Overclock With Q9450

[dark strike1]

I have been tweaking this little chip the past day or so and I have noticed that this chip doesn't need a lot of voltage to get it to 3.6. Here are my current setting for this CPU:

ImageShack - Hosting :: 69429992ng7.jpg

It seems weird though that the voltage could be so low as it runs at an fsb of 450x8(1800) and the temps are only low 40s. I am using the Xigamatek Rifle with MX-2. I ran Prime for about 20 minutes and temps hit high 50s, around 56C-59C. Does this look like a safe setup?

I just ran prime 95 again and one of the threads failed, so I moved the voltage up a notch on my q9450 and ran a 40 minute session in prime95 with no errors but my temps were reaching 65-68C. Is this normal for a 45nm???

I have the back plate accessory for my Xigamatek rifle, so I know the heat sink is seated cause the screws are as tight as they will go. I am just trying to get a stable 3.6Ghz so I can run my 1800 ram in sync 1:2 instead of going unsynced and having my fsb run at 1600 while my ram is running at 1800(900Mhz).

 

[Nagasama]

68*c is close to the upper limit. i wouldnt run mine that hot, no way. it will get away from you pretty quick with it running that high.
mine hits 60* with prime at 1.225V 3.5ghz
what voltage do you have it set too?
also, your PLL voltage, fsb termination (VTT) and NB vcore are more important than the cpu Vcore with these chips. you should be able to get that overclock with only a slight increase, if any. try lowering the cpu voltage back down and raising those 3 another notch.

 

[dark strike1]

Here are my bios setting:

CPU Core: 1.275v
CPU FSB: 1.30v
Memory: 1.9v
nForce SSP: 1.40
nForce MCP: 1.550

GTLVREF Lane0: +30mv
GTLVREF Lane1: -10mv
GTLVREF Lane2: 0mv
GTLVREF Lane3: 0mv

Cpu/Memory
Linked and Synced
FSB: 1800

Memory Timings:
8-7-7-21-2t
tRRD: 7
tRC: 46
tWR: 12
tWTR: 19
tFAW: 35
tREF: 7.7uS

Limit CPUID MaxVal: Disabled
Speedstep: Disabled
CPU Thermal Control: Disabled
C1E Enhanced Halt State: Disabled
Execute Bit: Enabled
Virtualization Technology: Disabled

 

[Nagasama]

i'd raise the cpu fsb (if that is your termination voltage, and i think it is) a little and see if that keeps it from failing.

 

[dark strike1]

Alright I increased the termination voltage, and decreased the core voltage to 1.225 like yours. The cpu is now running around 60-64C instead of the high 60s range. Anything else I can tweak to get lower temps and stable OC?

 

[Nagasama]

60* is what i get under full load, and i have the s1283 cooler from xigmatek. try testing that for a few hours and see if it fails. if it doesn't, you can try lowering the vcore some more.
also, DL coretemp and see what your VID is if you dont know it.

 

[dark strike1]

My VID is 1.1750

 

[Nagasama]

DUDE...holy crap thats frikkin a great VID...mine is 1.225, and im happy to get 3.5 at "stock" voltage. you SHOULD (this is a big probably) be able to run that chip at 3.5-6 at way lower voltage than mine....it may be something else holding you back...

 

[veedubfreak]

Wait wait wait. I thought the 45nm chips had way lower VIDs. Heck my q6600 is only 1.2.

 

[dark strike1]

Alright I did some changes:

Q9450 VID = 1.1750

CPU Core: 1.200v (1.160v CPU-Z)
CPU FSB(VTT): 1.35v
Memory: 1.95v
nForce SSP: 1.40
nForce MCP: 1.550

GTLVREF Lane0: +30mv
GTLVREF Lane1: -10mv
GTLVREF Lane2: 0mv
GTLVREF Lane3: 0mv

Cpu/Memory
Linked and Synced
FSB: 1800

Memory Timings:
8-7-7-21-2t
tRRD: 7
tRC: 46
tWR: 12
tWTR: 19
tFAW: 35
tREF: 7.7uS

Limit CPUID MaxVal: Disabled
Speedstep: Disabled
CPU Thermal Control: Disabled
C1E Enhanced Halt State: Disabled
Execute Bit: Enabled
Virtualization Technology: Disabled


When I run prime95 now thread 2 automatically has an error and stops, then followed by thread 4. Anyone know what may be holding me back?