??? I am the only one able to boot at 4Ghz (500FSB) and fully stable at 3.8Ghz with an X48 ?
Printable View
Is your board a true rampage or a maximus flashed to the rampage?
What would one think the highest 24/7 volts are for this thing on air? Temps only hit 64C on prime small FFTs. Gaming no matter what doesnt go over 51C...
Hmm. Not used to this BIOS... had ASUS boards too long :p:
Looks like I might need a hand.
Honestly I havent heard much success stories with flashed rampages overclocking the new quad cores.
One update in the Q9450 thread in case some people don't read the other one:
http://www.youtube.com/watch?v=kZSocwaQntA
Finally bootable at 524MHz. :up: Not 525, though... oh well.
What do you guys think about one of these with with an EVGA 750I FTW board and some 4gb of Crucial Ballastix. I am doing a new PC build (first time in a while) and will probably pick up both. I mainly just want to get a stable 3.2 or 3.4 for the little extra boost in gaming.
Nice results so far guys!
513FSB was the limit with an E6600 proc in the Rampage before N/B water cooling. I won't test the limit for the Q9450 yet p
Memory tweaked a bit now (I never noticed the "C1/M1" stepping before. Edit: Yes, M1 is reported by Everest for others too):
http://i224.photobucket.com/albums/d...way/memory.jpg
hey RUNAWAYPRISONER i know this is a little off topic but i notice your using a kingwin CPU HSF. Do you like it? How does it compare to the Ultra 120 Extreme or the zalman9700?
Well, I can run OCCT and Prime at 500MHz if that's useful at all? Gonna need a real graphics card because the onboard one keep on dying at high voltages after about an hour or so. Seems like overclocking the board would also overclock the onboard graphics?
This thing, by the way, is identical to the Xigmatek HSF except for the logo on the fan. Everything else is identical to the touch.
so would you consider it better then the ultra 120 extreme? cause im shoppin for a new HSF cause im questioning my zalman 9700 nt at the moment.
It's kind of decent, I guess. At least about 10-20C better (cooler) than the ACF7 Pro.
well yea i guess your right. 500Mhz kinda speaks for itself huh?
Indeed. Bootable up to 524. I could only boot at 513 with the ACF7 Pro. Hmm... that's only 6MHz short from WR. :) Maybe it can go higher if I volt mod the board... but WR is not what I want. The board itself might have won a WR for mATX overclocking, anyway.
so what are you up to? prime95 stable wise?
500. But I can't go beyond 30 minutes because the onboard VGA keeps dying and crashing the computer. I need the 9800GTX here....
Anyone... what is the general consensus on the highest 24/7 volts on high end air????
i don't have a 45nm quad but intel states max VID is 1.3625v. i would not go over 1.4v load, 1.4v vtt or 1.4 pll on any 45nm chip.
Ok, Im under the voltage on idle and load. VTT is 1.4 from 1.3 stock (not sure if I can lower that or not, I will have to check stability). And I dont believe my bios/mobo has a 'pll' option (or its not named pll).
So 1.4 is the highest 24/7 voltage (to get 2 years out of this chip possibly).
Wow, I'm in the same shoes as you...I'm not trusting my Zalman 9700NT anymore either and i'm actually surprised it isn't cooling better considering it's very shiny and straight bottom...I ordered the Xigmatek Red Scorpion cooler which is an updated version of the much hyped S1203 cooler...only thing I didn't like was the fan's cfm rating so I ordered a 110cfm fan to use instead of the stock one ;)
I'm trying to cool a QX9650 so I hope it does better than the Zalman.
Pish, seen higher output fans then that:
http://www.sidewindercomputers.com/de12tfexhisp.html
http://www.sidewindercomputers.com/de12gf.html
yes, so have i, i was mainly stating how loud those fans are. i have a silverstone fm-12x that i don't use and it is so loud @12v that i can't use the phone in the same room :shakes:
On water it's a 24/7 system @ 3.8GHz now for me after hours and hours... of stress testing /zzzZzz
High end air is not far from my water system.
It depends a lot on the cooling on the N/B too.
But 3.8Ghz too I guess.
Anyway, besides world records or such I don't see the point to go to 4GHz+ when we are already lucky to have a 100% stable 3.8GHz while staying in safe voltage ranges.
I burned too many boards and too many RAMs. I don't want this to be the first processor I burn ;)
Aww! Don't be... safe! Intel doesn't like it when you are! :lol:
But seriously, I think on either extreme air or water, 3.80GHz should be the limit. Unless you think you are adventurous enough to stand 4.00GHz (1.48v vCore!). 3.80GHz at 1.38v is really a lot to me already. And the X3350 chip only required 1.304v for 3.80GHz, too... :(
Got my stuff in a week ago, been running it at stock and decreased CPU voltage to 1.05 from 1.2. Have spent the last few hours trying to stabilize my overclock now. So far 3.6 stable at 1.375 and +1NB for 2 hours. Time to try 3.8. Only thing that kils me is the difference between the motherboard reported temp of 43C and realtemps 57C. Not sure which to believe as it is such a big difference. Realtemp reports my idle at stock as 42C and im on water.
Ouch... I just found out that this is all I need for 475:
vCore: whatever works.
vPLL: 1.50v in BIOS
vFSB: 1.20v in BIOS
vNB: 1.43v in BIOS (1.45v real)
So... almost all stock! vCore is a bit high, though. (But everything else is shockingly lower than what X3350 required!)
Well 3.8 was a no go. I am unwilling to up my voltages any more than I did for 3.6 so I guess I'm stuck here. Not too bad tho, now I'm just gonna work down the voltages a bit for my 24/7. Running 30 minutes at 1.35V or reported 1.31V in EasyTune.
What voltage are you guys finding is nesscary for 3.6GHz?
Rich, i have my voltage set to 1.356 in bios, it is showing as 1.3v in EasyTune and cpu-z(1.442) is reporting it at 1.28 which is probably bugged. This is prime stable for 4 hours. My temps are 41C reported by EasyTune and 56C reported by RealTemp. Not sure which to believe as RealTemp seems to report my stock temps @ idle somewhere near 45C.
I have it in 1.280v and runs very well
24/7 @ 3.6GHz
http://img241.imageshack.us/img241/410/oc01hj8.png
Look likes CORE#1 is not running ... Temps on other 3 cores are ~45*C while core#1 = 27*C ...
Might be working fine Runaway. But I am getting really bad temps at stock voltage and stock speeds. 40C idle, on watercooling. Then with increased voltage and overclock under full load it only jumps to the mid to high 50's seems like too little a jump to be reporting correctly. As far as Easytunes reading, i would never trust them as I believe they report the socket temp. On a positive note at 3.6 I ran a 13.588 1M SuperPi and a 19106 in 3dMark06 (750/1100 9800gx2)
These are what I got ... from 2 different boards:
Asus P5K-Dlx, OCZ DDR2-1066
vcore: 1.25v (BIOS)
vdimm: 2.10v(BIOS)
Temp idle: ~33*C (RealTemp)
Temp full loaded: ~48*C (RealTemp)
http://img136.imageshack.us/img136/7...4503600mk6.jpg
Abit IP35-Pro, OCZ DDR2-1066
vcore: 1.35v (BIOS)
vdimm: 2.10v(BIOS)
Temp Idle: ~42*C (RealTemp)
Temp full loaded: ~58*C (RealTemp)
http://img100.imageshack.us/img100/5...pro3600ib1.jpg
Nice chip, SBinh. I guess some are just nicer than others. :) In fact, I bet I can get 4GHz stable with yours at 1.42v vCore... whereas with this chip I have right now, 1.50v is required for 4GHz stable...
It could just be because the motherboard is trying to compensate for the PSU, though. I shall see to that later on.
3.6GHz on the Rampage board is really easy.
I just set strap to 333, FSB frequency to 450, the right voltage for my RAM and all the rest on Auto.
It's OCCT stable and I can see the CPU voltage fluctuate around 1.27V under load.
(I checked the other voltages selected by Asus with Everest and it's safe/ok).
Usually "Auto" settings was a NO for me when overclocking but it seems Asus improved a lot for that.
For 3.8GHz, it's close to boot in Windows just with FSB frequency set to 475 and all the rest on Auto.
The only setting to change to make it fully boot to Windows for me was to set N/B core voltage to 1.73V.
Then you need to tweak the settings to make it OCCT stable.
Yes, lowest I could go was 1.71v at some point.
Hence the fact my N/B has a 50€ dedicated water block now: I never saw the N/B over 40°c load whatever insane voltage I throw at it.
The advantage of your G35 board is that you can go at lower N/B & FSB Termination voltage than me but I bet I can get a stable 4GHz with lower CPU voltage: working on that for fun ;)
Haha, yeah. Just try to do that. :) I guess that if I get another chip that doesn't eat up so much vCore, I can probably get 4GHz stable at under 1.45v... I'd take it even if it was 1.440v, but 1.492v is just a bit too high for me.
Nope, you win ;)
It's not stable at 1.45V and I won't try higher on a 45nm processor.
And I got scared for a minute:
At some point with lower voltage the computer posted with "Speed: 4.01GHz - Processor count: 3" :eek:
I guess one of the cores couldn't handle the overclocking but still... I thought one core died of course.
Hmm... I think it's time for me to get another chip... and see if I can finally get 4GHz stable at 1.40v or so... ;) That would seriously be awesome. :)
I can't believe these things aren't readily hitting 4.0GHz on the newer mobos.
*Shakes fist at 8.0 multi*
It's looking like 3.4-3.6 is the safe range.
Well i like the fact i hit 3.4 a default Vcore, i *suspect* for me to get to 3.6GHz i'll need to really juice up the voltage...no hard evidence until i test such, just a gut feeling.
Newer Mobos like Maximus/Rampage are looking to have BIOS problems when overclocking 45nm chips. Supposedly Asus is working on a fix, according to another users post that he saw it on a dutch forum.
But I am not holding my breath on that, I havent seen anymore info regarding this issue.
Yeah I'm right behind you with my X3350, I'm at 3.36 at default vcore as well. I'm just hoping my mobo doesn't give wuss out while i'm trying to get past 3.6. One thing i haven't really seen much is people OC'ing these on P35 boards, course i could have missed all those posts...
Not sure if its going to help much but I am waiting for the first final revision to the bios for my motherboard. I am running 4x2gb ram and it may be causing some stability issues. More to follow....
IP35-E & Q9450 @ 3.48GHz (435 x 8)
vcore = 1.32
vtt = 1.39
NB = 1.53
It needs more vtt volts, but this is all the board has to offer.
I have raised the CPU vcore and VB vcore, but neither got me any more mhz.:down:
Thinks....new mobo or new CPU:shrug:
I have here a motherboard that can hit 4.00GHz with stock PLL and stock VTT (FSB termination volt) if you are interested in getting one...
But seriously, you need a new motherboard for this chip.
Just noticed that post now sorry. I have to post as my goal with such overclocking is not the same than for everybody.
"To get 2 years+" with the current Bios we get, I would stay at 3.6GHz for sure.
I only plan to keep that processor & motherboard for one year. They both can die if Nehalem motherboards, processors and cheaper DDR3 is out: I won't mind at all.
I hope we will get an improved Bios for X48 chipset where we can lower the VTT voltage but till then, I had to make that more clear.
It is true that applying over 1.45V to a CPU or VTT with 45nm technology will reduce the lifetime of your system.
Edit: "PLL" is a different story. The lowest manual setting in the Bios is 1.50V. It is suggested not to go above 1.60V for that setting.
hello,
So i got my evga 790i running with an oem q9450 with a Scythe ninja b heatsink/fan combo, with corsair ddr3 1333mhz ram.
these are my bios settings:
cpu fsb: 1800
memory: 1333
cpu core: 1.3875 (1.336 in cpu-z, lower under load)
cpu fsb: 1.35
memory: 1.675
nforce spp: 1.45
nforce mcp: 1.65
I can hit around 3.7 using this setup, but my problem are my temperatures. Idle temps are about 50-55c, under load, even worse 77c maximum, 75c average. comparing to what ive been reading on some forums, this seems unusually high. When i set to stock speeds, i idle at around 47c and max out at 60c.
75c is too high isnt it? could something be wrong with my ninja b? it doesnt get cold much, but i have tried setting it up different ways which usually results in the same temperatures (i used to run it stock, then i added a scurk1 bracket and i didnt install the copper spacer, then finally im running with the scurk1 with the copper spacer, temps have usually been the same using all these configs)
are my voltages ok? is something too high? or is this simply the overhead from overclocking this chip at this voltage?
is 75c safe (not for 24/7 but for gaming sessions). would it be a good idea to leave my CPU VTT at 1.4, and try to lower my cpu voltage, would this work and improve my cooling (although i always thought 1.4 was too high)?
thanks for any help guys i appreciate it
For 45nm processors use RealTemp if you don't already: http://www.xtremesystems.org/forums/...d.php?t=179044
Your voltages are ok and 1.4V (<=1.45) is safe for the VTT so yes if you really go up to 75°C for the CPU even with RealTemp it's better to raise the VTT.
(You will kill your processor faster with a 75°c temp than with a 1.45V VTT).
Any new batches or revisions for the Q9450?...pipe dream, who's got the latest pack date for these chips?
My chip does 3.8ghz with 1.38v 24/7 475fsb, 1.5nb for 790i.
Highest I'm willing to push on my Asus Rampage
Which version are you running? 25.4 and later should spin multiple threads with no issues.
Ahh - that could be why... I have v2414
Looks like I also have an old version of RealTemp too...
I found 25.5 here: http://www.brothersoft.com/prime95-66720.html
but there's no mention of support for Vista 64-bit? Will it be ok?
Ok - I think I understand about calibration using RealTemp.
Has anyone gone through the process, and if so, what adjustment values are you having to use to get accurate readings in RealTemp?
I use Prime 25.6
Yeah - I found that afterwards. Working fine now - cheers :)
After some further tweaking of voltages and GTLs I got this 9450 completely stable at 447x8. Ran OCCT for 2 hours and Prime for 13 hours without any errors.
450x8 would run 3DM06, Cinebench R10, DVD Shrink 3.2 without problems but Prime and OCCT would give errors after 20-30mins.
This was using the following voltages:
Vcore: 1.45V (could probably be lowered a bit more)
Vmem: 2.19V
Vnb: 1.66V
VTT: 1.438V
GTLs: 75 75 80
This was a far as I got with this 9450 before I returned it yesterday.
Unfortunately this CPU was a bit of a dissapointment as it's very limited by FSB and low multi. :down:
If I was spending my own money I think I'd rather go for a E8500 that might hit close to 5GHz with high FSBs.
Now I've got two QX9650s to try out. One week34 ES and one retail week 39! :D
Just got my Q9450. It's stepping is L811A887.
can you post a picture, including the numbers on the PCB.......
Nope. I've only got a camera phone and it can't pick up the codes.
The codes on the PCB are -
2V80713 OR 2V88713
3A0467
I think the 4th digit of the first code is a 0, tho it's possible it's an 8. An 8 would make more sense considering my stepping ends in 887, but the more I look at it the more it appears to be a 0.
Looking good
http://img205.imageshack.us/img205/6964/3800qv5.jpg
1.27v at 3.8Ghz egg... nice!
Just got mine. Will be about two weeks before I can get some hard numbers up, though, waiting on my RAM.
http://i165.photobucket.com/albums/u...d/Q9450005.png
http://i165.photobucket.com/albums/u...d/Q9450006.png
http://i165.photobucket.com/albums/u...d/Q9450007.png
Well, this is very odd... The vdrop on my motherboard seems to be gone... After my mod wore off, I raised the voltage to keep it stable 24/7 (too lazy to redo it.) I just lowered it back down to 1.32V and I seem to be getting real 1.32V on idle and load...
Needed a bit more NB to finally get 3.8ghz stable -
http://img368.imageshack.us/img368/6313/3800iq1.jpg
so 1.30 is a safe 24/7 voltage to last a few years?
I am new to this whole seen and have a question which may seem dumb.
what does vtt stand for. I am lost on most of the acronyms
Ben
i want to get one of these Q9450 but it costs SGD$540+.
Well i just tried my first overclock, and had to up Vcore quite a few steps to get the cpu stable @ 3.2GHz, but I left everything else alone. So there's no chance I might be able to back it off a bit using vtt (or anything else for that matter) ?
Heres a thread with some screen shots/info: http://www.xtremesystems.org/forums/...9&goto=newpost
You can try it and see.
Every chip and motherboard is different.
Yeah - I thought it seemed quite allot!
Thanks guys, I'll give it a go.
Is there anything else that might help me keep a bit of Vcore headroom for a potential 3.4GHz clock?
Not helping much before I start getting into borderline unsafe Vtt values :(
so whats the verdict with these chips
are most good for 3.6 or am i better off with a Q6600
3.6-3.8Ghz seems like the sweet spot on air with a solid mobo.
You think I can make 3.6 / 3.8 with this ?
Q9450
Asus P5e3 Premium
OCZ XTC Platinum 1600mhz
Cooler Master 850watt Modular
TRUE (with 2 Coolermaster 43cfm push/pull)
Antec Nine Hundred (fans on medium.. If it matters :P)
I know every mobo/processor is different.. But I would like to
get an indication. As im getting worried :rolleyes:
#Phreak
Likely 3.6+...yes.