Give the benchmark program in my sig a run
Give the benchmark program in my sig a run
My work and bench rig:
i7 920 D0 3939B303 @ 4.2Ghz + ASUS P6T Deluxe V2 w/ 0006 BIOS + 6GB Corsair Dominators CL7 @ DDR3-1600 7-7-7-20-1T + Intel X25-M 80GB SSD G2 + Powercolor HD5870 1GB @ 900/1300 + Dell U2410 H-IPS 24" Screen (pics and more)
My review site: Erodov
To me, stable is 30 runs of Linx with all RAM tested, maybe not to everyones taste but if it passes that then I never have any problems from anything relating to the overclock
A lot of people flaunt and brag about how little voltage there chip needs at a set speed, I can run OCCT, Prime95 and Orthos for hours but my machine won't pass Linx and requires more Vcore.
I also think that world record overclocks should only count if stable enough to use daily with no problems.
I remember when people were getting Crazy overclocks with Phenom 2 on DICE and LN2, it made people go wow and gave them hope that they could get decent clocks on air and water, only in reality most Phenom 2's struggle to get to 3.8Ghz in the real world. So a lot of people were disappointed because some 'overclockers' thought that just because it could run SPI and a few benchmarks that it's awesome and stable.
All world record overclocks should at least do 5-10 runs of Linx with all RAM tested before they even get considered to be praised in my opinion.
As some other people here already pointed out, there is no real use in discussing how to define stable. It is a subjective matter. Anyone more or less defines "stable" the way he thinks it suits him.
In theory a system can only be 100% stable under given circumstances, because there will always be a way to get it unstable by changing one or more of these circumstances.
On the condition that these said circumstances are met and in a certain degree of tolerance, a stable system should be able to complete all the given tasks (here working through error free -impossible- software) in a finite amount of time and without producing arithmetic errors. If it does exactly that it could be called stable. BUT, there most likely won't ever be neither completely error free software, nor hardware. So, as already said, the discussion is kind of moot.
But using test software, such as for example the Linpack library, that was coded specifically to test/benchmark a certain architecture, is a good way to get info on the subject of stability.
Quote from one of our professors:
"Reality is hiding in the imaginary part."
for me is stable too LinX for 24/7 using. For others its maybe only benchmarks as superpi 32M or 3Dmarks. I think, all from OCCT, Prime or LinX is good fot testing stability system.
ROG Power PCs - Intel and AMD
CPUs:i9-7900X, i9-9900K, i7-6950X, i7-5960X, i7-8086K, i7-8700K, 4x i7-7700K, i3-7350K, 2x i7-6700K, i5-6600K, R7-2700X, 4x R5 2600X, R5 2400G, R3 1200, R7-1800X, R7-1700X, 3x AMD FX-9590, 1x AMD FX-9370, 4x AMD FX-8350,1x AMD FX-8320,1x AMD FX-8300, 2x AMD FX-6300,2x AMD FX-4300, 3x AMD FX-8150, 2x AMD FX-8120 125 and 95W, AMD X2 555 BE, AMD x4 965 BE C2 and C3, AMD X4 970 BE, AMD x4 975 BE, AMD x4 980 BE, AMD X6 1090T BE, AMD X6 1100T BE, A10-7870K, Athlon 845, Athlon 860K,AMD A10-7850K, AMD A10-6800K, A8-6600K, 2x AMD A10-5800K, AMD A10-5600K, AMD A8-3850, AMD A8-3870K, 2x AMD A64 3000+, AMD 64+ X2 4600+ EE, Intel i7-980X, Intel i7-2600K, Intel i7-3770K,2x i7-4770K, Intel i7-3930KAMD Cinebench R10 challenge AMD Cinebench R15 thread Intel Cinebench R15 thread
I was going to start a new thread, as what I'm looking for is a bit more specific, and at the same time a bit broader...
...unfortunately, I still can't post new threads for some reason. So, if a mod sees this: PLEASE GIVE ME RIGHTS TO POST NEW THREADS!
Anyway, what I wanted to know is:
What are the best apps for stress testing CPU and GPU overclocks, while keeping all power saving options on? I've found overclocks that pass 100% load tests overnight can't pass general-use stability.
You can read my rantings on HardOCP for now:
http://hardforum.com/showthread.php?...post1037437644
Now you may even get an infraction for being a thread digger... (hope i do not get one too, lol)... Well, there are seeeeveral things "between heaven and earth". One could say that "processors" degradate over time and may start asking mooooar voltage for simple things, crashing when not @full voltage that's a point. Another way to see this problem is that despite an "burn" test stresses 100% generally it is a padronized load, during "normal" use you may be multitasking, randomly one error could result in a crash in a normal task and not crash a CPU crucher app. Overclock tend to mess with pooling times, if one thread is waiting for a data and another thread runs out of sync a crash may occur, the way software is implemented, if it interacts with the hardware abstraction layer, it's hard to explain and depends on several factors that make some processes more prone to crash while the system is overclocked. On the VGA side of things, e. g. if you run furmark in a "stable config" it will probably not "crash" and the same config it may show thousands of errors in an OCCT GPU test, this error could end up being a minor/unnoticed visual glitch, one texture in a frame or something like that. The only, and hard way to track those issues is saving the bsod and searching what had caused the exception.
Stable to me is 48hrs of large fft primes on all cores at once.
A run of s&m at normal length, the mem test.
As well as a 8hr run of y cruncher with the mem setting maxed.
Then 2hrs of each games I have pretty much, and days of use of generic things like web surfing and watching vid's.
If I can do everything without it screwing up in any way, then it's stable to me.
Linx I personally believe is a waste of time.
It's test of mem or cpu stability is low balling it by quite a bit.
It's never been no where near what I need to get 100% stable.
OCCT lmfao, that program is a complete joke.
Err I probably shouldn't put it so bluntly...
Edit:
Oh yeah, furmark, I run that for an hour to see if my vga cooling is ok.
It's not a test I rely on for stability, just for cooling.
Last edited by NEOAethyr; 06-27-2011 at 07:30 PM.
Well, I found Crysis 2 with the new Ultra settings to really push things... it has just replaced Just Cause 2 as my quickest-way-to-test-an-overclocked-gpu test. 835 core passed the mustard in Just Cause 2, but had to back it down all the way to 823 on the overhauled Crysis 2
Bookmarks