Q9450 OC / Temps / Settings

[Cob]

A lot of X38 boards won't get stable above 450mhz. My ASUS Maximus Formula X38 would only do 450mhz with my Q9450, yet the same chip does 500mhz easy in my P5Q Deluxe, and even did 480x8 Prime95 stable in my Abit IP35 Pro.

[jadeiceman]

Hi there, I have a question: What does it usually mean when the computer freezes after 6 hrs of Prime95 Blend?

[Cob]

Not enough voltage somewhere.

[Rise]

what is a safe temperature to run this chip at under load? (air cooled)

I'm running memtest and my temps are hitting 63C @ stock speeds... should I re-seat my heatsink?

[grenadier]

I'm currently running a Q9450 / P5Q Deluxe / TRUE combo and have been REALLY impressed by the stability I've gotten so far. From everything I read quads looked like finicky critters, but I'm sitting here typing this after 7+ hours of new prime (25.6.1.0) stability with 3.4 (8x425) & 425/4-4-4-12 - everything on stock volts except the memory at 2.1V . I do have a 40mm scythe mini-kaze fan cable tied to the nb heatsink which may be helping. I'm going to try to push it up to 3.6 tomorrow - if I have to increase the voltage at all to get to 3.6 I might just stick with 3.4, slap a 2nd scythe s-flex and turn all my fans down to 7V.

Max real temp temps while priming at 3.4 are 55-54-54-53. I'm using Shin-Etsu X23-7783D and was a bit worried about the application (spotty at best), but these temps look good to me. Could they be lower / should I reseat?

EDIT: Just opened up the case to touch the TRUE / NB heatsink - the eX-120 is much cooler to the touch than my oc'ed E6400 (which maxed at ~45C) & the nb heatsink / heatpipes are barely warm.

[LUCI5R]

Ave,

3.6Ghz (450 X 8) / DDR2-901 (1:1)

6 hrs Prime95 (Small FFT's) Stable & OCCT 2hrs Stable with pretty much all stock volts (including RAM) & 1.325v CPU ...



This is Awesome!! I was @ 2.8Ghz with my E6420/P5W/DDR2-667 setup .. and this is all out-of-box on stock volts. Stability is insane with the P45 chipset (P5Q Deluxe) and these 45nm Quads are ridiculous!!!

[Nano74]

that mother is me next one, preety nice mother, moder, not expensive at all, stable.
Any body tested the las bios 0704 with q9450?? any difference??

[The Nemesis]

Ave,

3.6Ghz (450 X 8) / DDR2-901 (1:1)

6 hrs Prime95 (Small FFT's) Stable & OCCT 2hrs Stable with pretty much all stock volts (including RAM) & 1.325v CPU ...



This is Awesome!! I was @ 2.8Ghz with my E6420/P5W/DDR2-667 setup .. and this is all out-of-box on stock volts. Stability is insane with the P45 chipset (P5Q Deluxe) and these 45nm Quads are ridiculous!!!

I take it your no longer disappointed

[rdv]

This is my first post and any advise will be appreciated. Through the information in this thread and elsewhere I have been able to reach stability (24hr Prime95 blended) Q9450@3.7Ghz.

What I am concerned about is that the vPLL BIOS was set to AUTO and the BIOS says that the vPLL is actually 1.648v. I am now trying to lower the vPLL to the guideline of <=1.6v. Unfortunately none of the usual utilities (CPU-Z, Speedfan or ProbeII) tell me what vPLL is under load.

A test run failed after 7hrs with a BIOS setting of 1.52v with BIOS real of 1.616v. I just started another test run with a BIOS setting of 1.54v with BIOS real of 1.632v.

My bottom line is a 24/7 +3yr rig so with that in mind am I taking too much of a risk being +0.032v over the guideline target values? Also is there any other settings I can adjust so that I can lower the PLL voltage?

I am running a higher vNB than required but in the safe zone (BIOS 1.51v real 1.54v) and it is my next target to lower. Under load all of the temperatures are within reason especially if I can lower the NB voltage.

Other real under load voltages: Vcore-1.336 to 1.344, FSB Term (VCC): 1.28 to 1.30, vDRAM:2.16 to 2.14, vNB:1.52-1.55, vSB:1.07-1.09 All of these ranges are seen during a 24hr prime95 test session.

Assuming that I can achieve stability with 24/7 safety I will post all the gory details, I am just not finished tuning yet.

Thanks in advance.

EDIT: The test run with vPLL BIOS setting of 1.54v with BIOS real of 1.632v failed at 20hr 7min. The next BIOS higher setting is 1.56v with real 1.648v. The real 1.648v is the same real setting if you use a BIOS of "Auto" and at the AUTO setting I had a successful +24hr test run.

So I revise my question to be: Am I taking too much of a risk being +0.048v over the guideline target values? Also is there any other settings I can adjust so that I can lower the PLL voltage?

[LUCI5R]

I take it your no longer disappointed

Not in the least!!
I haven't gotten to actual encoding yet, but I started a sample and before I even thought about doing something else while the sample was encoding - it was done! I couldn't believe it.

System is blazing fast - like I've never seen before.

[Cob]

[REVHEAD]

Originally Posted by LUCI5R
Ave,

3.6Ghz (450 X 8) / DDR2-901 (1:1)

6 hrs Prime95 (Small FFT's) Stable & OCCT 2hrs Stable with pretty much all stock volts (including RAM) & 1.325v CPU ...

What strap are you using for 1 to 1

[LUCI5R]

What strap are you using for 1 to 1

CPU Ratio Setting: 8.0
FSB Strap to North Bridge: AUTO
FSB Frequency: 450
PCI-E Frequency: 101
DRAM Frequency: DDR2-901 (1:1)

CAS# Latency: 5
DRAM RAS# to CAS# Delay: 5
DRAM RAS# Precharge: 5
DRAM RAS# Activate to Precharge: 18
RAS# to RAS# Delay : AUTO
Row Refresh Cycle Time: AUTO
Write Recovery Time: AUTO
Read to Precharge Time: AUTO

DRAM Static Read Control: DISABLED
DRAM Read Training: AUTO
MEM. OC Charger: AUTO
AI Clock Twister: AUTO
AI Transaction Booster: AUTO
Common Performance Level [13]

CPU Voltage: 1.325v
CPU GTL Voltage Reference (0/2): AUTO
CPU GTL Voltage Reference (1/3): AUTO
CPU PLL Voltage: 1.5v
FSB Termination Voltage: 1.28v
DRAM Voltage: 1.96v
NB Voltage: 1.20v
NB GTL Reference: AUTO
SBridge Voltage: 1.10v
PCIE SATA Voltage: 1.50v

Load Line Calibration: ENABLED
CPU Clock Skew : AUTO
NB Clock Skew : AUTO

[RRunner]

Cob,

You can post the used Voltages?

thanks!

[Cob]

They won't be much use to you as you've a different chipset

[grenadier]

I'm using g.skill pc2-8000 ram and am currently sitting at 3.4 (8x425) on stock volts with 1:1. If I want to take things up a notch would it be preferable to go with 8x450 or 7x500? Would the higher speed / lower fsb be easier to stabilize or is 500fsb attainable without too much tweaking?

[Cob]

You'll likely find it very hard to run a 7x500 FSB without using high chipset volts. 8x450 should be possible on minimal chipset volts.

[LUCI5R]

Anyone here running Q9450 @ 3666Mhz+ on Air 24/7 Stable?

I'm stuck at:

[Cob]

What's your PLL set to?

[jas420221]

Anyone here running Q9450 @ 3666Mhz+ on Air 24/7 Stable?

I'm stuck at:

I was @ 3.8Ghz stable on air... Now on water though.

[LUCI5R]

What's your PLL set to?

1.58v

These are my current settings:

CPU Ratio Setting: 8.0
FSB Strap to North Bridge: AUTO
FSB Frequency: 450
PCI-E Frequency: 101
DRAM Frequency: DDR2-901 (1:1)

CAS# Latency: 5
DRAM RAS# to CAS# Delay: 5
DRAM RAS# Precharge: 5
DRAM RAS# Activate to Precharge: 18
RAS# to RAS# Delay : AUTO
Row Refresh Cycle Time: AUTO
Write Recovery Time: AUTO
Read to Precharge Time: AUTO

DRAM Static Read Control: DISABLED
DRAM Read Training: AUTO
MEM. OC Charger: AUTO
AI Clock Twister: AUTO
AI Transaction Booster: AUTO

CPU Voltage: 1.3625v
CPU GTL Voltage Reference (0/2): AUTO
CPU GTL Voltage Reference (1/3): AUTO
CPU PLL Voltage: 1.58v
FSB Termination Voltage: 1.32v
DRAM Voltage: 1.98v
NB Voltage: 1.22v
NB GTL Reference: AUTO
SBridge Voltage: 1.20v
PCIE SATA Voltage: 1.50v

Load Line Calibration: DISABLED
CPU Clock Skew : AUTO
NB Clock Skew : AUTO

@jas420221: That's what I'm hoping to achieve. Can't say I won't be happy settling for 3.66Ghz ... but I'd like to 3.8Ghz 24/7 Stable is possible at all.

[rdv]

Anyone here running Q9450 @ 3666Mhz+ on Air 24/7 Stable?

Here is proof of a 3.7Ghz 24hr prime95 stable Q9450 on air. I am currently +16hrs into a stability test run to lower the vNB to 1.49v (real). If that passes the full 24hrs then I will have finalized my 3.7Ghz testing. The proof links where all done with a real vNB of 1.54v. My air components are in my signature. The ambient (room) temperature during this 24 hr run was 22-24C.



I will post more details when my tuning is complete and successful.

[LUCI5R]

Dude, that is awesome! I mean, seriously.
I can't say there is a drastic difference between 3.6Ghz & 3.7Ghz, but the fact that you have 24hr Prime95 stable @ 3.7Ghz and I have the same on 3.6Ghz makes the difference.

I've been trying to get that stability @ 463FSB like crazy - but it's just not working. I almost thought it wasn't possible on Air. Interesting thing is that your temperatures are absolutely identical to mine. So I'm not sure what's bottleneck(ing) my situation.

Any tips for me?

Running it on P5Q Deluxe (P45) / 4GB PC2-8000

These are the essential volts I'm giving:

CPU Voltage: 1.3625v
CPU GTL Voltage Reference (0/2): AUTO
CPU GTL Voltage Reference (1/3): AUTO
CPU PLL Voltage: 1.58v
FSB Termination Voltage: 1.32v
DRAM Voltage: 1.98v
NB Voltage: 1.22v
NB GTL Reference: AUTO
SBridge Voltage: 1.20v
PCIE SATA Voltage: 1.50v

We're on different chipsets/mobo so I'm sure things will be a bit different ... but any input you can give.

Thanks!

Here is proof of a 3.7Ghz 24hr prime95 stable Q9450 on air. I am currently +16hrs into a stability test run to lower the vNB to 1.49v (real). If that passes the full 24hrs then I will have finalized my 3.7Ghz testing. The proof links where all done with a real vNB of 1.54v. My air components are in my signature. The ambient (room) temperature during this 24 hr run was 22-24C.



I will post more details when my tuning is complete and successful.

[jas420221]

More NB volts is my suggestion. You have 4GB of ram and are trying to o/c a quad.

[Cob]

1.58v

How far did you get with the FSB Term and NB voltages on their lowest settings?