Thanks! Looking forward to 4600.
Printable View
Thanks! Looking forward to 4600.
Congrats Nkresho, that's a great cpu and you've done well with it.
I happened to pick up a Q9650 today as I've wanted to see what they're for quite some time now. None of my s775 boards are really up to the task of quad overclocking, but I threw it into my REX. I used a TRUE copper with 120 x 38mm scythe fan at 1700rpm, and I have to say, these chips run very hot, the TRUE cant control it.
I ran prime for a little over 2 hours but forgot to get a SS at that point so only have 1 hour SS. I ran a quick 10x Linx but only realized I hadn't selected max mem until afterward.
4005 MHz @ 1.232v (LCD Poster and DMM read ~1.217v) / 1.219 Vtt / PL8 @ 1.43 vNB / 712 MHz 7-7-7-20 1T
http://i720.photobucket.com/albums/w...LCDposterP.jpg
4005 MHz @ 1.224v (LCD Poster and DMM read ~1.204v) / 1.19Vtt / PL8 @ 1.41vNB / 712 MHz 7-8-7-20 1T
http://i720.photobucket.com/albums/w...4v119VttPL.jpg
^ thanks. I'll take it further next week when my new board comes. (had to replace ud3r since i'm a goober)
Wow that is warm for that clockspeed. As others have said, new, low vid chips need water I guess.
I don't even try linpack over 4300mhz since it takes me well into the upper 70's. Maybe when it gets really cold here and I can open a window.
Yeahh !!! wellcome to the Q9650's world Cryptik :up:Quote:
Originally Posted by CryptiK
Congrats
Sergio
CryptiK the true is ok for 4.1
you be amazed how much better a 26$ coolmaster 212 is
1 fan not enough for either cooler you need push pull
Nice, hope the new board is a good one. Yeah its a warm 'new' chip (L847C176) and going higher with clocks and vcore easily pushes temps into upper 70's. I'd need water to go much higher.Quote:
Originally Posted by nkresho
Yours being 1.15 VID and running upper 70's on water @ 4300+ is indicative of how hot these things really are.
Thanks Sergio.Quote:
Originally Posted by andressergio
At 4.13GHz / 1.256v it hit 80*C so I stopped the test. I will have to try and add another fan next time.Quote:
Originally Posted by zsamz_
I have a 48C153, 1.23vid, and it def does not run hot. I have pushed 1.38V into it and temps still go no higher then 65C when doing P95 smallffts. This chip runs cooler then my old E8600. MY CPU fan is set up to push only. Of course my chip doesnt seem to clock as high as some of u guys.
man thats hotQuote:
Originally Posted by CryptiK
i didnt hit 68c @ 4.25 with rex and i had it @1.36v
This morning I was testing a 9550 at 4G with 1.268v on water and after an hour of prime small temps where in the mid 70s, no problems though kinda toasty for me.
Congrats :up: Very good vcore.Quote:
Originally Posted by CryptiK
omg why are all Q9650's so good my Q9550 just sucks doesnt go over 4GHz... also my vid is not as good as it should be my min vid is 1.1 and my max is 1.275 -.- anny tips here to get it over 4GHz..?
Your gonna hear alot of people say that your OC is more about what mobo you have versus how good your CPU is. Many people agree that the UD3P/UD3R really make quads "shine" in OCing. With a lower multi chip like the 9550, mobo selection is that much more important.Quote:
Originally Posted by MrDucky
My mobo is a DFI Lanparty DK P45-T2RS+ i know it isnt that good for quads but there must be more in it cause my cpu's vid is 1.275 so i think my cpu is the bottleneck and not the bord. so maybe im gonna switch with my brothers Q9550 he has a vid of 1.150
in such a low vid mate you need the UD3P, the DFI will limit you againQuote:
Originally Posted by MrDucky
Cheers :up:
Sergio
I don't think so;)Quote:
Originally Posted by MrDucky
http://img15.imageshack.us/img15/5243/9550515lp.png
Nice clocks on that 9550. That's a pretty insane vcore heh.
***********************
After some more testing, it became pretty clear that my REX was not suitable to clock a quad with, so I am now using my P5Q Deluxe. Seems PL 8 @ 467 FSB on the 4:5 divider (DDR2-1167 MHz) is stable, and requires 1.36 - 1.40 vNB (haven't tried to test minimum yet, but 1.34vNB was not stable, at least using default NB GTL Ref as I'm yet to tweak this). It's cooler today and I have the 2000rpm scythe at full speed so temps are a little better but its still very hot and it's limiting the CPU's max speed.
I will post some results if I get anything decent stable, I wont have anything near as good as some of you guys with your low VID chips and watercooling though.
Well i dont have that cpu yet i still have my own one with vid of 1.275 -.-.Quote:
Originally Posted by andressergio
I realy want the cpu of my brother cause he isn't a OCer he will use 4GHz and thats easy with my cpu at 1.296 volts on my bord and he has a Gigabyte GA-X48-DQ6 so that wouldn't be a problem. and i dont want to blow another 150 euros for a new motherbord since now you have core i 5 core i 7 and soon core i 9 so i think its a little bit useless to upgrade my mobo for lets say 500MHz higher clock. First im going to test some more with this CPU en hopefully this weekend i have the CPU of my brother :)
@Loco23
Pretty High VCore for that clock :P.
But ok anny tips are welcome :)
EDIT://
My Bios Settings
Genie BIOS Setting
Exit Setup shutdown........................... - Mode 2
Shutdown after AC loss........................ - Enabled
AC Shutdown free.............................. - Enabled
O.C. Fail retry Counter....................... - 1
O.C. Fail CMOS Reload......................... - Disabled
CPU Clock Ratio............................... - 8.5X
CPU N/2 Ratio................................. - Enabled
Target CPU Clock.............................. - 4124 MHz
CPU Clock..................................... - 485 MHz
Boot Up Clock................................. - Auto
CPU Clock Amplitude........................... - 800mV
CPU Clock0 Skew............................... - 100ps
CPU Clock1 Skew............................... - 200ps
DRAM Speed.................................... - 333/667
Target DRAM Speed............................. - DDR2 961MHz
PCIE Clock.................................... - 100 MHz
CPU Spread Spectrum........................... - Disabled
PCIE Spread Spectrum.......................... - Disabled
CPU Feature
Thermal Management Control.................... - Enabled
PPM (EIST)Mode................................ - Disabled
Limit CPUID MaxVal............................ - Disabled
C1E Function.................................. - Disabled
Execute Disable Bit........................... - Disabled
Core Multi-Processing......................... - Enabled
DRAM Timing
Enhance Data Transmitting..................... - Auto
Enhanced Addressing........................... -Auto
T2 Dispatch................................... - Disabled
Flex Memory Mode.............................. - Auto
CAS Latency Time (tCL)........................ - 5
RAS# to CAS# Delay (tRCD)..................... - 5
RAS# Precharge (tRP).......................... - 5
Precharge Delay (tRAS)........................ - 18
All Precharge to ACT.......................... - Auto
REF to ACT Delay (tRFC)....................... Auto
Performance LVL (Read Delay).................. Auto
Write to PRE Delay (tWR)...................... - Auto
Rank Write to Read (tWTR)..................... - Auto
ACT to ACT Delay (tRRD)....................... -Auto
Read to Write Delay (tRDWR)................... - Auto
Ranks Write to Write (tWRWR).................. - Auto
Ranks Write to Read (tWRRD)................... -Auto
Read CAS# Precharge (tRTP).................... - Auto
All PRE to Refresh............................ - Auto
Clock Setting Fine Delay
DLL and RCOMP Settings........................ - By Model
CH1 DRAM Default Skew......................... - Model 3
CH2 DRAM Default Skew......................... - Model 3
RCOMP Setting................................. - Model 1
Voltage Setting
CPU VID Special Add........................... - +250.0mV is 1.328 V idle and 1.296 load.
DRAM Voltage Control.......................... - 2.123
SB Core/CPU PLL Voltage....................... - 1.55V
NB Core Voltage............................... - 1.3750
CPU VTT Voltage............................... - 1.30V
Vcore Droop Control........................... - Disabled
Clockgen Voltage Control...................... - 3.45V
CPU GTL 0/2 REF Volt.......................... - 0.67X
CPU GTL 1/3 REF Volt.......................... - 0.67X
North Bridge GTL REF Volt..................... - 0.63X
FSB Vref...................................... - Auto
No matter how high vcore, result is important;)Quote:
Originally Posted by MrDucky
And it's was Linpack x64, not Prime.
@Loco23
True here a validation of 500FSB x 6 so you can see its not my mobo :P
http://valid.canardpc.com/show_oc.php?id=721724
This is what I've managed so far. As you can see, it's a very hot cpu. Prime small FFT's run at 76*C max at the same settings. To go further I'd need water cooling.
4203 MHz @ 1.312v (cpuz) 1.302v (DMM) / 1.26 Vtt / PL8 @ 1.40 vNB / DDR2-1167 MHz
http://i720.photobucket.com/albums/w...26vttPL814.jpg
Cryptik: The good thing is it appears that your chip doesn't crap out at 65-70c, a lot of 9650's lose stability at anything above that. If possible put it on water and try the giga board. Nice vcore for 4200mhz.
cryptik most important i find are gtl settings on rex or on ud3p
i was fine with quad on rex but only with 2 gigs of ram
ok its just my chip it sucks i cant get it higher so im going to switch this weekend hopefully... :)
You must switch your mobo. Everybody told you about this.
I had 4 processors:Q9450, Q9550 and two Q9650. All of them could work at 500FSB.
Definitely. Applies to the other quad 45nm chips too I think, tested out a Q8400 and it went over 500fsb as easily as the Q9650... :clap:Quote:
Originally Posted by Loco23
it reached 500 FSB but not with a multi 8 or 8.5 it just doesnt work :/ and if the other chip doesnt work i will stop at 4GHz cause im just 15 so i dont have the money for a new mobo, so ican reach 500MHz clock higher..Quote:
Originally Posted by Loco23
Hello fellow P5Q-D user :)Quote:
Originally Posted by CryptiK
Nice results...what GTLs you using for that?
Also, does that do P95 Large stably?
The GTL Refs are set to 0.670x (0/2) and 0.690x (1/3), and NB GTL Ref is on auto at the moment.Quote:
Originally Posted by -n7-
I can't call it stable, as I have only run large FFT's at those settings for around 40 mins. I have not been able to spend much time on testing.
I did run HCI memtest to load the ram and NB and I did get an error but it may be the ram, as I'm using the sticks I benched on DICE at 3.00v+
I'll use some known good sticks and do more testing but I think it may take 1.44 - 1.46vNB for stability.
way too highQuote:
Originally Posted by CryptiK
i find it perfect around 65%
Perfect GTL's differ from cpu to cpu, and also depend on Vtt. Setting them to above settings improved stability for my chip.
So I just upgraded from my sweet 1.2125VID E8600 to a used Q9650 I got from a forum. Turns out the Q9650 is 1.225VID!
First, I'd like to say that... WHOA... Quads need so much more voltage for stability..........:shocked:
http://i144.photobucket.com/albums/r...0mhz-8x500.jpg
On the same board, my E8600 Did:
http://i144.photobucket.com/albums/r...50-200linx.jpg
Anyway, I was stuck below 2000FSB because anandtech suggested no more than 1.36vTT...
I stumbled upon this thread, set my vTT to 1.4v and 2000FSB is mine.
I haven't downtuned my cpu voltage yet, but that's up next after linx verification.
All my stuff is watercooled:
Swiftech Apogee GTZ
EK NB-MAX
Swiftech MCW30
3x HWLabs SR1-120's
Swiftech MCP655b
Swiftech MCRes Micro v2
Enermax Magma Twister Fans
MB: Striker II NSE BIOS 805 790i :)
RAM: 4GB Patriot Extreme Low Latancy 1600MHz (7-7-7-20-31 @ 7-6-6-16-26)
vCore: 1.5125v BIOS 1.38-1.46v HWMon
vTT: 1.4v BIOS 1.44-1.46v HWMon
PLL: 1.6v BIOS 1.60-1.62v HWMon
vNB: 1.72v BIOS 1.76-1.78v HWMon
I think your overvolting your CPU @ 1.392v. You should use a lot less to get stable at 4ghz. Heck I thought mine was high for 4ghz and we both have the same VID. It something else that is not stable.
My latest bench, took me a week to get there WITHOUT Acards ANS9010:
PCMark05 23.7K
Now I only need 689 point to consolidate the Q9650 ranking :up:
Very nice Jor3lBR, good work.
Congrats mate :up:Quote:
Originally Posted by Jor3lBR
Great work as usual Jor3lBR. Congrats :up:
Great work! :shocked:Quote:
Originally Posted by Jor3lBR
I have two OCZ Vertex 120gb SSDs that I can put in RAID 0 with my Q9650, how do you think they would do? I got 12 or 13k with my Q9650 at 4.5ghz and a regular 7200 RPM hard drive.... :(
First of all Jor3lBR nice work realy nice oc :)
Does annyone know's what max safe voltage is for P45 Chip?? cauze if i google i find lots of stuff but all diffrent vallue's. some people say you need more NB voltage aswell when you use X-Fire or extra ram so what's the max safe voltage for p45 chip :P? and what is the max safe vtt :P?
Quote:
Originally Posted by CryptiK
Quote:
Originally Posted by andressergio
Quote:
Originally Posted by OC Nub
Thanks guys! It took me a long time!Quote:
Originally Posted by MrDucky
Regarding max safe voltages that is a subject directly linked to core temperature. If you plan to keep it under 60oC it will last you a lifetime with 1.7vcore. It's very relative.
Thanks teammate (XS HWbot), you will definitely improve your scores as PCMark05 is very dependent on HDD scores like Xp Start Up, HDD General Use and Anti Virus Scan.Quote:
Originally Posted by Linuxfan
I used this formula to improve my scores > HDDERASE each SSD > Create Raid_0 array > Diskpart Align=128 > Vista32 n lighted > the rest only on our team thread :up:
Finally back together guys...
http://img12.imageshack.us/img12/5899/img0240k.jpg
http://img7.imageshack.us/img7/8079/img0239xf.jpg
http://img14.imageshack.us/img14/8382/img0238mj.jpg
http://img12.imageshack.us/img12/2705/img0237eq.jpg
http://img21.imageshack.us/img21/379/img0236wv.jpg
http://img12.imageshack.us/img12/319/img0235fd.jpg
http://img21.imageshack.us/img21/8960/img0234c.jpg
http://img7.imageshack.us/img7/1667/img0233pt.jpg
http://img12.imageshack.us/img12/9315/img0232hi.jpg
Heatkiller backplate facing the right way this time... LOL
Always good to maintain the positive attitude!
I fried 2 UD3P's and almost killed my golden quad with insulation issues.
Now back to your benches :up:
mate you have a very bad bend on the rad at the bottom :eek: its totally collapsed carefull on thatQuote:
Originally Posted by nkresho
Cheers and very nice rig :up:
Oh yeah, good eye.
I put a 45 on the bottom of the gpu before filling the loop because of that. Thanks for spotting it though.
glad to help mate :up:Quote:
Originally Posted by nkresho
Well my P5Q-D wasn't showing me a lot of love, so I dusted off my Maximus II Formula. I'm currently focusing on Large FFT's as its where this board has the most trouble, small FFT's and LinX are easy by comparison.
The vcore and probably Vtt are obviously much higher than I need for the clock, I'm just testing the FSB only. I haven't had much time with it, but this is what I've managed so far.
1 HOUR+ Large FFT / 467 FSB / 1.26 Vtt / PL 8 @ 1.30 vNB / 560 MHz @ 5-5-5-15
http://i720.photobucket.com/albums/w...t-130vNB-P.jpg
So, I got the new board installed and all my clocks from 4200mhz up to 4500 are unstable in prime now. The bios is the same (f9) as are all the settings and hardware (except I am now watercooling the nb).
I thought the second pump might be stretching the limits of my power supply (hx620), but the voltage is sitting stable if I check for droop with my multimeter on the 12v and 5v lines. No change from idle to load, and well within specs.
The board is the same revision as well (1.1) so I am at a loss. I have another board coming back from rma some day, so I will have to test it on that one then.
Any ideas what might be different between the two boards?
Might my gtl values be off?
I noticed a bit more vdrop with the new board today and yesterday as compared to the last one. I'll post them later when I get home, I made a chart for all bios voltages I usedand the corresponding cpuZ voltages I was seeing. This new board is one step lower, as in, I need one bump higher to get one higher value in windows.
I really want to get you guys some screenies of 4600 and beyone, but I have to start from 4200 again.
I hope I didn't hurt my chip with that backplate stunt I pulled, but I have a good feeling it's just that this board just isn't as good as the last one.
^^ Since you're now watercooling your nb, have you considered lowering the voltage a notch or two? I'd play with the vtt too a little. Have you hooked up extra stuff to the power-supply since the switch?
Yeah I really hope nothing damaged your chip, but some of the lines in the back of the mobo might just do that when they are shorten. Very rare to happen but it can.
I always use a multimeter to test if the cpu plate is not grounded accidentally (touching any pins), before I power it on for the first time.
Try what Zucker suggested first and work on your GTL's again, they change from mobo to mobo.
well heres me at 4.7...I basically just threw up the voltages to get there, no real tweaking, I think Im running a PL of 13. My MCH was at 1.6 and I was getting freezes during my Vantage run. I touched my NB and it was burning hot, I think that was causing the freezes. Ill have to find a way to rig a fan to blow on it. The CPU test in Vantage runs fine though, so I think CPU Vcore is adequate.
http://i36.tinypic.com/21b9c42.jpg
Wow, 4.7GHz, 523fsb and <10 1M Pi! Lots of volts :) Did you just get hold of your quad? Where from?
This screenshot is just for you...lower Volts :)Quote:
Originally Posted by In-Fluence
Ebay, hand picked, my 1st quad, 3wks old
http://i33.tinypic.com/243jlhv.jpg
Nice Viper im sure you can go lowerQuote:
Originally Posted by The Viper
this is my best SPI1M
http://img10.imageshack.us/img10/1682/spi20.jpg
Cheers :up:
Sergio
My chip is not as good as yours, but it does run cooler :)Quote:
Originally Posted by andressergio
yes thats why i tell you that you can do better this one is a hell but i achieved this on fresh air comming from outside, very happy with thisQuote:
Originally Posted by The Viper
Cheers and keep the good work i will resend the template :up:
Sergio
Heh Sergio...thanks for motivating me...u were right, I could go lower...but this is the absolute lowest, any lower=No GoQuote:
Originally Posted by andressergio
http://i34.tinypic.com/fbcmqg.jpg
Awesome stuff - I'm definitely getting one now! Depending on wether I can get interest free finance on a Q9650, 8GB of OCz Blades, TC PA120.1 and some other bits & bobs :) If not, it'll just be the memory this month and CPU next :)
And 4.9 GHz?! Crazy speeds! I somehow doubt my MIIF will match that though...
i finaly got the cpu of my brother does anny one had 500FSB x 8.5 on a normal voltage :|?
Nice, hope you get a good chip. I'd be interested to see how your M2F goes with it. Mine's ok for 24/7 clocks (~4 GHz) but its got nothing on these guys with the UD3P's.Quote:
Originally Posted by In-Fluence
Managed to get a 4.8 screenie...still getting freezes though...I think my NB temp is the problem.
http://i33.tinypic.com/2yy5lqt.jpg
Good SP1M result and overclock! Stick some fans on it and give it some more Vcore, thats what I did to get a 4.85ghz SP1M... :up:Quote:
Originally Posted by The Viper