Mobile Barton + NF7-S Rev 2 and CPUID

[Megatomic]

Originally posted by: myocardia
Were you aware that using small fft's is totally useless, unless you suspect that you have a bad cpu? What you're running now is only testing your cpu and it's cache, nothing more. You would be much better off running either the blend torture test for a few days, or the custom torture test for a few hours, with a minimum of 2048 and a maximum of 4096.

Well, I am stress testing the CPU. That was my idea. As I said before, I already tested my memory using Memtest86 so why bother with it again? Maybe I'm missing something here?

 

[Elcs]

Originally posted by: Megatomic
Ok, so this time I lowered the FSB back to 200MHz and upped the mult. to 11.5 (2.3GHz) and this is what I got for Super PI:

1M in 43s
2M in 1m 38s

So, it's 100MHz faster and only finishes Super PI 2 seconds faster. Memory throughput according to EVEREST is 200MB lower at 200MHz FSB, so I think I'm gonna stick with 210 x 10.5 (2.2GHz) for a while.

If you were so 'anal about things' then shouldnt you have been anal about 210x10.5 Vs. 200x11.5? Even though the latter has a faster speed, its fsb is lower. If SuperPi is FSB limited rather than clockspeed limited.... then you introduce uncertainty. :evil:

Well, I agree with you that 100mhz isnt much at the 2+ghz end for gaming at least. However, if 200x11.5 is faster than 210x10.5 and you can do both at stock voltage, Id rather use 200x11.5

 

[Megatomic]

Believe me, I am anal about details. My wife hates it.

I'll be experimenting over the weekend when I'm done with my Bob Villa impersonations. I'm guessing that I'll go back to 200MHz FSB in the end, but I may stay at 2.3GHz if it proves worthwhile.

 

[myocardia]

Yes, you are missing something here. Memtest only tests the memory, nothing else. Running Prime95 with small fft's only tests the cpu, nothing else. You were aware that there are more than 2 parts to your computer system, weren't you? If you run large fft's, it stresses the entire computer, besides of course the hd's and optical drives. Your chipset, for instance, hasn't played a part in either of the 2 tests you've run so far, yet it has as much to do with your system's stability as either of the 2 parts you've tested.

 

[Dug]

Person A at 2.2Ghz and uses his computer = work gets done.

Person B spending all his time comparing different speeds and testing = won't get work done as fast as Person A no matter how fast their computer is.

 

[myocardia]

Originally posted by: Dug
Person A at 2.2Ghz and uses his computer = work gets done.

Person B spending all his time comparing different speeds and testing = won't get work done as fast as Person A no matter how fast their computer is.

No doubt, Dug! Now, if you could just figure out a way to be able to tell that our overclocks were completely stable, in less than 5 minutes, we would all be very grateful.

 

[Megatomic]

Originally posted by: myocardia

Originally posted by: Dug
Person A at 2.2Ghz and uses his computer = work gets done.

Person B spending all his time comparing different speeds and testing = won't get work done as fast as Person A no matter how fast their computer is.

No doubt, Dug! Now, if you could just figure out a way to be able to tell that our overclocks were completely stable, in less than 5 minutes, we would all be very grateful.

:beer:

Regarding the chipset, it's rated for DDR400 so I didn't feel the need to prove it to myself. Games and such run very smoothly at that FSB, I can't remember the last time my motherboard was responsible for a crash. Now the VGA card/drivers on the other is another story. Perhaps if I go with a higher FSB I should stress the entire system. Thanks for the tip.

 

[Elcs]

Originally posted by: myocardia

Originally posted by: Dug
Person A at 2.2Ghz and uses his computer = work gets done.

Person B spending all his time comparing different speeds and testing = won't get work done as fast as Person A no matter how fast their computer is.

No doubt, Dug! Now, if you could just figure out a way to be able to tell that our overclocks were completely stable, in less than 5 minutes, we would all be very grateful.

Depends on what you do really. My PC tells me whether its stable within 1 hour usually. Prime might kick out at 2 seconds, PiFast might die before completing, Memtest errors horrifically.... yet my games are stable and thats all I care about.