just trying to run RAM stock it can be a tricky endeavor.
 
I've always been from the school of thought where no single test is the end-all-be-all test. For me it takes a battery of tests and lots of real-world testing (just using the PC as I normally would over time) before I will definitively call it 24/7 stable.

I've had instances where the system would pass small-fft for 20+ hrs, but then BSOD while just surfing the web.

I run all my tests. If I pass them, then I start using the PC normally. If I start running into problems w/ normal usage I make small adjustments until I find the culprit.



If your system won't pass IBT when the CPU is stock I would look to the RAM. Make sure you're giving it it's spec'd voltage. Also try lowering the RAM speed (below stock if you have to) and/or loosening the timings. If you have 6 sticks of RAM, try running just 3. Mobos do a good job of running CPUs when set to stock/auto, but RAM is a different animal. Even if you're just trying to run RAM stock it can be a tricky endeavor.

Find your OC with OCCT, as it's the quickest method, I've found, just like test 5 in memtest86+ for memory. You can dial it in with that 90% of the time and then come up with full tests afterwards and make adjustments once you get there. The ajdustments required should be small if you're doign 8 hour burns with occt at every interval.
 
Hello and thanks for your help. The RAM is getting 1.64v and the speed is set to auto in BIOS, so I think the maximum speed it goes to at stock is 1066mhz? so it isnt nowhere near the speed at which the RAM is rated at. I am using OCZ Reaper 6GB (3x2GB) DDR3 PC3-14400C8 1800MHz Triple Channel Kit 1800MHz RAM Speed, CAS 8-8-8-26 Timings, 1.6-1.65v VDIMM.

This is the first system I have built myself so I am unsure how to change the RAM timings and what too. Would it be a good idea to run MEMTest to test the memory? If so how long for?

When I have finally sorted the problem and start to OC, what is the quickest and most reliable way to see if it stable in between raising the OC? Do you have to use Prime95 for 24 hours everytime you up the OC or is it better just to go straight to 4ghz and guess the best vcore and just one test of Prime and leave it at that?

Sorry about asking so many questions. Appreciate your time. Thanks.
 
I have loosened the RAM timings to 10-10-10-30, set the QPI to 1.35v and also set the vcore to 1.3v and still cannot pass hvery high or maximum on IBT. I have used MemTest for 3.5 hours and it passed with no errors. Do you think IBT could not be compatable with something im running? Also tried Linx and that sent my PC crazy. Are there any other quick stability tests that are good? Thanks.
 
First off, memtest86+ is worthless with X58. Use it all you want but it just doesn't hit the IMC right.
Second, I am a user of LinX(which is Linpack) and find it the most reliable. I have yet to find something that passes 20 runs at full memory utilization fail in any other test whether OCCT or P95. I have been prime stable and have LinX shutdown my pc. In one system it was because the PSU could not handle the current load of LinX but could handle Prime all day.

You haven't mentioned your temperatures that I noticed. With HT on LinX will output a ton of heat, even at stock settings. Many have seen it reach shutdown temps of 100C internally on the stock cooling. It also draws more current than anything else known.

If stock speeds are failing: memory, shutdown temp, shutdown overvoltage/current protection are the 3 candidates to check first imo.

Try P95. See if that causes the error also. OCCT was also recommended just don't use the Linpack test, use the default.

GL and HF
 
First off, memtest86+ is worthless with X58. Use it all you want but it just doesn't hit the IMC right.
Second, I am a user of LinX(which is Linpack) and find it the most reliable. I have yet to find something that passes 20 runs at full memory utilization fail in any other test whether OCCT or P95. I have been prime stable and have LinX shutdown my pc. In one system it was because the PSU could not handle the current load of LinX but could handle Prime all day.

You haven't mentioned your temperatures that I noticed. With HT on LinX will output a ton of heat, even at stock settings. Many have seen it reach shutdown temps of 100C internally on the stock cooling. It also draws more current than anything else known.

If stock speeds are failing: memory, shutdown temp, shutdown overvoltage/current protection are the 3 candidates to check first imo.

Try P95. See if that causes the error also. OCCT was also recommended just don't use the Linpack test, use the default.

GL and HF

I'll keep the memtest thing in mind when I go i7 if I do.
 
Thanks N3buchan3zzar. I downloaded an older version of Linx 0.6.0. I had 0.61 and was doing all sorts of weird things. I have passed all the tests on the new Linx using all the memory so I think IBT may have not been correct. I will tighten back all the RAM timings and start OC'ing. Do you think 5 passes of Linx at full is enough to be stable in between changing settings then a final 20 passes when I got as far as I can go? Many thanks.
 
Thanks N3buchan3zzar. I downloaded an older version of Linx 0.6.0. I had 0.61 and was doing all sorts of weird things. I have passed all the tests on the new Linx using all the memory so I think IBT may have not been correct. I will tighten back all the RAM timings and start OC'ing. Do you think 5 passes of Linx at full is enough to be stable in between changing settings then a final 20 passes when I got as far as I can go? Many thanks.

That's probably a good strategy, as long as whatever you settle at is tested and proven stable, who cares how you get there? :shrug:

I usually work my way up like I said, small increments with lots of testing at each step. That way I get a better idea of how the chip runs.
 
Back