Prime95

[KNiGHTMaRe]

After getting my cpu up to around 2.6ghz i ran Prime95 one more time and it failed so i lowered it to 2.5ghz and it fails again. this is the error i receive.

 

[sun Nov 26 22:40:47 2006]

Compare your results to other computers at http://www.mersenne.org/bench.htm

That web page also contains instructions on how your results can be included.

 

AMD Athlon 64 Processor 3200+

CPU speed: 2009.13 MHz

CPU features: RDTSC, CMOV, Prefetch, 3DNow!, MMX, SSE, SSE2

L1 cache size: 64 KB

L2 cache size: 512 KB

L1 cache line size: 64 bytes

L2 cache line size: 64 bytes

L1 TLBS: 32

L2 TLBS: 512

Prime95 32-bit version 24.14, RdtscTiming=1

Best time for 512K FFT length: 23.761 ms.

Best time for 640K FFT length: 30.982 ms.

Best time for 768K FFT length: 37.642 ms.

Best time for 896K FFT length: 45.146 ms.

Best time for 1024K FFT length: 50.032 ms.

Best time for 1280K FFT length: 63.805 ms.

Best time for 1536K FFT length: 78.048 ms.

Best time for 1792K FFT length: 96.496 ms.

Best time for 2048K FFT length: 109.969 ms.

Best time for 2560K FFT length: 147.408 ms.

Best time for 3072K FFT length: 176.615 ms.

Best time for 3584K FFT length: 213.412 ms.

Best time for 4096K FFT length: 238.931 ms.

Best time for 58 bit trial factors: 5.987 ms.

Best time for 59 bit trial factors: 5.982 ms.

Best time for 60 bit trial factors: 5.969 ms.

Best time for 61 bit trial factors: 5.983 ms.

Best time for 62 bit trial factors: 10.749 ms.

Best time for 63 bit trial factors: 10.764 ms.

Best time for 64 bit trial factors: 13.888 ms.

Best time for 65 bit trial factors: 13.782 ms.

Best time for 66 bit trial factors: 13.886 ms.

Best time for 67 bit trial factors: 13.850 ms.

[Mon Nov 27 00:44:20 2006]

Self-test 1024K passed!

[Mon Nov 27 01:01:21 2006]

Self-test 8K passed!

[Mon Nov 27 01:17:05 2006]

Self-test 10K passed!

[Mon Nov 27 01:35:39 2006]

Self-test 896K passed!

[Mon Nov 27 01:41:51 2006]

FATAL ERROR: Rounding was 0.5, expected less than 0.4

Hardware failure detected, consult stress.txt file.

[Mon Nov 27 06:46:33 2006]

Self-test 1024K passed!

Self-test 8K passed!

[Mon Nov 27 06:53:30 2006]

Self-test 10K passed!

FATAL ERROR: Rounding was 0.5, expected less than 0.4

Hardware failure detected, consult stress.txt file.

[PhoetuS]

Are you asking something here, or just reporting to us what happened?

 

 

If you want to know why you got the error message, it is because your overclock is unstable. Need to tweek your settings until you don't get any error messages...

[Mr. Jules]

try lowing the fsb, or upping the vcore a tad.

[KNiGHTMaRe]

try lowing the fsb, or upping the vcore a tad.

 

i lowered it from 2.8 to 2.6 then to 2.5 with all the small increments in the middle. i did not try the vcore, but would that be the 1.425??? i changed it to 1.5 because i read its stable there, but i will change it back to 1.425 once i get back home.

[not_this_punk]

check your ram timings also. check your ram divider: lower it past 1:1.

but you could definitely get more help if you'd share more info on your ram. are they PQI Turbo or standard PQI? if they're the non-turbo version try increasing the CAS Latency (tcl) to 3 and see how they're holding up after that.

oh and make yourself a more accurate signature.

[not_this_punk]

it would also be wise to carefully read the amd overclocking guide, located here: http://www.dfi-street.com/forum/showthread.php?t=20823.

[KNiGHTMaRe]

i read that guide the best that i could. here's what i did.

 

when i first built the pc and i used a guide i found on here to get the dfi boards up and running.

 

the only setting i have been tweaking was the FSB setting. it was set too 200. in my pcwizard2006 it said my cpu was at 2009.11ghz. i then started going up with 5 - 10 increments. saved bios, restarted and made sure XP booted up etc. but i did not run Prime95 after each one. i had it up to 280 when i decided to run Prime95. it failed, so i lowered it to 270 and then 260. And each time it failed. I will try 250 and below when i get home. I didn't switch the voltages until i read somewhere about keeping the volts at 1.5 so i did that and thats when i got the error in Prime95. i reset the voltages and tried again and still failed. im gonna tweak with it some more, but any ideas what range i should be in for the FSB?

 

by the way, this is memory i have atm

http://www.newegg.com/product/product.asp?...N82E16820141167

[MACarter02]

probably more voltage is the easiest first try.

[not_this_punk]

oh man...look...it says ddr333...that means that this memory is supposed to run fine at 333 mhz - 2 * 166 mhz actually.

 

your ram speed is calculated like this: ddr speed = fsb * ram divider. note that ddr speed is not 333 mhz, but 333/2=166 mhz.

 

your standard fsb is undoubtebly 200 mhz. so your ram divider should be ddr speed / fsb = 166 / 200 = 0.83. ram dividers usually show up as fractions in bios, but this is not a problem, as you can do the math.

 

so, at a fsb of, let`s say, 250 mhz, your ram speed would be (using your normal ram divider) 250 * 0.83 = 207.5 mhz. that would mean something like ddr415. but remember, your ram is made to run in ddr333 mode. so you must find a divider, which, multiplied with the fsb, would result something a little closer to 333.

 

there is the possibility of ram overclocking though. my ram is ddr400, but right now i`m using it in something like ddr500 mode: i have a 300 fsb and the 0.83 ram divider. 0.83 * 300 = 249. (249 * 2 = approx 500 hence ddr500). the difference is that my ram is a little better at oc'ing, as it's made to run at a default speed of 200 mhz and with low (or tight) timings.

 

that`s where timings intervene: most manufacturers give you 4 numbers in the ram specifications. mine (at 200 mhz) are 2-3-2-5. without going in too many details, the lower those numbers are, the better. because when increasing the ram speed, if the ram isn't stable, you can increase those timings as well, and stabilize it. so you can see in my sig that i don't have 2-3-2-5; instead i have 2.5-3-2-5, in order to stabilize the ram overclock.

 

the problem is that your ram probably has high timings, and even worse, it's made to run in ddr333 mode. so if you increase the frequency, you must increase the timings also, but your timings are also high so you don't have anything to increase (probably). that depends on your stock timings though, but this is my hunch.

 

i hope i succedeed in helping you solve your overclocking mysteries.

[WintersFury]

I'm going to guess that's a 3200+ Venice? I have a 3000+ Venice and it's only stable to 2.4 with 1.425v Tried 1.475 at 2.5 and still couldn't get it stable. You may just not have that great of an OCing chip. Watch your temps as you up the volts. That's why I didn't do over 1.475 with mine.