I think our "frog" is missing the whole point. Let me set him straight.
First of all, you're not a moderator, so you don't get to set rules or guidelines, and you certainly don't get to moderate members of the moderation team. If you are displeased with that, you can feel free to contact a member of the forum community staff. Otherwise, you can sit back in your chair and act as though you belong with the rest of the group, or leave.
Two, Prime95 version 26.6 is one of VERY few programs out there (Or programs that use elements of Prime95 WITHOUT Linpack, x264 encoding or some form of AVX instruction set) that uses a 100% steady state workload, so unlike those OTHER programs out there you were talking about, you know, the ones that use Linpack and have cycles that peak at 110% workload, which aren’t suitable for CPU thermal testing, Prime95 is the Gold standard used by anybody who has the faintest clue about the reasons why you test in the first place, which we'll get to in a minute. The test utility OCCT runs elements of Linpack and Prime95, but will terminate the CPU tests at 85C.
I'm not sure what other utility you THINK is out there (Aside from those listed below that are OK for thermal testing, but are not adequate stability tests) that presents a 100% steady state workload, but I am unaware of it as are the three individuals I know of who have probably spent more time testing and writing about CPU architectures than anybody else on the planet that isn't an engineer with either AMD or Intel stamped on their paychecks.
Prime95 is a 100% workload. Not 110%. Not 115%. 100%. That means it will make full use of the core resources as outlined by the hardware configuration. If it cannot do so without causing an error, there is instability. If it cannot do so without exceeding the thermal standard, there is too much voltage. If you cannot achieve stability with an amount of voltage that remains within the thermal guideline, then the silicon is incapable of the parameters you are trying to enforce (Or you have some other issue affecting stability such as out of spec ripple/noise that is affecting the hardware, or other hardware problems such as unstable memory configuration) and you should reduce the expectation of the overclock you are trying to achieve. Simple as that.
Running a program that does not induce 100% workload, simply to avoid the possibility of an unwanted thermal situation, is merely giving yourself a pat on the back for nothing. You've proven NOTHING. You've protected yourself from NOTHING. AND, more importantly, if you have NOT run Prime95 for a period of 24 hours then you have ABSOLUTELY not ensured that there is not silent data corruption happening WHICH IS THE WHOLE REASON FOR CHECKING STABILITY IN THE FIRST PLACE.
The point of stability testing is not, primarily, to avoid obvious errors like crashes and blue screens. It is to eliminate the silent data corruption that happens when a processor or other hardware is overclocked beyond the configuration the manufacturer already assured stability at for the given clock/voltage configuration. It matters not AT ALL whether you only game on this system, or if it is used by NASA for scientific calculations. In ANY and ALL cases, silent data corruption WILL occur and will cause micro-errors and bit flips to accumulate in the operating system and any game related files that have been written or re-written during normal operation, updates and saves, or during the normal course of the operating system's duties, and will eventually cause the very crashes, blue screens and miscellaneous errors that you thought you had tested for with those other utilities in order to avoid.
FURTHER, no CPU has ever been damaged by running Prime95, because the system would either throttle if it exceeded TJmax or it would shut down if throttling was incapable of reducing the thermals back to the point of being within tolerance.
You cannot point to ANY instance of Prime95 EVER causing damage to ANY system. OVERCLOCKING might have caused damage to their systems, but running Prime95 did not. IF thermal damage could be caused by running a utility, then 90% of the quick testing utilities out there would be 100% more likely to do so than Prime95 since those are mainly Linpack or AVX based utilities that can exceed 100% workload unlike Prime95 OR as with some utililities such as AIDA64, fail to come anywhere close to 100% workload or full TDP, and are worthless except as a means of giving you a false sense of security.
The argument that if you only game it matters less whether your CPU is fully stable is the least accurate of all the statements I've read on this page for exactly the reasons I've outlined above. So that argument needs to get shelved next to the old magazines that used to say that smoking improved lung function.
Another often overlooked but incontrovertible fact is that the 24hrs of Prime95 Small FFT (NOT blend, NOT Large FFT) testing is not MERELY to ascertain whether the system is MOSTLY stable (No such thing as 100% stable without running the test indefinitely and being around forever to determine that it never errored out, but for our purposes 99.999% stable will be good enough for gamers and workstations running CAD or scientific software equally) but is intended that enough time be given for all the relevant FFT LENGTHS to have been run otherwise running it for only a short time or a couple of hours does not allow it to run through all the FFT lengths which is the REAL reason it was originally recommended to run it for 24 hrs. Without testing all FFT lengths or at the least, most of them, you've no idea if some random calculation or instruction will create a catastrophic or silent error, whatsoever.
About 20 hrs is generally long enough to have tested all relevant FFT lengths according to every reliable and exceptionally well informed source I've spoken with at length about the subject, but I have personally seen systems that errored out/crashed workers between 20 and 24hrs, so I, like pretty much every other accomplished overclocker on the planet, set 24hrs as the standard and anybody with any significant overclocking experience is likely to tell you the exact same thing. Just because some knucklehead, or ten of them, can run an off the wall utility they read about on Linustechtips or Reddit for ten minutes and then not apparently experience any blatantly obvious crashes while gaming is miles away from saying their system is stable or safe from corruption. You might not notice that once out of every 10,000,000 cycles your CPU is erroneously calculating that 2+2=1, but within 6 months or so your system WILL begin showing signs of it, if not sooner, depending on how much instability there actually is.
For testing any CPU, overclocked or not overclocked, and yes, even stock configurations should at least be thermal tested 915 minutes of Prime version 26.6 Small FFT) to make certain the CPU cooler mount job and cooling system have been correctly installed and configured, and whether it is an AMD or Intel processor, all of the following holds true as FACT.
(1) A steady-state workload gives steady-state temperatures; encoding does not.
(2) Simplicity in methodology; most users would find encoding apps unfamiliar and cumbersome to accomplish a simple task.
(3) Most users such as gamers never run any apps which use AVX / FMA, so adaptive or manual voltage aside, it makes no sense to downgrade your overclock to accommodate those loads and temps.
(4) Standardization; Prime95 has been around since 1996; many users are familiar with it.
For the minority of users who routinely run AVX / FMA apps, then P95 v28.5 can be useful tweaking BIOS for thermal and stability testing.
regardless of architecture. P95 v26.6 works equally well across all platforms. Steady-state is the key. How can anyone extrapolate accurate Core temperatures from workloads that fluctuate like a bad day on the Stock Market?
I'm aware of 5 utilities with steady-state workloads. In order of load level they are:
(1) P95 v26.6 - Small FFT's
(2) HeavyLoad - Stress CPU
(3) FurMark - CPU Burner
(4) Intel Processor Diagnostic Tool - CPU Load
(5) AIDA64 - Tools - System Stability Test - Stress CPU
AIDA64's Stress CPU fails to load any overclocked / overvolted CPU to get anywhere TDP, and is therefore useless, except for giving naive users a sense of false security because their temps are so low.
HeavyLoad is the closest alternative. Temps and watts are within 3% of Small FFT's.
Computronix
Author - Intel temperature guide
And this write up at the following link will explain everything else you need to consider, that I do not feel like taking the time to hand feed to you.
http://www.overclockers.com/forums/showthread.php/335813-Guidelines-for-Thorough-Stability-Testing
Facebook
Twitter
Instagram