TwoCables

This is all in LinX's Read Me file and it's in this thread, however, to summarize: 50-100 runs using all the available memory. Be sure to browse the Settings to customize it to your liking - especially to switch it to 64-bit.

__________________

freakb18c1

does any real application stress as much a LINX?

Laforet

Well, maybe 8 hours of small FFT works better however in real life few applications works without moving data between the cache and system memory so it is not as relevant. My personal experience has shown that 20 passes of LINPACK is more than enough to pick up any instablity on the CPU's part with the RAM set to run 1:1 at loose timings. After that I will try to OC and tighten the RAM, if it can pass memtest86+ it must be good.

I have always overclocked according to the aforementioned 20+8hr methodology and nothing have failed on me yet.

alex98uk

I'm running LinX 0.6.0.

What I see is "Problem size", "memory to use" and "times to run".

So, I have it in 64bit, I will set it to 75 runs for now (any idea how long that might take?). There doesn't seem to be an option to use all available memory, I have to choose a number. So, assuming I have 4GB of RAM, should I choose the max value of 3584? (It says I have 2542Mb physical memory available). What to do for problem size, it's currently on 5732?

__________________
Winamp FLAC > Xonar D2 > Optical Zero DAC > Little Dot MKIII > Sony MDR XB700

British Overclockers Club

Laforet

LinX is quite user friendly so just try using the default parameters with the maximum memory that it will allow you. However the only difference that makes is that it runs longer with more memory, RAM stability is always better tested using MemTest86+ outside an OS which adds its own overhead

TwoCables

No, but an overclock is unstable by nature. So, we need to provide the system with an overclock that's so stable that it's "practically" 100% stable for our needs. But if you were to run your system as a server, then it would never be stable enough no matter what.
Click the word "All" next to "Memory to use". For some reason, he changed the GUI so that this button doesn't always look like a button (I guess it's cleaner that way).

As for how long that might take: it could be a few hours, but it depends on how long one run takes. So time the first or 2nd run (the timer in the GUI might not be very accurate as to how long the run actually took). It should be exactly the same for each run. Then just multiply that time in te value of seconds by 75, and then divide by 60. But again, don't use the thing that shows how long the run took. It was inaccurate in 0.5.8, and his changelog doesn't show that he fixed it. My system does one run in about a minute and 45 seconds, but the thing says that it takes about 60 seconds, or some inaccurate reading like that.

As for the problem size: that will change automatically when you click "All".

__________________

alex98uk

Excellent, didn't notice the "all" button. I'm going to do some low numbered passes such as 10 to get an idea how long each one takes

__________________
Winamp FLAC > Xonar D2 > Optical Zero DAC > Little Dot MKIII > Sony MDR XB700

British Overclockers Club

TwoCables

Oh. You'll know after about 2-3. Or, you may be able to just time the first one and go by your clock instead of LinX's.

__________________

alex98uk

Heh, well, I did a 10 pass and got about 95% through at heard a click and everything turned off. Heat was also an issue, pushing 76c, so I guess 3.6ghz is probably a bit too far for my little old Q6600. Went down to 3.5ghz and have it relatively stable at 1.41v (only done 10 passes).

Do I now do more passes at the same voltage or keep reducing voltage until it fails?

__________________
Winamp FLAC > Xonar D2 > Optical Zero DAC > Little Dot MKIII > Sony MDR XB700

British Overclockers Club

jam3s

I still stand by IBT and Prime95 Small FFTs.

If I can do 30 passes and 24 hours on Small FFTs It's good for me.

Folding is also pretty intensive as well..

I'm stubborn that way I suppose

__________________