Originally Posted by d3v0
haha what I am saying is I ran P95 and memtest for the above times, but no linpack.
Pizzaman, firefox crashes.
If Firefox is crashing, then either the overclock isn't truly stable, or perhaps a previous version of Firefox might be more stable on your particular system.
Originally Posted by Tnlgg
I run 50 passes in Intel Burn Test & get no errors so i can say the system is stable.
Other programs like Prime95 & Orthos take too long.
But linpack testing is not superior to prime testing. It's just different. Linpack stresses the system differently than prime tests.
Many people have noticed that their overclock can pass 50 or more runs of linpack with the highest stress level (or if using LinX, then with the "All" button for "Memory (MB)"), but couldn't last more than a couple of hours of any prime test. But on the other hand, I have seen many say that their overclock can handle 12 hours of each prime test, but fail within a few runs of linpack. This just means that their overclock can be much more stable than it already is.
Therefore, it doesn't matter that prime testing takes longer. That's just the way it is. So, don't be surprised if you run into problems later on down the road; if you do, then that means your overclock was never stable to begin with. Unfortunately, it doesn't matter if it feels stable now. Using an unstable overclock for a month or 2 usually results in the user suddenly experiencing problems due to instability. But that's just because the use of the unstable overclock has degraded the CPU just enough thereby revealing just how unstable the overclock really is. But it's not as harmless as it sounds: if you wait until this happens, then it will be harder to make it stable than it otherwise would have been from the beginning due to the permanent degradation that has occurred.
Originally Posted by d3v0
Load temps during Linx with priority "high" and all memory used are only 80c, quite a bit less than I thought. I get the same temps with P95 small ffts anyway.
5 passes and counting, shooting for 50. Memtest86 had 7 passes, gonna start doing P95 small/blend/large overnight for the next three nights.
That is, if I get any sleep after Dragon Age: Origins comes out
Using the High priority is not recommended. Here's a quote from LinX's GUI when pausing my mouse over "Priority class": "Lowering this level will make the system more "responsive" during testing. Values higher than Normal are not recommended."
In other words, this is absolutely unrelated to IBT's "High" setting for the stress level.
The best way to get the absolute most out of LinX is to just click the "All" button next to where it says "Memory (MB)". Be aware though that if it's in 64-bit mode, then the test may not run at all if you have the "All" button clicked. So you may have to click the button, make a mental note of the amount of memory, and then unclick the button and manually enter a value slightly
lower. For example, if I click the "All" button, mine says "3207" right now. So, I have to click the button again, and then type "3200" and try again. If it fails, then I might try 3190. In fact, I tested it just now and I ended up having to lower it by about 80 MB. But unfortunately, it's different every single time. So, if you close LinX and open it right back up, then you might be able to get away with lowering it by less (you will also notice that clicking the All button gives you a totally different amount than before). Or, you might have to lower it by even more! It's a bit unpredictable, but this is how to get the absolute most out of it. And take notice how the Problem Size is dictated by how much memory you're using.
You can also adjust this amount in the Settings for "Memory for OS (MB)". It's roughly the same thing as manually lowering the amount of memory to use for the test. So, when posting results, you can make mention of whether or not you adjusted "Memory for OS" too. Edited by TwoCables - 11/2/09 at 8:01pm