ASUS P5W DH - Problems + Fixes Thread

[Polynikes]

Your CPU is only at 40c on full load? What kind of watercooling setup do you have?

[hoax]

Network Disconnecting - Upgrade Marvell Yukon NIC to latest driver; http://www.marvell.com/drivers/drive...?dId=116&pId=3 [Credit - BulldogPO]

guys I have just reformated my pc, and the above link is no longer working... Last time I had to format, I used that update and never had a problem....but I wasnt smart enough to backup the file..

this time windows keeps finding the second ethernet controller everytime I reboot... and device manager reports it as "This device cannot start. (Code 10)"

I have already installed (and later removed) the driver that came with the mobo cd, and atm I am using version 10.22.1.3 that I found at marvell webpage... still gives the same error

does anybody have any advice on how to get rid of this error? or a new link to the driver that works????

TIA

[nrs250]

Can someone explain to me the following concept.

If you push the FSB and compromise by lowering the memory clock and loosen timings what is the effective tradeoff in performance?

Example 1:
Dominator C5D PC2-8500

FSB 345 and PC2-5700 at 5-5-5-15

Vs.

Example 2:

FSB 333 and PC2-6400 at 4-4-4-12

What is the tradeoff? Which is the better performance all other factors being equal?


In Some ways the PC-26400 runs better because I can run tighter timings at a lower FSB.

If you push FSB at the expense of memory speed is the net-net in your favor or against.??

[John600rr]

If you push FSB at the expense of memory speed is the net-net in your favor or against.??

It depends on how much FSB you can gain by loosening your mem timings.

In my case, my board will not function properly with manual ram timings above 370FSB. Now I could run 370 and some nice and tight timings, or I could run 425 and pretty loose timings. With a 9x multi, the 500 extra mhz make up for the loose timings.

But maybe you've got a 7x multi, and the difference is only say 10 FSB. That only nets you 70mhz more in clockspeed. I'd take the tighter mem timings in that case.

[FabricioGS]

Hi Folks, just a few quick questions:

1) Microcode update: what does that do? What's better for OC stability?

2) What temps are ok for stock and load? I'm using Intel's TAT for measurement.

[MarcusTaz]

Can someone explain to me the following concept.

If you push the FSB and compromise by lowering the memory clock and loosen timings what is the effective tradeoff in performance?

Example 1:
Dominator C5D PC2-8500

FSB 345 and PC2-5700 at 5-5-5-15

Vs.

Example 2:

FSB 333 and PC2-6400 at 4-4-4-12

What is the tradeoff? Which is the better performance all other factors being equal?


In Some ways the PC-26400 runs better because I can run tighter timings at a lower FSB.

If you push FSB at the expense of memory speed is the net-net in your favor or against.??

Benchmark your system using Everest's "benchmark>quickreport" and you can see the diff yourself. Tight timmings do not necessarly mean better performance overall...

[gwolfman]

...moved my memory back to spd rather than manual which helped...

What FSB speed were you able to get to when you manually set it?

[gwolfman]

Example 1:
...
FSB 345 and PC2-5700 at 5-5-5-15

Vs.

Example 2:

FSB 333 and PC2-6400 at 4-4-4-12

Your example isn't a very good one or incorrect. In example one your stated ram speed is PC2-5700 which is slower than PC2-6400 in example two AND your timings in example one (5-5-5-15) is slower than in example two (4-4-4-12). It's just not right. Faster timings (lower numbers) are usually correlated with slower speeds, but you have slower speed with slower timing. Did you mean PC2-5700 @ 4-4-4-12 and PC2-6400 @ 5-5-5-15?

[BulldogPO]

Latest Marvell drivers can be found here:
http://www.station-drivers.com/page/marvell.htm

[aledsav1]

What FSB speed were you able to get to when you manually set it?

I found that with manual memory any fsb above 457 wouldn't run orthos at all just set it back to spd and well orthos ran...though still have not performed full orthos run yet. my memory ain't great Corsair 6400.

[aledsav1]

Your CPU is only at 40c on full load? What kind of watercooling setup do you have?

my e6750 stock temps are low as low as 8c.... at 3.7 with vcore 1.55 idle in coretemp is 19-20 load is 38-42 TAT show 23 idle and 36 load. Have got a AquaXtreme block can't recall exact setup without having a look as have mixed and matched over time.

[Polynikes]

8C? On water? Yeah, right. I have a pretty good watercooling setup and running at 1.3v on my CPU my idle temps are around 25/29c @ 68F ambient. Are you using a TEC waterblock or something?

[aledsav1]

Here check out 3.7...fsb@465...temp underload 36c! now thats at its lowest on orthos it does go upto 44c during tests.

Here is idle temp 1.5625vcore. 18c as i am saying at stock voltage my best idle temp has been 8c! but usually around 10-14c...
hope my images are usable as new to image hosting.
coretemp don't show correct voltage?

[aledsav1]

Goto say bought this board on saturday on a whim was using aw9d though couldn't get over 3.646 stable....wasn't sure what to expect but am quite happy with results so far...though as stated earlier my e6750 would like a higher fsb limit so considering a mobo move again??

[Polynikes]

The links don't work.

What kind of watercooling parts do you use and what was the ambient temp?

Here's a question for those of your smarter than I. Is it even POSSIBLE for a regular (not TEC) watercooling loop's temperature, and therefore a CPU's temp, to be lower than the ambient temp? That's physically impossible, isn't it?

[hoax]

Latest Marvell drivers can be found here:
http://www.station-drivers.com/page/marvell.htm

THANK YOU!!

[aledsav1]

http://img162.imageshack.us/my.php?image=capturetp3.jpg

http://img162.imageshack.us/my.php?i...apture2mn5.jpg

Hopefully these links work? as said 1.5625 vcore 3.7g = 18c idle.....load anywhere between 36-44 during test

Ambient temp? clueless.....wheres my abient temp, just looked at someone else oc and their temp on cores was 65c and their ambient temp was 77c?
water setup...aquaxtreme block and pump...rad top of case with 2 120m yates

[aledsav1]

So polynikes 'im a little slow at catching on' what your saying is I can't have the temps im reporting? well i have had this e6750 on another board aw9d and my temps on that where great too...could it be the cpu is sending incorrect info? on previous board coretemp and speedfan would report (can't recall exact) say 20c though bios would report 31c, speedfan,coretemp report same load temps Tat reports lower load temps my bios on p5w reports pretty close to Tat.

[~aoe~]

You will notice that those temps are based on a tJunction of 85C. I believe the correct tJunction for E6750 and E6850 should be 100C.

You will need to add 15C to the CoreTemp temps.

I would hope a later version of CoreTemp will correct this error (when it's released).

[aledsav1]

tahnks aoe makes sense now....did notice other posts else where refering to this....saying that why does speedfan report same as coretemp and TAT report lower than both under load though higher than both idle? and given that tj is 100 are higher core temps ok

[gwolfman]

...Is it even POSSIBLE for a regular (not TEC) watercooling loop's temperature, and therefore a CPU's temp, to be lower than the ambient temp? That's physically impossible, isn't it?

Correct. The cooling effect in water and air cooling is the air that is passing the radiator/heatsink. It is impossible on regular water/air cooling to get lower than the ambient temp.

[gwolfman]

...Ambient temp? clueless.....wheres my abient temp...

Ambient temp is the air temp in your room/house. Easy way to get an approximate temp is to look at the temp reading on your AC/heater thermostat and see what it says.

[aledsav1]

Correct. The cooling effect in water and air cooling is the air that is passing the radiator/heatsink. It is impossible on regular water/air cooling to get lower than the ambient temp.

Please forgive my ignorance....when my fans are pulling air through the rad its air thats the temp of the room(ambient) how come this air coming into the case feels so much colder than my environment also my water tubing is cold to the touch....in my mixed up logic this tells me my ambient temp dosn't dictate that my cpu can't be colder???

[~aoe~]

I'll have a go at trying to explain, but not sure I'm qualified...

Your skin emits liquid (sweat) which, when air passes over it, creates a cooling effect (due to evaporation of the liquid). This is why the air coming into the case 'feels' cooler (it's moving and passing over your skin).

The radiator fins don't emit a liquid (hopefully), and therefore does not cool by evaporation.

The reason the tubes feel cool is that they are cooler than your body temp (body temp ~37C).

It is possible in humid conditions that water condenses on the radiator, which would then assist the cooling effect (by evaporation), possibly allowing your temps to fall slightly below ambient. However, we're only talking 1-2C here at most.

[aledsav1]

I'll have a go at trying to explain, but not sure I'm qualified...

Your skin emits liquid (sweat) which, when air passes over it, creates a cooling effect (due to evaporation of the liquid). This is why the air coming into the case 'feels' cooler (it's moving and passing over your skin).

The radiator fins don't emit a liquid (hopefully), and therefore does not cool by evaporation.

The reason the tubes feel cool is that they are cooler than your body temp (body temp ~37C).

It is possible in humid conditions that water condenses on the radiator, which would then assist the cooling effect (by evaporation), possibly allowing your temps to fall slightly below ambient. However, we're only talking 1-2C here at most.

Many ta's......yes of course now my reason has been re-alined, makes perfect sense...