I didn't say FSB problems. I'm not having any fsb problems. I had a problem getting the cpu stable. I could dump 1.55volts on the cpu and it wasn't stable with 0311 but with 0501 it works fine at lower voltage.
Printable View
Anyone have a problem with their quadcore running 500mhz on FSB? Is it the boards problem where the POST restarts, or is that the chip? I cant seem to tell, it might be too much load on the chipset and everything. Anyone know if this sounds like a board limitation or FSB wall?
What voltages are people using to get 500+ FSB on a dual core for the following:
PLL
Northbridge
FSB Termination
Southbridge
Thanks
Heres what I ran on my Xeon 3060 (E6600 Equivalent)
PLL - 1.6-1.7v
NB - 1.55v
FSB Term - 1.4v
SB - Auto, but now for quad I leave it on 1.2v
Also set your NB to .67x or .61x depending on how much voltage your chip is eating, seems to of helped out my OC on my quad.
How are the actual voltages (measured reliably with a DMM
or something) the P5K Deluxe compared to the
settings in the BIOS?
Does VCORE actually match what's expected from the
BIOS settings for various versions of BIOS and
various VCORE values from 1.1 to 1.5x?
Does it droop a lot under load?
I notice that ASUS redesigned the RAM voltage regulation
on the P5K PREMIUM vs. the P5K deluxe and added several
more MOSFETS to their new board. Does the P5K DLX
generate a lot of instability / droop for RAM voltage
and so they found the need to change it radically?
I've heard lots of stories about many (most?) other ASUS
P5**** model boards having BIOS bugs or design
quirks where the actual VCORE is pretty different than
expected from the settings in BIOS.
I've also heard that their "x-phase" VCORE power on some
other models boards doesn't even do ANYTHING
unless you set the various VCORE options a certain way in
the BIOS AND tell it to give a 200mV Vboost.
Any reports on the P5K's voltages and power stability
under heavy load (Q6600 OC'd, 4 double sided RAM sticks) would be welcome, thanks in advance!
Is the stock ASUS P5K Deluxe cooling considered
trustworthy, or is one WELL advised to change it?
I remember hearing some comments by people about
things like NB / SB / MOS heatsinks with installations
that aren't actually even touching some of the chips
(e.g. the multiple FETs in a line), or with super-thick
gunky thermal tape so much that the overall cooling
is not so good, etc.
I also saw one article (which may have been biased)
that showed the back of the P5K PCB in the section near
the MOSFETs and CPU being VERY hot in the region of
100C under normal conditions.
I don't even know how well their heatpipes work in
practice whether the motherboard is vertical or horizontal
and whether there are mechanical/contact problems
with the NB/SB/FETS so much that it might not be doing
its job without reinstallation or replacement.
So I end up with some doubts about whether the
overall setup is "very pretty and theoretically impressive"
with the heatpipes, heatsinks, etc but maybe very
unreliable in the real world if the mechanical contact
and cooling is heat pipe action is not good.
Because of difficulty, risk, and not wanting to make things worse,
I would not want to tear off the existing heatsinks and
heat pipes if I can at all help doing it and they are
doing the job, but if they are actually blocking good
convection for the parts they supposedly cool I'd reinstall
them or replace them as needed/possible.
But probably wouldn't tear it ALL apart just for the heck
of it to replace their thermal tape with AS5 if there's no
well known major benefit to cooling of a large amount.
What can one do to cool the underside of the CPU area
of the PCB where there are bottom-mounted FETs and
the close confinement of the back side of the case
when you have no waterblock / heatsink designed for
that area, and your case hasn't got a back-side
fan opening? Any clever and cost effective ideas / products?
the stock cooling is OK better than the 680i motherboards. I'm watercooling my NB though. the motherboard temp sensor is at the top of the board so if you really want to know the temp of the board use your finger or get a temp probe and put it underneath the NB heatsink.
Jusrt got the P5K Premium to replace the P5K Deluxe that I erazed the EEPROM on with Asus flash. I booted straight up on delivery BIOS 0102 with 445x9 on 1.55v and memory at 1110 on air. I couldn't do that on the deluxe with the same CPU and memory so first impressions are that the different capacitors appear to make a difference. I will play with it a bit more and see what it takes to run orthos stable. Just wish I had more faith in the temp readings
If you want to remove and replace the thermal pads on the deluxe SB/NB with AS5 use the end of any click style pen barrel to compress the heatsink pins under the board. Its way better then trying to use long nose pliers as the end of the barrel fit neatly over the pins compressing them enough to allow easy release. That way you wont break any pins. You can also stick a 40mm fan on the NB to assist cooling . I did it on mine before I screwed the BIOS - I didnt bother with the mofsets. If you want to change the mofset on the side u could use a thermalright HR-09 mofset cooler. Check it will fit first. Overall changing the paste does help reduce MB temps and gives you confidence that the heatsinks are properly mounted with decent paste applied.
Never got to a P5K Deluxe, but am benching a P5K-Premium right now with the same X3060 cpu I ran in my P5B, and I can run SuperPi at quite a bit higher fsb in the new board.
Can't wait for the GO Q6600's to arrive here (New Zealand btw)
Anyone have resume issues? When my monitor goes into sleep mode, the computer will refuse to wake up. Anyone have a solution for this thing?
I've had that problem on a few different computers
in the past. I don't know if there's ever a totally
consistent solution.
One thing to help debug it is to make sure your
ACPI settings in Windows (assuming you use it)
and the BIOS are set to ONLY sleep the monitor
at that time, and not ALSO the HDDs, PC in
general, etc. Also change your screen saver to
something really basic like "none" or "starfield",
because sometimes it can be the screen saver
that crashes and not the monitor sleep itself.
Update your video drivers to the latest stable
thing, and look in their release notes for any
warnings about this issue.
Also set in the BIOS the ACPI monitor settings
for something like "DPMS enabled monitor" as
true (assuming it's so), and "All modes -> Off"
for the monitor power savings type of action.
Then in the BIOS power management settings
be sure to enable wake-up on several different
events like PS/2 keyboard, PS/2 mouse,
USB keyboard, USB mouse, etc.
Also try reprogramming your "power off" button
as functioning as a "shutdown" or "suspend"
function as opposed to "power off immediately"
or "power off in 4 seconds" because when all
else (keyboard, mouse) fails to wake a
PC, sometimes a short tap on the soft-off case
power button will enable a graceful shutdown
when it's so configured.
Also enable the BIOS setting to
do something like re-post VGA BIOS or
issue VGA interrupt on resume from sleep
state. It's called something like that, I forget
the exact terminology used. Sometimes that's
necessary for video cards to resume working
properly when they come back from sleep.
Also try unplugging the USB KB/Mouse from
whatever hub or port they may be on and
(re)plug then directly into one of the motherboard
ports and see if they reset and help you wake
up from sleep state that way.
Other times physically powering off the monitor
and reapplying power to it can help wake up
the video card it's attached it.
Be careful of using things like DVI->VGA
dongle adapters too since sometimes they can
interfere with the communication of the DDC
link with the monitor that helps the PC manage
the monitor's power settings etc. Use the dongle
that's the latest one that matches your specific
video card or try a new VGA/DVI cable etc.
Also if you have a spare DPMS enabled monitor
then try hooking it up and see if the problem
persists with that one.
The only BIOS difference I can see from the deluxe is that CPU PLL offers voltage increments upto 2.2v. Presume that is for people with chilled cooling to make use of it. Overall seems a little bit more stable than deluxe.
From conversation at P5WDH thead about P5K DLX soldered BIOS chip
http://www.xtremesystems.org/forums/...postcount=5463
Hey guys, Since I get a thermalright ultra-120 Extr, Soon , I have to sacrifice 1 Side panel fan @ my CM stacker 830evo, so there wont be air blowing @ the NB area, can I run NBv. @ 1.55v for 24/7 ? I got like 2 fans @ intake front, 2 side intake/ right side of the sidepanel both( RAM/ HDD area). 1 side exhaust @ the lower left ( GFX area). 1 Rear Exhaust, 1 Top exhaust.
[QUOTE=bichi;2345031]From conversation at P5WDH thead about P5K DLX soldered BIOS chip
[url]http://www.xtremesystems.org/forums/showpost.php? QUOTE]
I notice that this is also refered to on the first post of this thread. It also mentions the following re Crash Free 3: -
'But it has Crash Free 3 bios feature see page 4-7 of manual. Shove the included driver dvd disk the system and boot it up and it should attempt to do a recovery bios flash.'
Please note that Crash Free 3 will not work if the BIOS flash has erased the EEPROM - your board will not post - end of story!
What is actually the best BIOS atm, I am still on 0311? I Try to get my E6420 higher , for 1.475vcore ( bios)/ windows :1.446 load =1.440/ CPU PLL volts: 1.5 /NBv 1.4 FSBv 1.3/ the rest on auto, It is stable. However for 3.5ghz prime would do @ 1.5125 however to get pass 3dmark05 cputest ( 3dmark06 cputest it can run all day). I need up to 1.5325 vcore @ bios. Also another note for 3.4ghz I needed 1.4625 to get 19hours+ primestable, however it would crash during 3dmark05 cputest first loop, so I find out 3dmark05 is a good test. But what I was wondering for 438 fsb= 3.5 / 450 fsb 3.6 , I try to achieve those clocks with reasonable volts, what do i need to change or different BIOS , or my chip just plainly cant got futher, I see @ different forums people with the same PFO: as mine getting 3.5 /3.6ghz within 1.5-1.55vcore , so thats makes me wonder maybe I got wrong settings or should try new bios.....? Btw My temps arent of a concern. 3.4ghz @1.4725vcore 44-45 with peaks to 47 during prime 9hours+ with coretemp..... Any help?
can someone tell me if the Sata problem has been fixed, the system hangs when trying to detect my raptor and samsung drive on sata ports 1 and 2 but runs fine on ports 5 & 6.
Thanks.
Well one thing I have found for my 6750 is it only takes 1.45 vcore to do 475 x 8 3.8Ghz and will boot windows at 490 x 8 with 1.5 vcore but orthos's failed after 3 hrs. I don't know if it is worth it to me to go over 1.5v actual just to gain 100-200Mhz. I would like to get 500 x 8 4Ghz stable but I am very happy with 3.8Ghz because I was only expecting 3.6......:)
Forgive the somewhat trivial question, but it's not
covered in the manual and I can't play with the
BIOS since I'm waiting for my Q6600.
At a given CPU multiplier, is there a way to make
the ram frequency SLOWER than 1:1 over the base
system clock?
I.e. at 1:1 you'd have 266 MHz for both base CPU clock
and also actual clock frequency input to the RAM.
The CPU would run (say) at 9x multi 9x266 = 2400,
and the RAM being DDR2 would run
266x2 rate internally = DDR2-533.
However if my CPU multiplier is locked at x9 maximum,
I'd like to be able to clock the CPU around 430x9,
but if the RAM can't do 430x2=DDR2-860, I'd want
the RAM to run at say DDR2-667 or DDR2-533 so
I'd need to use a RAM:FSB divider of something less than
1:1 which may not be documented as supported
according to the manual (which isn't very clear anyway).
I've heard about "hidden multipliers" in some of these
chipsets to give more flexibility thay the 4 or 5
"typical" choices they talk about in the manual.
What can I do with those on the P5K Deluxe / P35?
Can I actually get a good CPU overclock and still re-use
old slow DDR2-533 RAM by use of such multipliers?
If I change the timing from say 5-5-5-7 to something
slower like 8-8-8-10 or whatever would that usually
permit me to operate the RAM at MUCH higher
frequencies than it's designed for even with slower
access times? e.g DDR2-533 RAM at 450 MHz input?