View Single Post
  #12  
Old September 9th 14, 12:51 AM posted to alt.comp.hardware.overclocking
Paul
external usenet poster
 
Posts: 13,364
Default Asus P5K Premium with Xeon E5472 and 771 to 775 adapter sticker??

Bob F wrote:
Bob F wrote:
.......
Otherwise, you may have to consider locating a Windows based
overclocker tool. Such tools, work with particular clock generators,
and it's the luck of the draw as to whether your particular
motherboard would work. I think I got lucky once, and managed
to find a tool that would overclock one of my older boards.
It bumps the FSB a megahertz at a time, and it would
take 30 seconds or so to get to the "target" frequency I
specified while sitting in Windows.

And 272MHz is a weird value. Not a canonical value. I wonder
where that is coming from ?

Interestingly, that is the same as the Q6600 that I had in it
previously, and that is back in the board working great now. I can't
imagine any way that value could have stuck through the processor
change/bios update/cmos reset/bios to default sequence.

I haven't gotten as far as windows on the IP35 Pro yet, so a windows
version of CPUID is out on that PC. I've basically just been trying
to get the new processor to run memtest at a reasonable speed. I does
run it at the 272. although the last test I left it on had 10 or so
errors, then crashed. Interestingly, the errors were complete bit
inversions.
One thing that really has me wondering is why I cannot change the
E5472 speed either up or down. This probably is associated with the
double boot, which always sets it to 272. No idea why that processor
makes it always double boot.

Is it possible that the bios is detecting a problem and restoring the
settings to the previous fully working settings of the previous
processor, even after resetting the bios? But then, why can't I
change the speed slightly, even down, from where it does work?

If it resets to the processor default, and doesn't like that, it
resets to the previous working settings? I'd be really impressed if
it was that smart, but then again, it could be dangerous with some
processor swaps.
Maybe I'll try putting the E5472 in after running the Q6600 at 300
instead of 272, and see what happens.


Interesting.

I tested Q6600 on this board for awhile with a 320MHz clock got 2880MHz
processor speed. I then shut it down, swapped in the E5472, reset CMOS, and
booted it to MemTest86+ V5.1. (My previous tests were on 4.1) The E5472 is now
running at a clock of 320MHz, showing a speed of 2400 MHz. So it is remembering
the clock settings of the previous working processor. I wonder if it is
remembering the voltage and other settings of the Q6600 also.

ESCing from Memtest86+, it does soft boot without the double boot , and going
into the BIOS, I see the 320 MHz clock, 7.5 multiplier in the active "User
Defined" settings. The CPU core voltage is shown as 1.2250V and CPU VTT is
1.10V, MCH 1.25V voltage is Auto, ICH 1.05 is 1.05, and ICHIO 1.5V is 1.5V. CPU
temp is 47C, System temp 30C, and PWM 38C.

It does seem to be running Memtest86+ just fine, with no errors. Memtest86+
shows the RAM running at 480 MHz (DDR2-960) - BCLK: 320m timings 5-5-5-18 @
128bit mode. The Bios shows DRAM speed as "Default (DDR2-768)"

Looking at the double boot in detail: Powering the board up with the E5472 after
powering down, and pushing the start button, it boots, displaying the numbers
83, several quickly I can't catch, then 90 and 99, then powers down and
re-boots. The on board display then displays 84,83,84,83,84,82, then hangs
there, with the green on board LED flickering off every half minute or so.
Powering down and back up and pushing the start button, It then starts up
correctly, but at the same settings each time. Number sequences are
abbreviated - some are too fast to catch.

Changing the speed from 320 to 300 in the BIOS and saving the change results in
the standard reboot, but letting it boot to MemTest86+ shows the speed has not
changed from 320.

Changing the speed


Most boards in that situation, not knowing what you've plugged in,
wouldn't attempt to start. They would have an appropriate beep code
and you'd be stuck.

To me, it seems a bit dangerous to be using the previous
settings for the new processor.

What I can't understand, is why your board doesn't apply
canonical values. Clock generator at 100Mhz, 133Mhz,
166MHz, 200MHz, or 200, 266, 333, 400. As selected by
the BSEL straps on the processor. Similarly, the VID
voltage can be specified by the processor, as it drives
out bit values on the VID pins, to set the voltage it
wants. And the board really should not be attempting
to program the clock generator, if it doesn't know the
correct value. It is supposed to rely on the processor
strap values to determine that.

As for your Port 80 codes, some motherboard manuals have
a table of values for those. They're "progress" codes rather
than error codes, so even if you decoded them, you might
not be any further along in understanding what is
going on. The last code on the display should be
"Booting OS", and after that point the BIOS won't be
writing any new values to the display.

The "progress codes" are of more value to the developers,
than to us. I haven't run into a scenario yet, where
knowledge of "progress" helped debug an error
condition.

If you wanted to capture more of the codes, you could point
a movie camera at the display. It's still going to miss
stuff, but at least you don't have to write them down
"on the fly".

Paul