Retail Q9650 Overclocking Thread

Printable View

Show 100 post(s) from this thread on one page

10-02-2009, 03:10 PM
deachus

Quote:

Originally Posted by OC Nub

You have very little experience with a c2q it seems. Using LinX, my cpu require less vcore to pass than when testing with blend. This was not the case with my e8400. If you go back a few pages and read you will see where several other had the same findings.

You started insulting people in your first post in this thread by stating that we use spi as a means of stability testing. You have shown no respect to the guys who have spent months oc'ing these cpu's and posting results in this thread, in return you will receive the same. You then dictate to us the definition of stability and imply that anything that differs from this is not acceptable to you (like we care). All anyone was trying to say in the beginning before you went off the deep end was that your board, however nice it may be was not designed for overclocking a c2q, that you would have better results on a more appropriate board for the job.

Well said, my thoughts exactly:up:
10-02-2009, 03:15 PM
The Viper

Quote:

Originally Posted by OC Nub

You have very little experience with a c2q it seems. Using LinX, my cpu require less vcore to pass than when testing with blend. This was not the case with my e8400. If you go back a few pages and read you will see where several other had the same findings.

You started insulting people in your first post in this thread by stating that we use spi as a means of stability testing. You have shown no respect to the guys who have spent months oc'ing these cpu's and posting results in this thread, in return you will receive the same. You then dictate to us the definition of stability and imply that anything that differs from this is not acceptable to you (like we care). All anyone was trying to say in the beginning before you went off the deep end was that your board, however nice it may be was not designed for overclocking a c2q, that you would have better results on a more appropriate board for the job.

That sums it up perfectly OC Nub
10-02-2009, 03:16 PM
zsamz_

first off that that 790i gonna need much more voltage than any other board
to even get that 790i to 500fsb is verry hard
but the board wont even last long before it :explode2:
more than an hour of prime is way too much imho just a waste
stop :brick: sozon
when your 790i dies we know u be on a ud3X:rofl:
10-02-2009, 03:17 PM
Zucker2k

Quote:

Originally Posted by zoson

There you two go putting words into my mouth again. At no time did I say you could pass LinX and then have Prime fail.
I've maintained this whole time that LinX is by far a better stress tester than Prime. Prime can pass 24 hours while LinX still fails after a short period of time. I've seen this behavior in the last four CPU's I've owned. I'm quite convinced that people like you bash it because you can't pass it stable. Complete denial of the fact that your OC is not stable.

Now you guys are fronting like a bunch of kids. Clearly I use 200 LinX loops to test my stability, as seen in my screenshots. I asked sergio to prove his OC with 500 loops max mem. If you guys can't see that he's deflecting instead of proving, well I feel sorry for you. If he were truely stable, this conversation would have been over long ago when he posted his screenshot.

I thought XS would be a haven for people who were serious about overclocking, but instead I find another forum full of fanboys defending their false champions.

Right there, is where you're wrong - prime IS NOT small ftts alone. The real deal is large ftts (especially when you're running high clocks with high fsbs and ram around 1200mhz). The reverse is true - you need more juice to pass blend and large ftts than linx. Yeah that 200 passes with your DUAL CORE is nice, BUT THIS IS A QUAD THREAD; and that's why you came here for help because that same board that took your dual to 550 can hardly take a quad to 460 fsb stable! You need to listen and listen well; to play with the big boys in this thread, you'll need to discard that NV mobo and get a p45 board - hell, get a UD3x board and you can come back and thank us later.

To summarize, many here can run linx all year and fail prime blend/large ffts in five minutes, unless they add more voltage; and that's the truth.

Edit: Nub beat me to it.
10-04-2009, 05:17 AM
CryptiK

Got time for a little more testing. I lapped my TRUE Cu as it was not giving good contact, which reduced max temp - this allowed me to lower vcore. Also got an hour blend done, no time for anymore until later in the week perhaps.

Linx max mem / 4203 MHz / 1.296 vcc / 1.26 Vtt / PL8 / 1.32 vNB

http://i720.photobucket.com/albums/w...6vtt132vnb.jpg

1 hour blend / 4203 MHz / 1.296 vcc / 1.26 Vtt / PL8 / 1.336 vNB

http://i720.photobucket.com/albums/w...t134vnbpl8.jpg
10-04-2009, 05:37 AM
andressergio

Quote:

Originally Posted by CryptiK

Got time for a little more testing. I lapped my TRUE Cu as it was not giving good contact, which reduced max temp - this allowed me to lower vcore. Also got an hour blend done, no time for anymore until later in the week perhaps.

Linx max mem / 4203 MHz / 1.296 vcc / 1.26 Vtt / PL8 / 1.32 vNB

http://i720.photobucket.com/albums/w...6vtt132vnb.jpg

1 hour blend / 4203 MHz / 1.296 vcc / 1.26 Vtt / PL8 / 1.336 vNB

http://i720.photobucket.com/albums/w...t134vnbpl8.jpg

Congrats mate :up: well done
10-04-2009, 10:00 AM
-n7-

Nice work CryptiK.

I gave up on 4.2 GHz...i believe i could do it, but going from 4 to 4.2 GHz seems to require a huge jump in CPU & motherboard voltages, so i settled @ 4.1 GHz.

http://img11.imageshack.us/img11/722...vllconp.th.jpg
10-04-2009, 01:23 PM
andressergio

Quote:

Originally Posted by -n7-

Nice work CryptiK.

I gave up on 4.2 GHz...i believe i could do it, but going from 4 to 4.2 GHz seems to require a huge jump in CPU & motherboard voltages, so i settled @ 4.1 GHz.

you got my same batch n7 isnt yours L828A446 vid 1,1875 ?
10-04-2009, 09:02 PM
Davit

Lowered the vcore a tad and doing an overnight prime run... again... My vid is a lil high 1.25. though I have seen you fellas push the 1.25 past 4.5ghz so that is what I am aiming for. Temps seem pretty decent on water. Any ideas on what can get me 4.5? I did a 4.5 run just before this BSOD after 4hrs, upping vcore 1 core dumped after 2hrs... Have been fighting with this thing for 3-4days now to get 4.5ghz. It will do anything up to 4.3 without even sweating... run to 4.4 kicked my arse, and 4.5Ghz has been a pain as well.

http://i195.photobucket.com/albums/z...1979/44Run.jpg
10-04-2009, 09:06 PM
thezfunk

I see you have quite a temp variation between cores as well. I have two cores that have a 5-7 degrees difference from the other two. I am going to try reapplying the thermal paste and see if that helps. It was my first try with this new paste and I might have done it wrong. I lapped my QX6700 and I am not afraid of doing it to this one if it gets me lower temps.

(lapping my QX6700 saved 10 degrees C!)
10-04-2009, 09:53 PM
Davit

When I took off my heatkiller thermal paste was not spread even...
You read my mind, I ordered a lapping kit from frozen cpu yesterday... I think it should help some, cuz no matter what I do and how I apply the paste it does not look right.
Never did actually, even with my Sunbeam CCF.. go figure. Oh and I am using the 24-carat diamond compound. Supposed to knock off couple of degrees of the AS5 scores, but who knows...
10-05-2009, 12:46 AM
Davit

7hrs and change.
http://i195.photobucket.com/albums/z...1979/447hr.jpg

I'll try lowering vcore some more to see what happens I guess.
10-05-2009, 06:16 AM
CryptiK

Quote:

Originally Posted by andressergio

Congrats mate :up: well done

Quote:

Originally Posted by -n7-

Nice work CryptiK.

I gave up on 4.2 GHz...i believe i could do it, but going from 4 to 4.2 GHz seems to require a huge jump in CPU & motherboard voltages, so i settled @ 4.1 GHz.

Thanks, I'm reasonably confident this cpu will go higher, but I won't be using LinX to test stability due to temps. I would probably need a UD3P to do so however.

Quote:

Originally Posted by Davit

When I took off my heatkiller thermal paste was not spread even...
You read my mind, I ordered a lapping kit from frozen cpu yesterday... I think it should help some, cuz no matter what I do and how I apply the paste it does not look right.
Never did actually, even with my Sunbeam CCF.. go figure. Oh and I am using the 24-carat diamond compound. Supposed to knock off couple of degrees of the AS5 scores, but who knows...

You can see in the following 2 pics the contact was originally quite bad. I did a quick lap job on the True (3rd pic) and it has much improved contact and temps. I'll finish it off when I get some finer grade sandpaper.

Congrats on the 4.4 stable, very nice work.

CPU bad contact

http://i720.photobucket.com/albums/w...badcontact.jpg

True Cu bad contact

http://i720.photobucket.com/albums/w...badcontact.jpg

Lapped True Cu

http://i720.photobucket.com/albums/w...ueculapped.jpg
10-05-2009, 07:11 AM
andressergio

Quote:

Originally Posted by CryptiK

Thanks, I'm reasonably confident this cpu will go higher, but I won't be using LinX to test stability due to temps. I would probably need a UD3P to do so however.

no doubt about it Cryptik you will go up and up on a UD3P :up:

now you lapped you are ready to te giga world ;)
10-05-2009, 07:42 AM
Morgoth

Untweaked pi run on mine, I will remount and fiddle with voltage more, see what happens
http://i6.photobucket.com/albums/y21...h2/10640pi.jpg
10-05-2009, 11:04 AM
Davit

Quote:

Originally Posted by CryptiK

You can see in the following 2 pics the contact was originally quite bad. I did a quick lap job on the True (3rd pic) and it has much improved contact and temps. I'll finish it off when I get some finer grade sandpaper.

Congrats on the 4.4 stable, very nice work.

CPU bad contact

http://i720.photobucket.com/albums/w...badcontact.jpg

True Cu bad contact

http://i720.photobucket.com/albums/w...badcontact.jpg

WOW. That is exactly how mine looks... I tried the drop in the middle method, spread even method nothing worked... Nice to know that I am not entirely crazy. I will lap my 9650 asap, see what I can get out of it. Temp drop would be a pleasant bonus, I would even settle for temp stability between the cores lol. I have read mixed reviews about lapping the heat-killer, so I will let it be... Funny don't want to ruin a $90 waterblock, instead I am taking a sandpaper to $349(what I paid for it) CPU....

Thank you for the kind words as well! :up:
There will be more shenanigans to come.
10-05-2009, 11:28 AM
OC Nub

Davit, I have the DQ6 also. My best results were with the following ref's. CPU ref 1-.775, CPU ref 2 default.800, MCH ref-.795. Try setting default vtt 1.20v and then set these ref's, then bump vtt back up to where you need it leaving ref's alone. These same ref settings worked on 3 45nm quads. Going up a few notches on your vtt to around 1.36v combined with better reference voltage settings may help you lower vcore some.
10-05-2009, 11:28 AM
deachus

I was told that the base of the True's are concave so I lapped mine and got much better temps. I have recently retired my Q9650 to my Media Server and have the True running on a 920 D0. Temps are great. Try lapping your True. By the impressions on your CPU and True in looks like it is concave, as it's making contact in the center and not on the sides.
10-05-2009, 12:26 PM
Davit

Quote:

Originally Posted by OC Nub

Davit, I have the DQ6 also. My best results were with the following ref's. CPU ref 1-.775, CPU ref 2 default.800, MCH ref-.795. Try setting default vtt 1.20v and then set these ref's, then bump vtt back up to where you need it leaving ref's alone. These same ref settings worked on 3 45nm quads. Going up a few notches on your vtt to around 1.36v combined with better reference voltage settings may help you lower vcore some.

Oh that is fantastic! I was looking for some input on that, thanks man!:up:
I saw you hit 4.4 on DQ6, that what actually inspired me lol. Did you ever hit 4.5 on that board???

Does this look about right?

CPU Vcore….………......1.250v*........: 1,50(lets say, man can dream lol)
CPU Termination…..... 1.200V*.......: 1,360v
CPU PLL…………………....1.500V*.......: 1,560v
CPU Reference.………….0.760V*.......: 0.775
CPU Reference2………..0.800V*.......: 0.800
MCH/ICH
MCH Core…………….....1.100V...........: 1,300v
MCH Reference….…….0.760V...........; 0.795
MCH/DRAM Ref.…......0.900V...........: Auto
ICH I/O……………….....1.500V............: Auto
10-05-2009, 12:34 PM
OC Nub

Quote:

Originally Posted by Davit

Oh that is fantastic! I was looking for some input on that, thanks man!:up:
I saw you hit 4.4 on DQ6, that what actually inspired me lol. Did you ever hit 4.5 on that board???

Does this look about right?

CPU Vcore….………......1.250v*........: 1,50(lets say, man can dream lol)
CPU Termination…..... 1.200V*.......: 1,360v
CPU PLL…………………....1.500V*.......: 1,560v
CPU Reference.………….0.760V*.......: 0.775
CPU Reference2………..0.800V*.......: 0.800
MCH/ICH
MCH Core…………….....1.100V...........: 1,300v
MCH Reference….…….0.760V...........; 0.795
MCH/DRAM Ref.…......0.900V...........: Auto
ICH I/O……………….....1.500V............: Auto

I was able to hit 4500mhz stable on the q9650, 4400mhz stable on the q9550.
I think you will need more MCH voltage for 4500mhz all depending on the strap you use. I needed 1.4vmch for 1:1-200d and 1.46vmch for 5:6 240b 1200mhz. You may be able to lower vcore more than you think with these ref's but it also depends on the chip. 4400mhz with default ref's is pretty good. Remember that the ref's are linked to vtt so when you set these 3 ref's make sure you default your vtt to 1.20v first, then raise vtt back up.
10-05-2009, 12:40 PM
OC Nub

Also what are your advanced clock control settings? Napalm had suggested using 1000/1000/0/0 so i tried it and that worked best for me also.
10-05-2009, 01:04 PM
Davit

Quote:

Originally Posted by OC Nub

Also what are your advanced clock control settings? Napalm had suggested using 1000/1000/0/0 so i tried it and that worked best for me also.

Yeah that is what I read as well, thank you for the tip.

Here's the setting I have now:

MB Intelligent Tweaker(M.I.T.)
Robust Graphics Booster ..................: [AUTO]
CPU Clock Ratio ..............................: 9
Fine CPU Clock Ratio.........................:
CPU Frequency ...............................: 4.40Ghz(489 x 9)

Clock Chip Control
Standard Clock Control
CPU Host Clock Control...................: [Enabled]
CPU Host Frequency (Mhz) .............: [489]
PCI Express Frequency (Mhz) ..........: [100]
CIA2 - disabled

Advanced Clock Control [Press Enter]
CPU Clock Drive...........................: 1000mv
PCI Express Clock Drive.................: 1000mv
CPU Clock Skew (ps)....................: 0
MCH Clock Skew (ps)....................: 0

DRAM Performance Control
Performance Enhance....................: [STANDARD]
Extreme Memory Profile (X.M.P.)........: Disabled
(G)MCH Frequency Latch.................: [AUTO]
System Memory Multiplier ..............: [2.00D]
Memory Frequency (Mhz) ................: [978]
DRAM Timing Selectable ................: [manual]

Standard Timing Control
CAS Latency Time..........................:5
tRCD ...........................................:5
tRP'.............................................: 5
tRAS............................................:1 5

Advanced Timing Control
tRRD............................................:A uto
tWTR.......................................... :Auto
tWR.............................................:A uto
tRFC............................................:A uto

tRTP...........................................:Au to
Command Rate (CMD) ....................:2

Channel A
Static tRead Value.........................10
tRD Phase0 Adjustment...................:Auto
tRD Phase1 Adjustment...................:Auto
tRD Phase2 Adjustment .................:Auto
tRD Phase3 Adjustment..................:Auto
Trd2rd(Different Rank)....................:Auto
Twr2wr(Different Rank)...................:Auto
Twr2rd(Different Rank)...................:Auto
Trd2wr(Same/Diff Rank)..................:Auto
Dimm1 Clock Skew Control...............:Auto ps
Dimm2 Clock Skew Control...............:Auto ps

Channel B
Static tRead Value.........................:10
tRD Phase0 Adjustment...................:Auto
tRD Phase1 Adjustment...................:Auto
tRD Phase2 Adjustment .................:Auto
tRD Phase3 Adjustment..................:Auto
Trd2rd(Different Rank)....................:Auto
Twr2wr(Different Rank)...................:Auto
Twr2rd(Different Rank)...................:Auto
Trd2wr(Same/Diff Rank)..................:Auto
Dimm1 Clock Skew Control...............:Auto ps
Dimm2 Clock Skew Control...............:Auto ps

Motherboard Voltage Control
Voltage Type.………... Manual
CPU
CPU Vcore….……….......1.250v*........: 1,5250v
CPU Termination…..... 1.200V*.......: 1,300v
CPU PLL…………………....1.500V*.......: 1,560v
CPU Referen.…………....0.760V*.......: Auto
CPU Reference2………,,0.800V*.......: Auto

MCH/ICH
MCH Core…………….....1.100V...........: 1,300v
MCH Reference….…….0.760V...........; Auto
MCH/DRAM Ref.…......0.900V...........: Auto
ICH I/O……………….....1.500V............: Auto

DRAM
DRAM Voltage ……....1.800V............: 2,220V
DRAM Termination .…0.900V............: Auto
Channel A Reference 0.900V............: Auto
Channel B Reference 0.900V............: Auto

Advanced Settings
Limit CPUID Max. to 3 ------------------ : [Disabled]
No-Execute Memory Protect -------------- : [Disabled]
CPU Enhanced Halt (C1E) ---------------- : [Disabled]
C2/C2E State Support ------------------- : [Disabled]
x C4/C4E State Support ----------------- : [Disabled]
CPU Thermal Monitor 2 (TM2) ------------ : [Disabled]
CPU EIST Function --------------------- : [Disabled]
Virtualization Technology -------------- : [Disabled]
10-05-2009, 01:09 PM
Davit

Sorry almost forgot, and pardon the stupid question, to set the default vtt to 1.20 do I just set it to that, save and boot and then reset it to what I need to, or do I just load optimized defaults, save and then adjust what I need?
10-05-2009, 01:09 PM
OC Nub

Quote:

Originally Posted by Davit

Yeah that is what I read as well, thank you for the tip.

Here's the setting I have now:

MB Intelligent Tweaker(M.I.T.)
Robust Graphics Booster ..................: [AUTO]
CPU Clock Ratio ..............................: 9
Fine CPU Clock Ratio.........................:
CPU Frequency ...............................: 4.40Ghz(489 x 9)

Clock Chip Control
Standard Clock Control
CPU Host Clock Control...................: [Enabled]
CPU Host Frequency (Mhz) .............: [489]
PCI Express Frequency (Mhz) ..........: [100]
CIA2 - disabled

Advanced Clock Control [Press Enter]
CPU Clock Drive...........................: 1000mv
PCI Express Clock Drive.................: 1000mv
CPU Clock Skew (ps)....................: 0
MCH Clock Skew (ps)....................: 0

DRAM Performance Control
Performance Enhance....................: [STANDARD]
Extreme Memory Profile (X.M.P.)........: Disabled
(G)MCH Frequency Latch.................: [AUTO]
System Memory Multiplier ..............: [2.00D]
Memory Frequency (Mhz) ................: [978]
DRAM Timing Selectable ................: [manual]

Standard Timing Control
CAS Latency Time..........................:5
tRCD ...........................................:5
tRP'.............................................: 5
tRAS............................................:1 5

Advanced Timing Control
tRRD............................................:A uto
tWTR.......................................... :Auto
tWR.............................................:A uto
tRFC............................................:A uto

tRTP...........................................:Au to
Command Rate (CMD) ....................:2

Channel A
Static tRead Value.........................10
tRD Phase0 Adjustment...................:Auto
tRD Phase1 Adjustment...................:Auto
tRD Phase2 Adjustment .................:Auto
tRD Phase3 Adjustment..................:Auto
Trd2rd(Different Rank)....................:Auto
Twr2wr(Different Rank)...................:Auto
Twr2rd(Different Rank)...................:Auto
Trd2wr(Same/Diff Rank)..................:Auto
Dimm1 Clock Skew Control...............:Auto ps
Dimm2 Clock Skew Control...............:Auto ps

Channel B
Static tRead Value.........................:10
tRD Phase0 Adjustment...................:Auto
tRD Phase1 Adjustment...................:Auto
tRD Phase2 Adjustment .................:Auto
tRD Phase3 Adjustment..................:Auto
Trd2rd(Different Rank)....................:Auto
Twr2wr(Different Rank)...................:Auto
Twr2rd(Different Rank)...................:Auto
Trd2wr(Same/Diff Rank)..................:Auto
Dimm1 Clock Skew Control...............:Auto ps
Dimm2 Clock Skew Control...............:Auto ps

Motherboard Voltage Control
Voltage Type.………... Manual
CPU
CPU Vcore….……….......1.250v*........: 1,5250v
CPU Termination…..... 1.200V*.......: 1,300v
CPU PLL…………………....1.500V*.......: 1,560v
CPU Referen.…………....0.760V*.......: Auto
CPU Reference2………,,0.800V*.......: Auto

MCH/ICH
MCH Core…………….....1.100V...........: 1,300v
MCH Reference….…….0.760V...........; Auto
MCH/DRAM Ref.…......0.900V...........: Auto
ICH I/O……………….....1.500V............: Auto

DRAM
DRAM Voltage ……....1.800V............: 2,220V
DRAM Termination .…0.900V............: Auto
Channel A Reference 0.900V............: Auto
Channel B Reference 0.900V............: Auto

Advanced Settings
Limit CPUID Max. to 3 ------------------ : [Disabled]
No-Execute Memory Protect -------------- : [Disabled]
CPU Enhanced Halt (C1E) ---------------- : [Disabled]
C2/C2E State Support ------------------- : [Disabled]
x C4/C4E State Support ----------------- : [Disabled]
CPU Thermal Monitor 2 (TM2) ------------ : [Disabled]
CPU EIST Function --------------------- : [Disabled]
Virtualization Technology -------------- : [Disabled]

For 8gb of ram you may need a lot more mch. I had only tested using 4gb kit. With 4gb you should be able to lower Static tRead Value to 9, may work on 8gb as well but am not positive.
10-05-2009, 01:14 PM
KRich

I got a lot of inspiration from all you amazing chaps here and am so excited to tell you that I got my dream Q9650 :clap:

http://img97.imageshack.us/img97/325...4400l847cq.png

I have improved watercooling in the works as I just know that this chip has plenty more to give and which I am ready to receive :cool:

4.4GHz at 1.328v actual in CPU-Z feels so good and just rocks - I am so happy :D

Thanks so much to you all for all your continued contributions here and I wanted to give this back :up:

Show 100 post(s) from this thread on one page