ASUS P5W DH - Problems + Fixes Thread

[Gig4moller]

But other than the ratio stuff it ran fine and you were able to overclock it?

Awesome news if they can make this true.

[Falkentyne]

All, this might be an old news, just thought I'd share anyway
Just got reply from ASUS support on my question about QX9650 support:

----------- start quote ------------------
-------------------- end quote -------------------------

Take it for what it is worth. These guys are pretty responsive and from my experience (may be I am just lucky) they eventually deliver what they promise. It might take time though :-). My take that it won't happen until Intel releases new 45nm CPUs early next year

I did had a chance to play with 45nm CPU on this board using BIOS 2403. I managed to get it up and running and in general it works as advertised - ~20% or more performance gain vs my old QX6700 (B3), ~30W less power consumption, etc. BIOS is clearly not there though - it required CMOS reset each time it did not like the changes I made in the CPU configuration in the BIOS. In particular, it hated me fiddling with the ratio multiplier.

Eh? What CPU? O_o

I don't have that problem on my QX9650 and multipliers work fine. I've been up to 15x.

2406 (and 2403) both support the QX9650. However I don't know what Asus is talking about. They MIGHT be talking about the BIOS not identifying any of the cache in the cpu properties (it works, obviously), -or- they might be talking about the lack of a .5 multiplier, which the Penryns support.

But yes, you should have no problems getting a QX9650 running with 2406. How far you can push the FSB is another question, though. Highest I could hit stable was 370 FSB 370*10., and 400*10 was a NO POST, but 333 *12 worked fine @ 1.3625v O_O

[DarknessRevenge]

p5w dh deluxe (1602 bios version)
c2d E6300@1,86 Ghz
3 GB RAM
WD 320 GB SATA
ati radeon x1650 ...etc
I have some problems (like slow boot etc) and questions.

1) Update to a latest bios !
2) 3GB ? That means 3 modules ? Four ?? Try to use only 2 modules of the same brand and capacity (example: 2x1GB) in the dual channel slots (the orange ones are fine).

Slow boot can be related to these two things. Try to fix them. Bye

[atko]

@DarknessRevenge, thanx first

I know, I have to update to the latest bios

3 GB - 2x1GB and 2x512MB kingston , I'm not sure about its ordering in slots, I can't remember. I'll check.

[Mr.X]

I just put in the Q6600 I had ordered but I the multi doesn't want to change for some reason. I set it to 8 in the BIOS but it's still 9. Why is this happening?

Any idea why it's displaying my clock speed incorrectly? I want to run at 400x8 but it it says 3.6Ghz (400x9) in the post screen. I've updated to the latest BIOS version but it's still the same.

[Unbornchild]

1) Update to a latest bios !
2) 3GB ? That means 3 modules ? Four ?? Try to use only 2 modules of the same brand and capacity (example: 2x1GB) in the dual channel slots (the orange ones are fine).

Slow boot can be related to these two things. Try to fix them. Bye

You do not HAVE to update to the latest bios.

Slow boot can be related to these things:
-not enabled quick boot.
-not set all fans to [ignored] in hardware monitor
-not disabled all unneeded devices (except hdd as the first, obviously) in boot order

try these first.

Oh, and if you put 3 modules in, you lose dual channel mode.

[Lee13168]

w t f
So my p5w was always unstable around 333 fsb even if i put the nbridge voltage at 1.85. I update my bios to 2406 and to my surprise its now stable around 350. However i need to disable hyper path. So i OC my cpu from 2.9 to 3.1 and my ram loses maybe 25mhz. I run 3dmark06 and to my surprise i lost 600points. Does hyperpath really make so much of a difference to overshadow .2ghz cpu gain? The ram shoulnd not have anything to do with this loss since i tested how ram effects 3dmark06 and 200mhz barely took 100 points off.

[atko]

You do not HAVE to update to the latest bios.

Slow boot can be related to these things:
-not enabled quick boot.
-not set all fans to [ignored] in hardware monitor
-not disabled all unneeded devices (except hdd as the first, obviously) in boot order

try these first.

Oh, and if you put 3 modules in, you lose dual channel mode.

Generaly, I wont newer bios than this one which I have currently.

I have 4 modules, 2x1GB and 2x512MB.

Of course I'll make changes which you have recommended.

[bichi]

Hi Bichi,
Would you mind sharing your FSB/Multilpier settings for your 3.0 OC? I just got my q6700, and i am running 266x10, but when i first booted, it was running 266x8. With BIOS 2406, i am able to change the Q6700 multiplier between 6 to 10. Am i doing this correctly?
Thanks
Tom

Here ya go
- BIOS 2406 setting screen shots
- 2.3mB in zipped format:
http://www.esnips.com/doc/548739a4-d...6-Screen-Shots

[romfi]

Eh? What CPU? O_o

Yep, you got me, it was an ES :-). The stepping was C0 though - just like the production QX9650.

2406 (and 2403) both support the QX9650. However I don't know what Asus is talking about. They MIGHT be talking about the BIOS not identifying any of the cache in the cpu properties (it works, obviously), -or- they might be talking about the lack of a .5 multiplier, which the Penryns support.

Right. On top of that I also noticed that SpeedStep option in BIOS was not functioning correctly. Have you noticed anything?

But yes, you should have no problems getting a QX9650 running with 2406. How far you can push the FSB is another question, though. Highest I could hit stable was 370 FSB 370*10., and 400*10 was a NO POST, but 333 *12 worked fine @ 1.3625v O_O

I personally decided to wait for the BIOS update before I throw $$$ to buy a Penryn based CPU. I need the system for 24x7 with power consumption optimization features working.

[Troop]

In order to fully utilize 4GB ram, you MUST be using Windows XP 64bit/Vista 64bit. Next you need to enable "Manual" overclocking mode in the BIO then enable "Memory Remap Feature". Then you will see 4GB ram in boot screen and Windows.

Steps:
1) BIO > JumperFree Configuration > AI Overclocking > [Manual]
2) BIO > Advanced > Chipset > Memory Remap Feature > [Enabled]

Thanks a lot for the information. I had thought I would need 64 bit as well, not sure what I was thinking when I posted my question. I am using Vista 32 bit Business edition at the moment. Next year (after SP1 is released) I plan to go with Vista Ultimate x64.

Another quick question. This is the RAM I have. I had to go into the BIOS (just as some reviewers stated on Newegg) to manually set the timing to DDR2/4-4-4-12. My question is, do you think it would be safe to bump up the voltage to 2.0 or 2.1 like some reviewers stated here?

If I do bump it to 2.1, it shows up in red in the BIOS which I assume cannot be good. I have never really overclocked before (and I just have stock cooling on my rig) so I just want to be sure that I am not going to fry my system by doing so.

If you need more information, please let me know.

Thanks guys and Happy Holidays.

P.S. Memory voltage as of right now is set to Auto.

[w2richwood]

the manufactors say 2.1 it wont hurt the ram to run it a that volt
Rich

[Falkentyne]

oh no I never have used power saving features. I always ninja shut them off immediately.

I do remember (on 1602 bios) that sometimes, speedstep (or maybe EIST? the voltage was EXTREMELY LOW) was occasionally on when it was supposed to be disabled; my C2D sometimes had a vcore of 1.2v or something while idle (when it was set to 1.35 in bios).

[lawrywild]

oh no I never have used power saving features. I always ninja shut them off immediately.

I do remember (on 1602 bios) that sometimes, speedstep (or maybe EIST? the voltage was EXTREMELY LOW) was occasionally on when it was supposed to be disabled; my C2D sometimes had a vcore of 1.2v or something while idle (when it was set to 1.35 in bios).

speedstep = EIST (Enhanced Intel Speedstepping Technology)

Eist rules... everyone should use EIST to save power and lower temps..

[firsttimmer]

Hi Bichim

Thanks alot for your help!

Tom

Here ya go
- BIOS 2406 setting screen shots
- 2.3mB in zipped format:
http://www.esnips.com/doc/548739a4-d...6-Screen-Shots

[Snake_Eyes]

speedstep = EIST (Enhanced Intel Speedstepping Technology)

Eist rules... everyone should use EIST to save power and lower temps..

I have tested useing EIST and C1E and i see no difference between usage in watts..
Only 1 a 3 watts difference
So EIST are not that what you say about it..
C1E are better for me..
if i enable it and run 3ghz the temps are iddle better about 3-5c
the C1E set the vcore iddle to 1,14v and stressed it goes to 2,23v
But with EIST i see no difference in Watts usage or temps.

So C1E are the rules !!!
But with higer overclock over 3 Ghz than you cant use C1E becource you have more vcore needed..

[romfi]

Actually it is the EIST feature that allows processor to drop to a lower frequency (by changing the bus ratios) and voltage, conserving power while maintaining reasonable responsiveness. You can use CPU-Z to monitor this behavior.

C1E stands for Enhanced HALT state support in the CPU. When system is idle, OS will try to halt CPU cores in order to save power. If I am not mistaken, C1E means that while a core is halted it shuts down some of its functional units in order to further improve the power conservation.

[fordf250]

w t f
So my p5w was always unstable around 333 fsb even if i put the nbridge voltage at 1.85. I update my bios to 2406 and to my surprise its now stable around 350. However i need to disable hyper path. So i OC my cpu from 2.9 to 3.1 and my ram loses maybe 25mhz. I run 3dmark06 and to my surprise i lost 600points. Does hyperpath really make so much of a difference to overshadow .2ghz cpu gain? The ram shoulnd not have anything to do with this loss since i tested how ram effects 3dmark06 and 200mhz barely took 100 points off.

Over 333 you need to disable hyperpath to oc,e6600@425fsb no prob I never noticed a difference in benchmark scores without, you might need to bump some voltage, your score should go up for 200mhz extra.

[Crow-]

Hi,

I have a little problem with my SATA DVDRW and P5W DH Deluxe.

On the 3 SATA ports I have 3 HDD's, so I've connected my DVDRW (SATA) to the JMicron port, but when it's reading a DVD the system becames to slow
If I change the DVDRW to another SATA port it's all ok, and if I connect a HDD to the JMicron port it's also slow when it's reading/writing

What can I do to fix this? Or, can I use the RAID SATA ports to connect one HDD or the DVDRW?

Thanks

[genish]

all asus systems bios can't come back between 2406 2004
afudos az bios and update with option "old enable" failed.
how can I change the bios date 2004.
i had test to change the date later then 2406 it' s ok to read .
but it is dedecting than too old .
the 2004 is genial to change timing @3.4 and memory timing speed.
i can do that with the new bios black screen when reboot.
excuse me for my translate.

[lawrywild]

all asus systems bios can't come back between 2406 2004
afudos az bios and update with option "old enable" failed.
how can I change the bios date 2004.
i had test to change the date later then 2406 it' s ok to read .
but it is dedecting than too old .
the 2004 is genial to change timing @3.4 and memory timing speed.
i can do that with the new bios black screen when reboot.
excuse me for my translate.

Use an older version of afudos like 2.17

[MacClipper]

Hey guys. I've been using the P5WDH for almost 6 months now, thanks to Lawry and this great thread.

A few days ago, my PC did not POST after shutting down from the night before. I managed to trace it to 1 of my RAM sticks (I have Crucial Ballistix PC-8500 tracers). I know it's the RAM because, 1) I pull out the offending stick and the PC boots. 2) I put the offending stick in, and it doesn't POST. Nada. Nothing. 3) I trade out the RAM sticks and nothing, again.

After putting all my settings back to default, I ran memtest85+ on the 'working' RAM. I am getting errors on 2 blocks. It seems those two blocks are faulty. I'm currently in the process of getting an RMA for the pair, but had the following question:

In all my months of tweaking, I never pumped up the mem voltage, or freaked out the timings. These chips are rated at 2.2v, at default CAS reads 5-5-5-15 according to memtest and CPU-Z. Could the motherboard have damaged the RAM? It's hard to believe that it's my 2nd pair of defective crucial sticks. Has anyone else had similar problems?

Thanks for any input.

BTW aside from memory problems, I'm pretty happy with the Ballistix PC8500.

Just experienced the same problem today, PC running well a few minutes ago and I did a shutdown. It failed to POST when I tried powering it up a while later, a stick of Crucial 10th Anniv. in DIMM slot B2 had somehow died in that short downtime. All 4 sticks had been running well for the past year and at the moment of death, it was at stock FSB@DDR667@2.05v (with a 80mm fan cooling the 4 sticks) as it had been doing ever since I got the Q6600.

Put in my spare stick and rig's back up and running but an unwelcome hassle as lawrywild puts it.

[genish]

Use an older version of afudos like 2.17

I try 2.15 without underbios
but with 2.17 it's ok thanks......
but i try to overclock with my " older test" cpu 333 1.48v ram 834 with 2.1v .....overclocking wrong it's easy ok before
i test the mini overclocking
cpu+ram just 266 to 267 and spd to memory default ( manual)
i can't boot i shut my pc and when i reboot i must go to change bios with f1

but when I test automatique overclocking 10 20 30% it's ok
perhaps my old new bios have no change something in the bios .
I you a idea?

[w2richwood]

did you clear the bios after bios flash pull cmos battery and clear with the jumper on 2/3 be sure to unplug power cable
Rich

[Chrysalis]

I am on 2406 now no problems, also I read what was said earlier about EIST (speedstep) and C1E and decided to disable speedstep in the bios, the result was my voltage still drops when idle alongside the temp but I dont lose cpu clock speed so now I am sitting here wondering if speedstep (and coolnquiet) is just an overhyped feature since its C1E doing the power/heat saving. The only conclusion I can come to is just that by lowering the clock when idle reduces the chance of there been a stability issue when undervolted as the cpu itself is throttled but as soon as the cpu comes under load the voltage is upped so I reckon its a very low risk.