Printable View
hey :)
my G.Skill 4GB 800MHZ Pi cant get stable on 4-4-4-12 with my OC to 890MHZ.
is this is so hard to get this timing on 890MHZ? i think it will be easy for him.. :down:
the VRam is on 2.1V, he need more voltage to get this timing?
what will help me to get stable on 4-4-4-12 at 890MHZ?
tanks.
help?
when i set 4-4-4-12 at 890MHZ on 2.1V i cant get a post.
which mobo ? CPU ? FSB ram ratio ? have you tried 5-4-4-12 ?
sorry, i always write my rig.. i forgot this time.
e8400 stepping CO on TRUE120
ASUS P5Q PRO bios 1613
G.Skill 4GB 800MHZ Pi CL4
1:1
i try very soon, i got some problem with the OC.
i put these setting and its crash after some hours, but pass 1 and half hour on OCCT. any help? tanks.
here are the setting:
Ai Overclock Tuner : Manual
CPU Ratio Setting : 9
FSB Frequency : 445
PCIE Frequency: 100
FSB Strap to North Bridge : Auto
DRAM Frequency: 890MHZ
DRAM Timing Control: 5-5-5-15
DRAM Static Read Control: Auto
Dram Read Training : Auto
MEM OC Charger : Auto
Ai Clock Twister : Auto
Ai Transaction Booster : Auto
CPU Voltage : 1.366
CPU GTL Reference: Auto
CPU PLL Voltage : 1.50
FSB Termination Voltage: 1.30
DRAM Voltage : 2.10
NB Voltage : Auto
PCIE Sata Voltage: Auto
Load Line Calabration : Enabled
CPU Sread Spectrum : Disabled
PCIE Spread Spectrum : Disabled
CPU Clock Skew: Auto
from this guide: http://www.computerlounge.co.nz/foru...x?g=posts&t=50
Try the above plz and set TRFC to 50 or 55Quote:
Originally Posted by PakaPaka
Does the CPU pass small FFT ? I didn't like the 1613 bios on the P5Q Pro, I reverted back to 1406
what is it TRFC?
its this number after the 5-5-5-15-50-2T that usually need to be 36? its cal TRC, this is what you meed?
well, its stable i think.
http://img410.imageshack.us/img410/3...9134301su9.png
looks good?
the MB temp in ASUS probe is the NB temp?
tanks :)
Mobo temp is mobo temp, it's NOT NB temp !! also try to run prime95 plz small FFT, your orthos seems to test 8k all the time which is not really a sign of stability as it's the easiest test...
TRFC is row refresh cycle time in the bios try 45-50
ok i tast now with prime95 small FFT, but how much time is it test? until i stop it?
at least a few hours mate, if it fails, up Vcore a bit more
i don't think i need more Vcore, i get stable with 1.328V on 450*9=4.05GHZ.
prime95 shows only FFT length 8K.. its test 11 and still 8K.
i run blend and see if its stable for one hour maybe, after that i do test overnight.
so if i am stable now that the ram is on 5-5-5-15 at 890MHZ at 2.1V.
how can i get post on 4-4-4-12?
what will help me to set this timing? tanks.
I have no clue why prime and orthos stop at 8K on your rig, you can select custom too when you start prime and set 20k then
don't matter i test with OCCT.
so if i am stable now that the ram is on 5-5-5-15 at 890MHZ at 2.1V.
how can i get post on 4-4-4-12?
what will help me to set this timing stable?
and which setting i need for 4.5GHZ? if i can make it, i want to try.
now its like this in bios:
Ai Overclock Tuner : Manual
CPU Ratio Setting : 9
FSB Frequency : 445
PCIE Frequency: 100
FSB Strap to North Bridge : Auto
DRAM Frequency: 890MHZ
DRAM Timing Control: 5-5-5-15
DRAM Static Read Control: DISABLED
Dram Read Training : Auto
MEM OC Charger : Auto
Ai Clock Twister : Auto
Ai Transaction Booster : Auto
CPU Voltage : 1.336V in cpu-z at load
CPU GTL Reference: Auto
CPU PLL Voltage : 1.56
FSB Termination Voltage: 1.26
DRAM Voltage : 2.10
NB Voltage : 1.26
PCIE Sata Voltage: Auto
Load Line Calabration : Enabled
CPU Sread Spectrum : Disabled
PCIE Spread Spectrum : Disabled
CPU Clock Skew: Auto
give me some numbers for 4.5GHZ or less, that i now from where to start, tank you very much my friend.
AI clock twister to lighter and trfc to 55 will prolly make them stable at 900mhz 5-5-5-15 2T (it all depends on ya rams ofcourse)
Disable DRAM Read Training, Mem OC Charger.
Command rate = 1N, 4-4-4-12-2-50-5-3 7-3-4-4-5-4-5 13-4-1-5-5
Try timings along those lines, and manually set AI Transaction Booster to 9. Raise Vdimm upto 2.2-2.24v and see how you go.
Personally P5Qs I've used don't like CAS4 on 2N command rate, even with 1N command rate they didnt like them much but there is a fighting chance if you can get the skews right. Seems the 1:1 divider is crappy in my experience.
tanks a lot mikeyakame, but 2.2V-2.24V isn't way too much for 24/7 work?
2.1V didn't enough for 4-4-4-12 on this speed?
how many i can get with the cpu?
i got 55C max temp at 1.336V 4.00GHZ 445*9.
so how much voltage i need for 4.2GHZ or higher? tanks.
i just need some numbers to now from where to start.
Depends on what the rated voltage is the memory, and how well it scales with voltage. Some sticks scale poorly and adding more voltage just makes more errors, others crave voltage. You need to figure this bit out as all memory sticks are different, and I can't say I have any experience with your G.Skills.
Most CAS4 sticks need 2.2V+, hell I haven't seen many that don't in my experience. Corsair Dominator CAS3 1GB DDR2-800s needed 2.35-2.40v just to run those timings.
Ive run 2.4-2.45v 24/7 with a Corsair Dominator Ram Cooler and other than shortening the life of memory to less than a year, it was fine its all about the air flow and cooling mostly.
well i cant get this timing mikeyakame :(
i set everything you write and try little bit more but i cant get a post.
so i stay on CL5 :( :(
now tell me guys, i am now on 4.00GHZ.
is 4.2GHZ give me a lot of more performance?
and why e8400 is such a good overclocker if he cant get over 4.5GHZ on air for 24/7? mostly of the e8400 cant.
Most E8400 do 4-4.2Ghz daily mate, the 4.5 24/7 ones are rare believe me...benching they can do 4.8Ghz
I would give 1406 bios a try and maybe also one of Ket's modified biosses ( they have better ram support, but I hardly noticed anything usefull with my boards and rams)
The values I gave you are a starting point. I can't just wave timings and hope they post, CAS 4 is hard to achieve on a P45 none the less. They aren't all that suited to low CAS timings. CAS5 needs much less voltage so stick with that.Quote:
Originally Posted by PakaPaka
PakaPaka,
may be your ram simply cant handle CAS4-4-4 at the ram clock speed.
well i got some problem.
this is the setting for 4.00GHZ:
Ai Overclock Tuner : Manual
CPU Ratio Setting : 9
FSB Frequency : 445
PCIE Frequency: 100
FSB Strap to North Bridge : Auto
DRAM Frequency: 890MHZ
DRAM Timing Control: 5-5-5-15
DRAM Static Read Control: DISABLED
Dram Read Training : Auto
MEM OC Charger : Auto
Ai Clock Twister : Auto
Ai Transaction Booster : Auto
CPU Voltage : 1.336V
CPU GTL Reference: Auto
CPU PLL Voltage : 1.56
FSB Termination Voltage: 1.26
DRAM Voltage : 2.10
NB Voltage : 1.26
PCIE Sata Voltage: Auto
Load Line Calabration : Enabled
CPU Sread Spectrum : Disabled
PCIE Spread Spectrum : Disabled
CPU Clock Skew: Auto
after 5 hours of stable test that pass, its crash like after 3 hours.
i just don't now what to do :(
this CPU give me 4.05GHZ on DFI P45-T2RS PLUS with 1.328Vcore and 1.26VTT and 1.200VNB and 5-5-5-15 2.1Vram stable 100%.
but here on the ASUS P5Q Pro its not stable on 1.336Vcore and 1.26VTT and 1.26VNB and 5-5-5-15 2.1Vram :(
OK update.
i think i wasn't stable but now i think i am.
http://img201.imageshack.us/img201/3...ableibtyk4.png
the settings:
Ai Overclock Tuner : Manual
CPU Ratio Setting : 9
FSB Frequency : 450
PCIE Frequency: 100
FSB Strap to North Bridge : Auto
DRAM Frequency: 901MHZ
DRAM Timing Control: 5-5-5-15
DRAM Static Read Control: Disabled
Dram Read Training : Disabled
MEM OC Charger : Disabled
Ai Clock Twister : Auto
Ai Transaction Booster : Auto
CPU Voltage : 1.3500V in bios, 1.336V at load, 1.328V at idle.
CPU GTL Reference: Auto
CPU PLL Voltage : 1.56
FSB Termination Voltage: 1.26
DRAM Voltage : 2.10
NB Voltage : 1.26
PCIE Sata Voltage: Auto
Load Line Calabration : Enabled
CPU Sread Spectrum : Disabled
PCIE Spread Spectrum : Disabled
CPU Clock Skew: Auto
so now how many times should i run IntelBurnTest?
am i stable on 5 run times with 100% successful?
tanks :)
If it's stable mate in ya daily apps and usage it's okay... most users tell at least 20 IBT runs, but that linpack is sooo hard on all ya components that I don't recommend it to be used for long term usage... also it can be 100 run linpack stable yet cap out in prime blend under a 1 minute... if ya daily applications don't crash you are fine !!
Good you got it going mate !
tank you my friend :)
i running WCG 100% 24/7, so i see if i am 100% stable in a few days.
so what do you say? should i going to the 4.2GHZ~ ? i mean 24/7 stable.
my rig can caring that? the temps and mobo?
i don't want to get over the safe range, 1.4Vcore.
and one other question, what is giving me more performance?
Ram speed 1080mhz~ cl 5-5-5-18 or 900mhz~ cl 5-5-5-15?
tank you very much.
I think 1080mhz will do better mate as you can use a lower PL setting... when you run ya rams in sync on P5Q it uses a high PL level , when you use a divider you can go at least one or two values lower and increase bandwith and co (but don't expect miracles of it)
Nice to see you using cycles on WCG mate, top job, I'm a folder myself :)
WCG is great for stable test, run 100% 24/7 and you are 100% stable ;)
and help the world too :D
well i can't get stable 1080mhz on 5-5-5-15\18.
how can i get more stable with this ram speed?
more low timing? :-/
more NB Voltage?
other setting?
tanks.
You're completely right there, Lee.Quote:
Originally Posted by Leeghoofd
I run 50 IBT runs from time to time, maybe because I have a compulsive disorder, to make sure my QX9650 is stable at 4 ghz @ 1.375v (1.360-1.368v load; max temp 68-71C) but I really should stop... who knows what damage I may be doing?
I don't even keep it at this speed; I run it 24/7 at 3 ghz -3.71 ghz @ 1.25v.
Besides, if you know your system is stable at a certain temp (that you will -never- reach in day to day usage anyway), why is it really necessary to heat it up 10C *PAST* That temp? (comparing prime95 @60c max vs IBT @ 70c max?). Lets take a situation like, you know your CPU is fully prime stable (small/blend) if its at 65C or lower at that vcore; Prime never exceeds 60C, but IBT goes up to 70C and therefore it fails. Does that mean you're unstable? Of course not. You're unstable in linpack, but not anything else since nothing heats up the CPU that much.
Now it's nice to say "yeah, i'm ROCK solid stable and nothing on the planet can crash (the CPU...RAM is another matter), but you get the point.
The main reason I see linpack being good (even though you are actually prime stable) is because you run linpack for a far shorter time, which therefore gives you less downtime, and your CPU is stressed less *IF* you can keep the the temps down. But the compromise is that linpack heats your CPU hotter than anything else, even past its "stable" temp point. So people may wind up pushing their vcores *higher* than necessary for 24/7 usage just to get linpack stable. Linpack is an official Intel tool though so you know if you pass it, nothing "official" can bring the CPU down.
My question is, what if someone (similar to what the author of BurnBX/BurnP6 did) manages to write a program that heats your CPU 5C hotter than Linpack? Would people flock to that as the new ultimate stability test?
IBT get me 100% stable, before that i crash in 100% WCG running and playing Burnout Paradise Ultimate Box.
after i tested with IBT i now sure i 100% stable this time, after 5 runs in IBT.
the max temp is 67C but now on WCG 100% running is only 50C.
so fine only for 18 minutes like my CPU is hottest but i think its worth that now i 100% stable.
i hate to wait 5 hours or more with Prime95 or others, IBT is short and good.
but is it IBT 5 runs make some damage?
i run this like 6 times maybe in different settings.
i am not stable yet :(
how?
maybe the row refresh cycle time? its on 50.
and i pass 5 run on IBT.
but if i in OC the games are crashes and i got 2 BSOD 1037.
those are the setting:
Ai Overclock Tuner : Manual
CPU Ratio Setting : 9
FSB Frequency : 450
PCIE Frequency: 100
FSB Strap to North Bridge : Auto
DRAM Frequency: 901MHZ
DRAM Timing Control: 5-5-5-15-50
DRAM Static Read Control: Disabled
Dram Read Training : Disabled
MEM OC Charger : Disabled
Ai Clock Twister : Auto
Ai Transaction Booster : Auto
CPU Voltage : 1.3500V in bios, 1.344V at load, 1.328V at idle.
CPU GTL Reference: Auto
CPU PLL Voltage : 1.56
FSB Termination Voltage: 1.26
DRAM Voltage : 2.10
NB Voltage : 1.26
PCIE Sata Voltage: Auto
Load Line Calabration : Enabled
CPU Sread Spectrum : Disabled
PCIE Spread Spectrum : Disabled
CPU Clock Skew: Auto
what should i change?
Are you still on 1613 bios ? Try one of Ket's modified biosses mate if the problem still persists
i not the guy for modified biosses, i prefer retails.
so get the 1406 bios from ASUS site?
he's much better stable from the 1613 bios?
??
so should i go back to 1406 bios?
can i go go back in bios version? from 1613 to 1406?
yes you can via eg Koda's flashing tool , but you need an USB stick for that or a CD ... read the instructions how to setup the stick or cd carefully pz... it worked fab every time link here mate :
Koda's ASUS Flashing tool
NEVER FLASH OC'ED PLZ
yes i now, i update bios before.
first i clear CMOS then i update.
can i just update with EZ Flash 2 from bios tools?
Don't think 1613 allows downgrading, you just have to try...
ok so i go with the Koda's.
now i try one more time to get stable, if not.. go back to 1406.
i set NB to Auto and PLL to 1.52V, if its no good please tell me.
tanks man for all your help :)