"ASUS P5Q Deluxe discussion".

[jagt]

Basic setup in BIOS is enable the jmicron controller then drive expert.

you say,"Windows hangs while loading". Did the orange drive have windows already installed on it and it hangs when booting from it? or does windows on your other hard drive fail because you enabled the orange slots? Did you install windows on your orange slot drive and now it won't boot?

-=Mark=-

No, I have Windows on a RAID 0 array that runs off the Intel ICH10. Anyway, I got it working by uninstalling the Marvell SATA drivers and let Windows install its own drivers. It's recognised as "Standard Dual Channel PCI IDE controller" now, but both my IDE DVDRW and the SATA HDD that's connected to the orange port works fine now. eSATA however doesn't work anymore, but I guess that's the price to pay.. Unless anyone has any suggestions to what drivers to use on the Marvell controller for this to work?

[Big Lar]

are you guys running the m1406 or the 1406

I personally an running the non mod'd 1406 here. I have not had any issue's with it so I have not run Kets bios yet. Very stable bios IMHO.

Larry

[Marvin]

Asus released a new p5q deluxe bios. Any difference between the 1406 ?
regards

[Leeghoofd]

Bandwith wise regarding my daily clocks not much difference at first sight... voltage wise I can run now with 1.25 instead of 1.256 for the same 4ghz, seems a positive evolution to 1406...

Only weird thing I had : I flashed the bios, reset to defaults, set my OC... first boot it was stock FSB (333), second boot FSB okay but ram in sync and third boot all was like set in the bios...very weird indeed... Linx'ed it , OCCT done and gamed upon it, seems stable...

EDIT : further testing allows me to lower PLL, NB and FSB Term voltages :

Old bios -- New bios
PLL 1.56 --> 1.52
FSB 1.26 --> 1.22
NB 1.26 --> 1.22

I can leave Static Read Control enabled now with clock twister MODERATE (before Light) with the 2GB dimms... This should give a small advantage in bandwith and a slight drop in Latency... but nothing earthshaking More testing later.

Stability testing continues : survived a 4 instance 32Mb Hyperpi already, linX and OCCT tomorrow....

[playground]

Only weird thing I had : I flashed the bios, reset to defaults, set my OC... first boot it was stock FSB (333), second boot FSB okay but ram in sync and third boot all was like set in the bios...

same here =] nothing new in this BIOS

[Big Lar]

Bandwith wise regarding my daily clocks not much difference at first sight... voltage wise I can run now with 1.25 instead of 1.256 for the same 4ghz, seems a positive evolution to 1406...

Only weird thing I had : I flashed the bios, reset to defaults, set my OC... first boot it was stock FSB (333), second boot FSB okay but ram in sync and third boot all was like set in the bios...very weird indeed... Linx'ed it , OCCT done and gamed upon it, seems stable...

EDIT : further testing allows me to lower PLL, NB and FSB Term voltages :

Old bios -- New bios
PLL 1.56 --> 1.52
FSB 1.26 --> 1.22
NB 1.26 --> 1.22

I can leave Static Read Control enabled now with clock twister MODERATE (before Light) with the 2GB dimms... This gives a small advantage in bandwith and a slight drop in Latency... but nothing earthshaking.

Stability testing continues : survived a 4 instance 32Mb Hyperpi already, linX and OCCT tomorrow....

Hello Leeg. Did the ram voltage stay the same as 1406, or has that changed for the better/worse? Thanks.

Larry

[JohnLoph]

EDIT : further testing allows me to lower PLL, NB and FSB Term voltages :

Old bios -- New bios
PLL 1.56 --> 1.52
FSB 1.26 --> 1.22
NB 1.26 --> 1.22

you meant v.1611 using lower voltage than old bios but it still STABLE ??

[angelreaper]

Bios 1611 rock stable here:

[Leeghoofd]

Well Vcore acts a bit differently then before : Think Martin.vr already mentioned that : my board tends to overvolt a bit more then before when being loaded...

With 1406 bios I had to set 1.256 to get 1.24 idle and 1.26 loaded under LinX ( CPU-Z readouts )
With 1611 bios I have to set 1.25 to get 1.24 idle and 1.26 loaded under LinX ( CPU-Z readouts )

Under prime it's funny that the board seems to give less juice as it tops at 1.256

The PLL, FSB Term and NB voltage needed to be that high to remain stable under IBT of LinX with 1406, 1611 allows in my case (with the QX9650) to lower them voltages considerably...

Maybe I'm an isolated case, the Vcore I'm pretty sure it acts up like that on several boards... for the rest you guys have to find out... Issue for me is that I cannot verify all the readings with my DMM as I've built the board into a case with all wiring neatly modded... and I'm not eager to pull everything out again... (before I had one board on a testbed ) so for the Vdimm I'm still at 1.92 will try to go lower there today...

Second thing : I mentioned before was that I could run Static Read control enabled this time... After some testing it seems that the bios engineers reversed it's function... look at the below results at 4ghz (in bold the better ones ) : (scores are run after a fresh restart, best out of three)

400FSB - 1002Mhz ram speed - 266 strap - Clock Twister Light - PL 9

SRC Enabled : R 8210 W 8471 L 63.3
SRC disabled : R 8255 W 8474 L 63.1

400FSB - 1066Mhz ram speed - 400 strap - Clock Twister Light - PL 8

SRC Enabled : R 8289 W 8448 L 62.1
SRC Disabled : R 8308 W 8479 L 62.0

450FSB - 1080Mhz ram speed - 333 strap -Clock twister Light - PL 9

SRC Enabled : R 8748 W 9568 L 59.4
SRC Disabled : R 8831 W 9415 L 59.3

Conclusion on my board is that Static Read Control Disabled yields slightly better scores now...

Can anyone else confirm this behaviour ?

[ThugsRook]

Conclusion on my board is that Static Read Control Disabled yields slightly better scores now...

Can anyone else confirm this behaviour ?

SRC always looked like that on my rig.

i havent tested 1611 for lower voltages yet, but its stable on the same settings as 1406. im not seeing any performance difference. 1501 didnt work too well for me.

[Leeghoofd]

With Static Read control ENABLED before I got better bandwith and lower latency and enabling this function alwys made my rig more unstable...

[ThugsRook]

With Static Read control ENABLED before I got better bandwith and lower latency and enabling this function alwys made my rig more unstable...

i am able to enable it, but my scores have always been worse with it enabled. i would concider that normal. everything is usually backwards for you quad guys so maybe they did fix it so it behaves the same way as a dually does

[JohnLoph]

i try Static Read Control to Enable / Disable it have the same result (nothing change)

but the change from 1306 ---> 1611 to me is Lantency improved from 55.9 to 55.1 (Clock Twister from Moderate to Strong)

E8400 E0, 550x7.5, 1100 CL5, PL 11

And another try to lower vPLL, vFSB, vNB but i seem doesn't work at all (can't lower than 1306)

[x3mgamer]

i used to have blend test fail with all 4 worker threads within 5 mins but now with this 1611 bios only one core fails within a minute and the rest keep going on forever, whats going on?

[Leeghoofd]

i used to have blend test fail with all 4 worker threads within 5 mins but now with this 1611 bios only one core fails within a minute and the rest keep going on forever, whats going on?

Is the CPU small FFT stable ? prolly still some more NB and ram tweaking required

[-n7-]

I was actually doing to some testing with static read control yesterday.

For me it doesn't seem to do anything, on or off or auto, Everest results are so close i'd call it margin of error.

Does anyone actually know what static read control does?

But even moreso, what is read training & OC charger?

I also tested with those on & off, & same thing, i really could not find any measureable difference.

[x3mgamer]

small fft stable, ill try playing with voltages, who knows maybe that "faulty" stick aint so faulty afterall

[gurusan]

Does anyone know of a pencil mod for this board? For some reason I am getting better clocks with LLC disabled, but the vdroop is just...atrocious. I'd really like to vdroop mod it if its possible.

I searched but could not find any info for the P5Q series.

[platojones]

Does anyone know of a pencil mod for this board? For some reason I am getting better clocks with LLC disabled, but the vdroop is just...atrocious. I'd really like to vdroop mod it if its possible.

I searched but could not find any info for the P5Q series.

Well, I hate to tell you this...but LLC *IS* the pencil mod...only it's been put into the BIOS for you by ASUS...so the answer is no...no pencil mod.

[3dchipset]

Does anyone use the top PCI slot above the PCIe 16x slot? Was wondering if any conflicts occur when using that PCI with a graphics card?

[MarkJohnson]

Does anyone use the top PCI slot above the PCIe 16x slot? Was wondering if any conflicts occur when using that PCI with a graphics card?

I use that slot for my sound card without issues. I don't think any of your ports should have issues of any kind at all. Unless, of course, your cards take up two or more slots.

-=Mark=-

[Porksmuggler]

Does anyone actually know what static read control does?

But even moreso, what is read training & OC charger?

I also tested with those on & off, & same thing, i really could not find any measureable difference.

Best I can determine, but not likely any new info...

DRAM Static Read Control: Enabled should tighten the Performance Level/tRD/MCH Static Read Control Delay, however it seems like a negligible setting compared to the manual options of the Ai Transaction Booster. Previous ASUS boards, and likely this one, are unstable with Static Read Control enabled and Ai Transaction Booster set to manual.

Ai Clock Twister: Determines the number of memory access phases (CHA-B,PH1-4) that are tightened/pulled-in to the next lower PL/tRD, of course you could manually enable each under Ai Transaction Booster.

DRAM Read Training: Enabled should tighten memory settings, how, I have no idea, seems negligible in testing.

MEM. OC Charger: Enabled, I have seen explained as a signal quality compensation mechanism, how, again no clue.

[x3mgamer]

ok guys now i really need ur help, this is some of my recent communication with RAMGUY over there at corsair forums:

hi Ramguy,

ive bought 2 x TWIN2x4096-8500C5D G kits like 3 months ago and immediately ive started getting spontaneous system restarts once every two days or something, i thought it was my vista ultimate 64 acting imperfect like it is but then i started clocking the system and i noticed theres always failure in blend mode in prime 95, yesterday i tested all sticks separatelly in blend and all came thru for coupla hours xcept this one stick which always failed within first 2 minutes, now the question is, even though i can boot and play games and do everything with this "faulty" dominator stick, should i RMA or are occasional restarts and OC instability something i should live with?


Ramguy:



I would recommend using www.memtest.org to test your memory configuration. Load setup defaults in the BIOS and then manually configure the following settings:

Memory Votlage: 2.1v
Memory Frequency: 1066MHz
TCL: 5
TRCD: 5
TRP: 5
TRAS: 15
Command Per Clock: 2T

Test each module one at a time in the slot closest to the CPU, if you get 3 modules which pass, and a single module giving you errors then you likely have a failing memory module. If all the modules pass or all the modules fail, then there may be some other issue. Let us know your results!

Here is a guide to help you with memtest: so i followed his guide and burned a memtest iso to a bootable cd and setup bios exactly as he said and left memtest86+ overnight and when i woke up there were 24 passes and not one error, 24 damn passes with 8 tests each and no errors! and then i boot windows and start prime 95 and voila it fails after less than 1 minute, this stick was always failing blend after less than 1 minute while all the other sticks were stable, and now this stick does 24 passes with no errors??? how the hell am i to interpret that???

[Dalten]

Anyone have a recently purchased P5Q Deluxe? What is the revision number on your board? Mine is 1.03G for example.

[pangingIII]

450FSB - 1080Mhz ram speed - 266 strap -Clock twister Light - PL 9

SRC Enabled : R 8748 W 9568 L 59.4
SRC Disabled : R 8831 W 9415 L 59.3

you mean 333 strap ?

anyway i have a question, can you set PL 8 at the 450FSB 5:6 DDR2-1080MHz and also get a stable system ?
i'm just wondering because at the CPU FSB and memory setting i always and be able to use PL 8 in P5Q-D's BIOS with 2x1GB, 4x1GB, 2x2GB and 4x2GB RAMs and all version of bios available since the first bios that came with the board.