Is there any ACC value to be set? Try that.

-pWs-

this mobo have acc value in all core and per core acc setting.
which setting to use all core or per core?
how much value to set?..there have -/+ 0,2,4,6,10,12.

Ermm. You have to find out yourself about the value.
Normally use -2% o -4%

Find out the weakest core and adjust the ACC.

-pWs-

Try set ACC to:

Core 0: 0%
Core 1: 0%
Core 2: 0%
Core 3: 0%

Then when u prime, if a core fails, note it down and on the next reboot, start to use -2% and try again... if still same core fails try -4%, -6% and so on

hey, y mine one htlink can run at 2ghz although unlock?

becoz it is unlinked....or because the Bios was programmed so that ur HTLINK runs at 2Ghz

latest bios?
i have the same issue..
can set at higher multi but the result still the same after boot into windows. 2ghz....
better then previous that can only set 1.6ghz

usually those that will work (in my case) are either +2%, +4%, -2% or -4%. higher no. don't work so well...

i found that weaker core(s) need +% while healthier core(s) are ok w/ -%, it varies from chip to chip thou.

what's d nb freq ur runnin atm? lack of cpu-nb volt (when nb freq is oc'ed) will contribute to prime failure as well.
exactly...

u can always make use of htt oc bro, ~250htt won't hinder d mileage much from what i discovered. neway, ht link freq won't affect performance much, but nb freq, yes, run it as high as possible. < 1.4v is ok. my chip capped @ 2750MHz w/ 1.4v, but i'm more comfortable runnin it @ 2500MHz w/ 1.3v for 24/7.

This post has been edited by ah_khoo: Nov 16 2009, 08:54 AM

imma noob..so i limiting my nb freq to this value only... 2.6 is no where to get with <1.25v..so i just make it 2.4ghz
pls note that i only oc with multi atm since i have some issue with NB that yet to be solve


u found any improvement if u run more then the table suggest?

2.4GHz is pretty decent for 24/7 + moderate cpu clock imo. what kinda imc issue ur havin?

i dun bench much nowadays so i can't tell u whether there's any diff. perhaps few run of 3dmark06 can give u d answer.

blend -fail
large-fail
small-pass
memtest-pass around 200%

so i conclude my NB have some issue..
happend when unlock to x3 and use HTT to OC
eg 244 x 15.5 sure failed blend within 6mins / failed large in 1min.
200 x 19 can run freely...(but need extra volt compared to above )

lowering the NB, DRAM.... increase vdimm.. NB volt not helping..


AMDAthlon able to clock NB up to 5ghz iinm on the 9150e..he said faster unzip/cmpress files..if my brain serves me well la...
so like that mebbe exceed the table suggestion also dont mind.

This post has been edited by kuntawakaw: Nov 16 2009, 09:46 AM

could be ram or imc to me.

ram & nb freq, timing & divider @ 244 * 15.5 & 200 * 19?

nb freq @ 5000MHz? i'm skeptical of that, could it be software error? if phenom 2 can prime stable @ 3k i'd consider it as very gud ledi. i don't think agena have better imc than deneb.

some chips/boards don't really work well w/ ram @ higher divider (or i can say d other way round), has yet to try any amd edition ram but normal to performance ram (those that i've tried like kvr, reaper 1066 or current pair geil ultra 2133) havin issue @ highest ram divider aka 1066 & 1600. if run @ those setting, timing need to be loosened & d mileage is not as high as lower divider. it could be those subtimings issue but never am i a ram tweakin expert so either i take it as it is or use lower divider (usually i'll choose d latter).

This post has been edited by ah_khoo: Nov 16 2009, 10:33 AM

Looks like either you are not feeding enough volts to ur IMC n NB...for the IMC, try put 1.4v and for the NB 1.35v (although stock is 1.3v) try blend see how it works...

Choose custom, input the min box starting with 448 and the max box 2048, ram just fill in how much you want...preferably >1/2 of your memory...

Test around 3 hours with short testing time (such as 4-5min) to enable the looping to be reached quicker...

Any instabilities will pop up quick and if it do occur, lower ur CPU-NB values in an decrease of 3-4mhz each time....just leave the volts that I mentioned just now

Check if your board has a known issue for overvolting or undervolting first....

any particular reason for such setting sir? i've been using min @ 8 & max @ 4096 since d gud old day of A64, i'm kinda outdated ledi i guess...

8 n below 448 is kacang for the IMC to test, and can be ignored (if you have run small FFTs for hours)...the really stress is around 448 and above...this is to troubleshoot quickly without having to wait for the same 8k and the gang to finish its test first....

So when it fails, it confirms that either NB or IMC is at fault with the voltage (assuming if your rams is good to go after 12Hours of memtest)

so can i say this setting is specifically meant for imc testing? how bout d blend u u always use (for stressin all > cpu, imc & ram)? mind to share d setting? for blend i always use 8 - 4096, w/ minute @ 1 > i read that it's faster to detect error this way instead of usin default @ 15 minutes.

The 'usual' blend for 12 hours more is when i want every calculation cycles to be run continuously and stress the 3 main component (CPU, IMC and Ram) to ensure total stability....after it pass that i usually saves the settings and will proceed to entirely new settings (but still based on the voltages fundamentals of the previous successful stress result) such as drop Ram multi and play with FSB for a change...

---

Added on November 16, 2009, 12:48 pmThe 'usual' blend for 12 hours more is when i want every calculation cycles to be run continuously and stress the 3 main component (CPU, IMC and Ram) to ensure total stability....after it pass that i usually saves the settings and will proceed to entirely new settings (but still based on the voltages fundamentals of the previous successful stress result) such as drop Ram multi and play with FSB for a change...

This post has been edited by antonio: Nov 16 2009, 12:48 PM

icic... thx for d enlightenment...

@ d FSB part...

Yap..such as what I'm doing now...playing with FSB HTT part

drop divider to 1333 and push 240HTT making the same speed (1600) with the same main timings CL7....

and important infos to 790FX-GD70 ppl

make sure you check your overvolts and undervolts....

Mine jumps around 0.03v for the CPU and IMC is around 0.07-0.09 jump!

Make sure you have a multimeter and tap on the points to check...

My board OV the rams by 0.1v

here i dig up old thread for the SS..

i dont keep ss since it failed within minutes..
i tried lower down the divider...still not working ( i lazy to tweak back the timing...so take faster way, lower divider somemore...not working)

NB i did increase it up...but only 1 notch since after change to PII, never touch NB volt until yesterday (since no involve too much HTT OC using PII BE)..the error still there...
OV and UV...last time i havnt encounter this..but now when priming it will increase around 0.01v...normal isnt it...new bios make this happens perhaps.

edit. guys..can share me ur trfc value in cl7..1600mhz preferably.. i think this might be the culprit then...mine at 110ns last time..now been loosen to 160ns
found the culprit..kompom my ram cant handle the trfc 110ns

This post has been edited by kuntawakaw: Nov 16 2009, 02:37 PM