• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Best stress test utilities?

Gerard

Gerard

Gerard

Caporegime
Joined
18 Oct 2002
Posts
40,804
Location
Ireland
Thought my 6700k was stable @ 4.5 but last night got a "whea uncorrectable error" which is apparently linked to cpu voltage. That's despite it passing several hours of aida 64 a few weeks earlier.


Seen some sites say that the latest versions of prime 95 are more of a power virus and aren't worth using, is intel xtu rated as being good for stability testing?
 
You might be better off asking this in overclocking and cooling.

My answer would be, it depends on your tolerance of errors.

I still stand by linpack (Intel Burn Test or Linx), because it's the most intense and sensitive stress test I've ever used, and because I want to guarantee as few errors as possible. Note that it's not obvious how to use linpack correctly so most people don't and then blame it when their machine passes but later crashes.

On the other end of the scale some people say that if you're gaming then games are the best stress test, but they don't do any explicit error detection (some errors may be silent).

In the middle are programs like Prime95, and 'blends' like Aida64 and OCCT.
 
Last edited:
RavenXXX2 said:
Real bench, a proper real world test.
Click to expand...
'Quoth the Raven', Realbench seems to be the one recommended by the guys that do this for a living. I personally just game and encode at the same time for a ~30mins or so; which has been a bang on indicator for me.
 
Intel have their own software that stress tests, other than that most of them do the job any way. I find some to be a bit over the top though
 
JamesM said:
Just crashes I think.
Click to expand...
No it doesn't just crash. ;) It gives you an error readout. Realbench really seems to be the best thing for people to use. It doesn't put unnessary extra heat into the CPU than has more potential to damage.
 
Last edited:
I do a run of real bench, then just use it, if it crashes add a bit more vcore. Although I have dropped to 4.6ghz now and reduced vcore a bit, as I didn't see the point in the extra V core for 100mhz, although it hasnt really made any real difference to my temps (low 60s)
 
Gerard said:
Thought my 6700k was stable @ 4.5 but last night got a "whea uncorrectable error" which is apparently linked to cpu voltage. That's despite it passing several hours of aida 64 a few weeks earlier.


Seen some sites say that the latest versions of prime 95 are more of a power virus and aren't worth using, is intel xtu rated as being good for stability testing?
Click to expand...

p95 has always been the best stress tester available and depending on settings can basically test cpu and memory+imc seperately, ibt and occt avx linpack can stress the cpu one hell of a lot too. all will hit max tdp. all will be a good +10c hotter than anything else. ibt and occt may get the cpu hottest.

realbench, well, its a "real world" test that doesnt stress the system very hard. the cpu wont be hitting max tdp.

imo aida, xtu and realbench are abit of a joke
 
zipp0r said:
p95 has always been the best stress tester available and depending on settings can basically test cpu and memory+imc seperately, ibt and occt avx linpack can stress the cpu one hell of a lot too. all will hit max tdp. all will be a good +10c hotter than anything else. ibt and occt may get the cpu hottest.

realbench, well, its a "real world" test that doesnt stress the system very hard. the cpu wont be hitting max tdp.

imo aida, xtu and realbench are abit of a joke
Click to expand...

Why not just remove the cooler, that will make it hotter still...... :rolleyes:



To the OP, if you class yourself as an average pc user, gaming, videos etc then just use Realbench as stated, couple more clicks of vcore may be required to be stable enough for everything, really no need to murder your cpu using the likes of xtu/ibt etc
 
zipp0r said:
p95 has always been the best stress tester available and depending on settings can basically test cpu and memory+imc seperately, ibt and occt avx linpack can stress the cpu one hell of a lot too. all will hit max tdp. all will be a good +10c hotter than anything else. ibt and occt may get the cpu hottest.

realbench, well, its a "real world" test that doesnt stress the system very hard. the cpu wont be hitting max tdp.

imo aida, xtu and realbench are abit of a joke
Click to expand...

A stress test for overclocking is about working out if the system is stable, not about making it as hot as possible.

Further more, if you use an offset voltage then stress tests that use AVX code (like P95, IBT, etc) use higher voltage than non AVX code does. This is normal because the CPU sets a higher SVID when it's asked to do an AVX workload. However while your CPU might be stable in P95 with the higher voltage, once you run a game which will come with a lower voltage you can run into stability issues. This is the number 1 cause of issues in systems that have been found 'P95/IBT stable' then quickly bomb out once games are played.

P95 or IBT are great stress testers if your workload is going to be heavy AVX code, but if you want your system to be stable in games then they are next to useless. Most people won't even be stress testing at the same voltage.
 
I agree with the above. I can stress test my 4770k at 4.7 with p95 and Asus realbench and it passes both but 30 minutes into gta v it will crash. Knock it back to 4.6 and it will run gta v all night
 
All stress test utilities should be used as a rough guide ONLY, use the ones that suits your needs and then use your system normally. If it crashes make notes of codes etc to determine if its cpu/vcore related or memory, there is lots of info around to work out the codes etc or just ask!
Also, just because you use XMP, never think it cant be memory related, it can! As this is still overclocking.
 
You must log in or register to reply here.
Back
Top Bottom