That's about were my vcore is. Mine is currently 1.3040
Still trying to adjust trace routes with the memory channels etc. Otherwise, i get almost 2rhs prime blend.
Printable View
That's about were my vcore is. Mine is currently 1.3040
Still trying to adjust trace routes with the memory channels etc. Otherwise, i get almost 2rhs prime blend.
After a week of fun in the Extreme, now it's time to apply some gained skills on the UD3P :up:
What I thought it was impossible is so far holding strong on my UD3P. This board and cpu just keep going and going! The hard part is to use the least amount of volts for the cpu and highest mem strap and timings possible.
Noticed I changed voltages from Team Xtreem to G.Skills a bit but that was only because of larger RAM used for testing.
All this in cheap water cooling system.
I try to send as many pics of the entire testing process (beginning/middle/ end) because I noticed many fake SS's around the web of people adjusting voltages and/or cutting blend tests as customs trying to deceive people here...
[IBT&HCI Memtes: EP45-UD3P + Q9650@4500Mhz + 2Gb_DDR2@1334Mhz PL7]
Beginning:
http://i42.tinypic.com/jh4odl.jpg
Middle:
http://i40.tinypic.com/9r69ms.jpg
End + 100% HCI Memtest:
http://i44.tinypic.com/2zxqtt3.jpg
[Prime95 Blend: EP45-UD3P + Q9650@4500Mhz + 4Gb_DDR2@1250Mhz PL8]
http://i41.tinypic.com/14wvkbb.jpg
...
I will post more results and templates during the next week, this was very very very hard to achieve and needed some voodoo magic to work! :up:
Now lets move on to get closer to 5000Mhz stable ;)
...
Quote:
Originally Posted by Jor3lBR
Just amazing, I love to watch for your posts :up:
can't wait to see your templates, i have a susprise comming for the UD3P :surf:
I'll be back to rocking hte UD3P and can't wait, been havign fun with the
BLOODRAGE
Lately, but i think the UD3P will be back in the "IGLOO" this weekend for a little suprise fun.... :shrug:
But keep up the great work, always love to see your expert work :up:
best regards/steve (SV)
http://img.techpowerup.org/090202/svLogo1.png
just so you know, that batch is ace... Mine is from it doing 4.15ghz on 1.24vcore with only 1.15vtt on a 790i ultra... You must be Canadian too... Some folks have ordered to get this batch from overseasQuote:
Originally Posted by Trigger5521
Here's an Everest and quick SuperPi_1M. I'm still trying to find the sweat spot.
wow guys, nice work. How are you able to get 4.0 Ghz with only 1.24 or 1.26v in BIOS.. and if it's 1.24 in bios, it's even less in CPU-Z huh.. thats just not fair... My board sucks. I'm sure I'd do much better with a better mobo, but I don't really need the extra 3 or 400 Mhz.. 4.3 Ghz is doable with my system, but I've yet to get it stable. I can bench with 4.275 Ghz.
Nice work with the 790i ultra there.
oh and to trigger, any Q9650 will be able to hit 4.0 Ghz with a faily decent motherboard. one as cheap as mine does it with no problems. just a lot of Vcore.
I almost want to mail my Q9650 off to one of you guys like Jor3 or someone to see how well it does in there board over mine.
Anyways, I'm posting in here to tell you all that I've finally done it, I got water-cooling. I have a HAF that is currently in my bathroom leak testing with a Apogee GTZ, MCP-655 D5 Vario, Feser TFC360 Rad, and the Micro res, all with Danger Den High flow barbs and Masterkleer tubing. I'm very exited for it. I can't wait to throw this Q9650 under some water and see how well I can do, the only problem is I need 1.312v to be stable at 4.0 Ghz, and thats in CPU-Z.. in Bios It's like 1.34v for 4.0 ghz stable.. but with 1.355 I'm able to bench with 4.25 Ghz. But I think my problem is that my motherboard is limited on the fsb/vtt voltage, I'm only able to raise the vtt voltage to 1.45v. I can't go any higher, and like Jor3, and Napalm said.. the higher vtt/fsb voltage sure helped them. Or it did with napalm. so we'll see what I can do with only able to go that high.
Results coming soon, I'll leak test tonight, and throw system in case tomorrow.
Jor3lBR, was there any advantage going from the F6 to F8a bios?
Quote:
Originally Posted by Trigger5521
Oh for the love of god, is my luck us unreal. I decided to order a chip from my trusted computer shop to see if I could get a B series instead of grabbing one of the L828A446's local to me. Would you believe I got L834B631...same damn batch that PortCitySlim( post) had all stuck senors and worthless OC.Quote:
Originally Posted by 5010
Thank god these guys I bought it from are top notch..he said he would just send it back and get me another one, I didn't even open it.
Most of these chips have stuck sensors, my new chip has stuck 34-34-32-34 sensors. I wonder if yours would have sucked ass like the one I had. I have seen chips from the same batch number act completely opposite of each other.Quote:
Originally Posted by Trigger5521
We'll never know I didn't want to take a chance after what you went through. Thanks for posting that BTW probably saved me a headache.Quote:
Originally Posted by PortCitySlim
Nope, same results.Quote:
Originally Posted by Hoss331
Quote:
Originally Posted by Trigger5521
It's official. All Intel $335 Q9650 are processors that ddin't make the cut to go to the shelf's on the first place. Same thing qith the new Q6600's...
All of them have stuck sensors.
I have 1 stuck sensor at idle but is right in line with the rest at full load. Here's my best 1M score in the little time I've had to play with the new setup. Much more benching to come over the weekend.
*UPDATE*
I was able to get some lower SuperPi 1M and 32M scores and gain some ground with Vantage PhysX off.
For some reason this Board will not run 1:1 at all. I have 2 sets G.Skill 2x1GB 1600Mhz D9GTS that run DDR3-2040Mhz on my P5E3 Premium a set of Mushkin 2x2GB 1600Mhz D9GTS and a set of G.Skill 2x2GB 1600Mhz Samsung HCF0 and none of them will run 1:1 on the P5Q3 Deluxe. Not sure if I'm missing something here. Settings are a little different then my P5E3 Premium.
Q9650
SLB8W
L829A405
VID:1.2375
Pack Date: 1/30/2009
Purchased: Newegg on 2/13/2009
Base line, as low as she can go....for now.
http://img25.imageshack.us/img25/5486/30min.jpg
Code:
Extreme Tweaker
Ai Overclock Tuner : Manual
CPU Ratio Control : 9
FSB Frequency : 433
FSB Strap to North Bridge : 333
PCI-E Frequency: 100
DRAM Frequency: DDR2- 1040
DRAM Command Rate : 2T
DRAM Timing Control: Manual
CAS# Latency : 5
RAS# to CAS# Delay :5
RAS# Precharge :5
RAS# ActivateTime :15
RAS# to RAS# Delay : A
Row Refresh Cycle Time : A
Write Recovery Time : A
Read to Precharge Time : A
Read to Write Delay (S/D) : A
Write to Read Delay (S) : A
Write to Read Delay (D) : A
Read to Read Delay (S) : A
Read to Read Delay (D) : A
Write to Write Delay (S) :A
Write to Write Delay (D) : A
DRAM Static Read Control: DISABLED
Ai Clock Twister : STRONG
Transaction Booster : DISABLED
===Relax Level = 0
CPU Voltage : 1.2375
CPU PLL Voltage : 1.5
North Bridge Voltage : 1.55
DRAM Voltage : 2.0
FSB Termination Voltage : 1.2
South Bridge Voltage : 1.050
Loadline Calibration : ENABLED
CPU GTL Reference : .63x
North Bridge GTL Reference : .67x
DDR2 Channel A REF Voltage : DDR2REF
DDR2 Channel B REF Voltage : DDR2REF
DDR2 Controller REF Voltage : DDR2REF
SB 1.5V Voltage : 1.5
NB LED Selection : NB Volt
SB LED Selection : SB Volt
CPU LED Selection : CPU VOlt
Voltiminder LED :
CPU Spread Spectrum : Disabled
PCIE Spread Spectrum : Disabled
Advanced CPU Configuration
CPU Ratio Control :9
- Ratio CMOS Setting :
C1E Suppport : Disabled
CPU TM Function : Disabled
Vanderpool Technology : Disabled
Execute Disable Bit : Disabled
Max CPUID Value Limit : Disabled
USB Configuration
USB Functions: Enabled
Legacy USB Support : Disabled
check out my new cpu.the two stronger cores really give it a great boost.
http://i44.tinypic.com/66jdb4.jpg
Okay, what's going on there? :)Quote:
Originally Posted by radaja
i have no idea?except that i had an E8500 in then swapped out for the 9650,cleared cmos,loaded setup defaults booted into windows then it said it had drivers for the 9650 and said to restart and then i got this.pretty strange though.
Just delete the processors and then reboot.Quote:
Originally Posted by radaja
Problem should be solved.
yea i got fixed.thanks
Latest on my MFII....12 Hours+ Prime screenshot
CPU GTL's are +10/-40/+10/-40
NB GTL is +60mV
http://i16.photobucket.com/albums/b3...504_3ghz-1.jpg
3Dmark06 - 22051 with single GTX285 FTW, no tweaks
http://service.futuremark.com/compare?3dm06=10092979
I am amazed at how tight this board allows me to run my Mushkin 2x2GB D9GTS. This was not at all possible on my P5E3 Premium. I am running DDR3-1568Mhz @ 6-6-6-18-2T-60~1.98v. These sticks ran DDR3-1600Mhz 7-7-6-18-1T 1.94v on P5E3 Premium. Bandwidth looks really good as well:)
Hey guys
A friend of my is going to buy a Q9650 and he has the EVGA I790 ULTRA SLI, how are this mobos to clock this chips ?
Cheers :up:
Sergio
This Q9650 and GTX 285 are still scaling nicely:D Haven't hit a wall yet on the CPU or VGA. I just finished up some benching, now time for some dinner. I am still amazed at the memory performance with these tight timings but low mem speed. I was able to run this set close to 2000Mhz on my P5E3 Premium but the Everest and SuperPI results were no were close to these:eek: I'm loving this new setup and I believe I got a great Q9650 from Newegg.
I have one, but it needs lots of volts to do 4ghz even at 1.4 vcore in bios it fails prime instantly.
settings are
vcore 1.4
cpu pll 1.50
FSB term 1.36
dram 2v
NB 1.71v
SB 1.1
CPU GTL 0.63
NB GTL 0.67
SB 1.5v voltage 1.50v
this seem pretty high to me.
can get in windows at 4.2ghz with much higher volts but still unstable.
Do I just have a rubbish cpu, board is an X38 P5E which has been rock stable at the 467fsb with my previous Q9300 with much lower vcore than 1.4v
What's the voltage under load (not in the bios?)
That matters.
If you're dropping to like 1.32v under load, there's the problem. You may need to do a vdroop mod or use loadline calibration (or equivalent) to reduce vcore droop.
Running OCCT now, have had to increase the volts in bios to 1.4250v and CPU PLL to 1.56v.
done 5 mins so far with no errors, will see if it pass's the full hour. vcore in this program sais 1.34 though so board undervolts quite abit.
Using loadline calibration in bios.
failed after 8:32
shall I keep on increasing vcore.
temps got to 74c on one core, pretty high considering I am under water.
water consists of 120.3/D-Tek fuzion V1 with quad nozzle, XPSC NB block, Asus 3a mosfet blocks, solid copper SB heatsink. Pump is 18w DDC with flexitop.
vcore listed as 1.38 in OCCT.
testing again 1.43150v in bios, this drops as low 1.34v at some points under load.
Failed after 11:56 on latest run.
Have dropped NB volts down to 1.51v now though.
if you're setting 1.42vcore and getting 1.34 load then it means loadline calibration is NOT enabled.
Its definetly enabled, just checked.Quote:
Originally Posted by Falkentyne
Just stopped OCCT after 30 mins, now brought down my NB volts to 1.45v and VTT to 1.3v from 1.36 see what it does.
All good fun on a Sunday morning this is, lots of work finding the right settings for a new CPU, thats why we love it though.
http://img11.imageshack.us/img11/418...x510na5.th.jpg
enfin la barre fatidique des 500 est passée :up:
Thats a great overclock, dont think I will get anywhere near that, I'm afraid.
But Banny,Quote:
Originally Posted by banny
What is your idle voltage (in cpu-z?)
If your idle voltage is .06 volts HIGHER than your load voltage, then loadline isn't active for some reason. Loadline is supposed to make your idle and load voltages very, very close.
Disable loadline and then compare the two voltages for me.
Wow Dark, 4.59... 4.6 Ghz. thats sweet. I'm able to hit 4.3 Ghz, but I've never tried for higher, I've also never took my CPU over 1.3625v in Bios yet either. so I'm sure I can get some better clocks out of it.
Anyways, I had a question for everyone in here. If anyone has a Q9650 under Water-cooling Could you Please, Please post your idle and load temps for it.. I just put my Q9650 under water for the first time, and so far I "Think" I have good results. My temps are pretty good. But I'm just wondering what others have, I know loops differ, and some may have GPU's in there loops, but post anyways please.
I just threw a GTZ, Feser 360 and a D5 Vario on my Quad. and I'm now idling around 26c at 4.0 Ghz with 1.312v (that is my 24/7 atm, but now I got water :) )
So 4.0 Ghz with 1.312v and idling around 26c, and the important temp, Prime 95 Load temps were 40/41, 37, 38, 38 after 30 minutes of Prime, not bad at all I'd say. But this i why I wanna compare to others.
Thanks.
Oh and here's a peek at the new Loop and case that's cooling my Q9650... I'm exited to enjoy my water loop.
http://i727.photobucket.com/albums/w...leteLoop-1.jpg
Idle voltage is 1.4v, 100% 1.352 this is in CPU-Z, gonna do the pencil mod anyway.
Have 26mins to go on this OCCT have increased bios volts to 1.44375v now, VVT 1.32 and NB 1.41.
Had vcore 1.436v before and failed after 41 mins, so increased one notch and re-testing.
BTW once stable in OCCT can I reduce vcore for everyday gaming
My idle is quite high in the 45c mark, according to OCCT, thats with water, but could be due to the fact I am also cooling NB and mosets in same loop, mind you I do a PA120.3 and 18w DDC pump, maybee need to do the thermal paste again.
Under 100% load max temp has been 77c
Passed an hour at these settings.
Gonna drop PLL down to 1.52 from 1.56 see if it works.
With loadline disabled
Vcore idle 1.40v 100% load 1.36-7v so its the same, seems like loadline doesn't do anything.
Time to get the pencil out then.
wow.. under water your idle is at 45c. something is not right. Thats to high. you should try a reseat. what thermal paste did you use, and how much? NB and mosfets shouldn't make that big a difference. Maybe a GTX 285 would, but not the NB..
Also, your needing 1.436 just to almost pass OCCT with 4.0 Ghz.. thats a very high vcore? I think it's because your x38.. if you got a P45, you'd hit 4.0 Ghz with less than 1.36v unless your chip is really that bad, but I've not heard of any that bad yet. get the P5Q Deluxe, or the Gigabyte UDP3 and you'll get some better clocks.
I agree the P5Q deluxe or the Gig udp3 seem to like quads, I have the P5Q deluxe running Prime again with a 9650 at 4320 at 1.375 bios for almost an hour now with out issue.
did the volt mod, reseated the CPU now it wont start up, just keeps switching on spins the fans then shuts off again, in one loop.
wiped the volt mod off, so probably killed it now.
Now thats overclocking
resetting the CMOS never made any difference
I dont beleive his x38 will be the problem, my Q9650 only needed 1.216v for 4Ghz stable on my maximus formula, my P5Q and UD3P were not stable at such low vcore. Maybe he has a bad chip.
Got it back up and running, I dont believe its the board, as it is not a bad overclocker, had my Q9300 at 480fsb for benching
This really stange now my load temps are only reading 3.5/15/15/12 and getting lower, into single figures now my reseat cant be that good. now 6.5/4/4/2 this was OCCT
realtemp sais its 42/41/43/40
Pretty sure it is seated right, as for paste using Akasa 450 silver stuff, just put a rice grain in middle of cpu and bolted down.
Here's my last SuperPi 1M run last night before I switched gears and started working on Prime Blend stability. This is my best 1M score yet. I still can't get over this P5Q3 Deluxe running my Mushkin 2x2GB D9GTS 16Mhz under DDR3-1600Mhz (1584Mhz) @ 6-6-6-18-2T-Trfc 60 @ 1.94v. Not sure if this is pretty common but def. wasn't possible for me on my other Board and I can't say I see these timings at these speeds frequently.
I was able to run prime for 5 hours so I stopped it, rebooted, went into bios and dropped vCore, Gtl's, Vnb, vTT vDimm each one notch and started prime blend again.
Just woke up this morning and everything was still running great Prime95 Blend 8 hours stable. I can achieve Prime stability at higher FSB but I was looking for a nice speed while keeping temps under 60C on air.
Nice,:cool: I am still running Prime right now as well at 4320, I will post up the results later today.
good to see everyone is getting nice results.
banny...what is the vid of your cpu ??
We just had a major power blackout here and I just lost over three hours of Prime.:mad: I am starting over again and should have something good by this afternoon as 1.375v seems to hold 4320mhz no problem with Prime95.
1.2500vQuote:
Originally Posted by bito
http://valid.canardpc.com/show_oc.php?id=513880
Valid:
http://img23.imageshack.us/img23/3721/4882validjo1.jpg
PiFast:
http://img12.imageshack.us/img12/4198/pfst1955ki7.jpg
WPrime 32M:
http://img9.imageshack.us/img9/6892/wprm8672ug7.jpg
WPrime1024:
http://img24.imageshack.us/img24/8771/wprm292328gk3.jpg
Why did you post a ss of an E8600? :confused:
mistaken, keep cool :)Quote:
Originally Posted by Hoss331
That last screenshot isn't for a q9650, or any other quad core. WTF?
Let's see the quad core at that speed! Godspeed!
"TheGanG" what vCore are you setting in BIOS? CPUZ 1.76v is really high, are those subzero runs?
Yes subzero, it was a max mhz run and juiced high. Expecting 4.9+Quote:
Originally Posted by deachus
VTT, NB and PLL was as low as possible (via GTL Refs/PL)
Good to know. The reason I ask is because I was running -5C ambient last night 4500Mhz @ 1.44v. I wanted to gauge how much more I have to up the vCore to get 4800Mhz - 5000Mhz. I hope I don't need that much but we will see. I see alot of people using LLC but I do not use it and my droop is practically nothing with it off. 1.440v set in Bios 1.432v CPUZ Next everything is going in my A/C Chiller -30C:)
These settings are the max I have run at my highest achieved FSB
4500Mhz @ 1.44v
vTT 1.40v
GTL 0.68/0.69v
PLL 1.58v
vDimm 2.00v
Performance Level 7
Clock Twister Stronger
DDR3-1584 - 6-6-6-18-2T-60
What was your ambient/surface temps at that run?
Load was -38°C, idle about -58°C. Ambient doesn't play a role with my setup. Actually i don't need so high VCore...Quote:
Originally Posted by deachus
PL:8 may be good for you. I believe u hit 4.7-4.8 GHz if u have a good CPU :) I haven't a good one...
Maybe this will help, for 4.8-4.9 I use about 1.48 volts on air, S1283.Quote:
Originally Posted by deachus
http://i131.photobucket.com/albums/p...apImage4-1.jpg
http://i131.photobucket.com/albums/p...tmapImage4.jpg
Yeah good job friend, i haven't a good chip so it wants fire.
Another reason is DFi LP DK P45-T2RS PLUS (with rev. A2 chipset), big headache with Quads :)
"Hoss331" can we see some Memset shots with that run. What divider, speed, timings, voltage, PL? Thanks.
I dont remember the exact settings for those runs but here is a quick one for you. Nothing fancy.Quote:
Originally Posted by deachus
http://i131.photobucket.com/albums/p...mapImage2v.jpg