Re: ASUS P5E-VM HDMI over clocking
- From: Paul <nospam@xxxxxxxxxx>
- Date: Sat, 15 Nov 2008 21:06:35 -0500
x@xxxxxxxxx wrote:
High, I can use some help with this. I got an overclock error at boot at pressed F2 to set to default values. I have a Q6600 which the default is 2.4Ghz. My problem is after resetting back to defaults the PC is running 1.6 Ghz.
In the Asus AI Suite it shows the FSB at 1600 mhz 267.26 x 6 with cpu voltage at 1.15
I have used AI Booster to set the CPU freequency to 333 and the mutiplier to 9 and PCIE memory to 100. I am pretty sure I had ran stable at this setting for over a month previously geting 3.0 ghz out of the cpu. But every time i reboot it says OC error and reverts to the above.
I am running XP PRO SP2
Any help with settings for this board is appreciated.
Thanks,
Brian
Check the BIOS for the EIST setting. Disable it, to see full speed
operation all the time.
Your processor has two operating conditions. Intel uses the second
setting for power saving reasons.
266 x 9 = 2400MHz, VID = 1.3 volts
266 x 6 = 1600MHz, VID = 1.15 volts
If EIST is enabled, and the OS has ACPI support, the OS is
able to tell the processor to throttle back when there is no
computing load. The processor throttles back in tiny steps,
between those two sets of operating conditions. So the Vcore
voltage is change dynamically (D-VID), and the multiplier used
(FID), is also changed. That makes a big difference to waste
heat output, when the computer is idle.
If you want to read "2400MHz" constantly in your Windows
utilities, then you can disable EIST. Or, you can load a program
which presents a computing load (Prime95 4 threads), and then
check things with programs such as CPUZ or your Wizard.
The "333 x 9 and instability" could be due to your choices of
Vcore voltage. What I've learned from my modest upgrade,
is Intel enforces the VID settings in hardware, so as far
as the BIOS writing out a value of VID to the regulator,
the processor limits the values to 1.15 to 1.3 volts in your
case. (The codes are digital, but the values are bounded by
Intel factory settings.) I think my particular processor allows
1.35V before no further increases happen. I suspect, when a
person goes into the BIOS and sets the Vcore in there, that
the motherboard may use another mechanism for boosting the
Vcore setting. On my board, I uses the OFFSET pin on the
regulator, and got enough voltage to run at 3.5GHz :-) But
some motherboards integrate a function like that, so it
appears seamless to the user. (For some people it is a
BIOS setting, but for me it took my soldering iron to
fix it.)
This is my best guess as to what it looks like. I don't know,
how they're doing the boost function in your case, because
your user manual shows pretty fine resolution for the voltage
steps. It would be more normal for boosted values to have poorer
resolution or accuracy. (The VID is a digital value, sent on a
parallel set of data lines, from the processor output pins to the
Vcore regulator.) I can think of ways of implementing this
function digitally, but the motherboard makers are quite
price conscious, and wouldn't want to add yet another chip
to the motherboard to do it.
+-------+
clamped | Vcore |
VID (set by BIOS --- Register inside -------------| Reg. |
or ACPI in OS) Processor 1.15-1.30V | |---CPU
| |
motherboard_maker_cheat_using_boost_method -------| Boost |
+-------+
So start by disabling EIST, and see if you get a consistent
result that way. Then use a tiny bit more manual Vcore setting
in the BIOS, to try to get stability at startup. Remember that
the BIOS has to achieve your desired settings in stages, because
when the processor starts, the default values are used initially.
The BIOS has to boost the stuff gradually early in the POST, and
then try and complete POST.
Also note, I'm not an "AI Booster" kind of guy. I like to set
things to "Manual" in the BIOS, and set my overclocking conditions
there. I have a dirt cheap motherboard right now ($70 CDN), and
it has virtually nothing in the BIOS I can use, which is why
I had to hack up the board with my soldering iron. I learned
a bit about the EIST and the processor clamping thing, while
trying to use RMClock, SETFSB, and CrystalCPUID415 to crank it.
The tools were refusing to let my Vcore go over 1.35V. Not the
fault of SETFSB, as it worked flawlessly to set the clocks.
SETFSB takes over, where ClockGen left off.
In terms of the amount of Vcore boost to use, you have to know
whether your processor is 65nm or 45nm technology. For 45nm
chips, don't go over about 1.40V for long term. People have
burned out 45nm processors while overvolting. Since both
you and I are using 65nm chips (Q6600 is 65nm), you can go
higher than that (and I don't know what the upper limit is for 65nm).
My nominal Vcore is 1.35V, and I'm running at 1.45V right now
(using my home-made boost). Mine still isn't "Prime Stable"
at that voltage choice at 3.5GHz. I've stopped working
on my overclock for now, because I really don't need it
at the moment (it would be a waste of electricity, for
virtually no benefit). Someday I'll need it, and it won't
take long to finish tweaking it.
http://processorfinder.intel.com/Details.aspx?sSpec=slacr
Paul
.
- References:
- ASUS P5E-VM HDMI over clocking
- From: x
- ASUS P5E-VM HDMI over clocking
- Prev by Date: Canned Responses
- Next by Date: Re: Top 10 Money Makers
- Previous by thread: ASUS P5E-VM HDMI over clocking
- Next by thread: ie 7 and firlefox only allow 6 charactors on login screens
- Index(es):
Relevant Pages
|
