Went too far

[BobRoss]

Well it happened, It looks like i went too far. I had been pushing my 3.2ghz Prescott and ECS p4m900t-m all week. A little at a time then running 2 instances of prime 95 for a few hours while looking at the graph on speed fan then running 3dmark03 to see if there was an improvement. I have also been intermittantly checking out GPU OC settings to try to maximixe my 3dmark score. well last night it hung on 3dmark. so i rebooted didabled the oc setting, returned the memory speed and the voltage to auto, hit F10 to save and restart and .....no restart. it does not post now. It looks like it is in a loop where it powers on and gives power to the drives but no beep and nothing on the screen and then it pauses and does it again.
So at this point do i do the clear cmos jumper and give it power and then take the jumper off and it works again? Or does it sound more expensive than that?

[Gizmo]

Try clearing the CMOS. You may have to pull the battery, too.

[Samuknow]

Do you see anything on the screen?

[BobRoss]

Originally Posted by Samuknow

Do you see anything on the screen?

nope, the screen stays black the whole time. and there is no beep.

[Daniel ~]

Are the HD drives ticking away like during boot, are your fans running? Maybe the Video card...? Sorry re-read your post...I'm sitting on the same nickel myself.
Waiting on a friends Video card to diagnose weather it's the mother board or the Video card that's checked out on me. Have you a spare video card you can try to boot with?

[BobRoss]

There is power going to the drives.
and the screen switches from the "no input" to black. So some signal is getting to the monitor. (a dark force)

I do have another video card though.

So I will try the clear cmos thing and then switch cards.

[Daniel ~]

The short path to knowledge! ":O}

[Samuknow]

Unfortunately it is the board or the video. If you pull the card out, does it act exactly the same. The memory? The CPU?

[richvanrooyen]

Not familiar with that motherboard... Did you have a config setting to force it into the bios?

[BobRoss]

Good News!
I used the clear cmos jumper and it did the trick.
I reset the defaults and ran 3dmark twice, which is what it was doing when it froze up in the first place, and it seems good to go.

I was looking at the default settings in cpuZ and i noticed that the default memory ratio is 3:5. I thought I read somewhere that 1:1 is the optimum ratio. Which leads me to wonder why it would default to 3:5.
Am I mixed up about this?

[Daniel ~]

Well done Bob! ":O}

(I've always run 1:1 as this seems the best for my overclocking... not sure what defult should be...Guys?

[MUff1N]

From what I've read all over the place, is the reason 1:1 is used is so that you can get your highest FSB clock speeds without the memory coming into play (for stability) because of the lower divider (frequency) used.

Now, IF I had a better CPU cooler and was going to go up to 3.6Ghz or higher I'd use a lower memory divider so that my memory wouldn't be so overclocked that it couldn't keep up anymore.
This is the reason I now have my memory at 4:5 (667) to keep it around 800Mhz which it is designed to run at and is stable, not to mention I found no benefit (added throughput) to overclocking my memory to 960Mhz. But I did notice much better throughput once my FSB went up & my games played much smoother.
That's my take on it anyways...Correct me if I'm wrong though guys! hehe

Btw, just so I could see I did a little experiment & I put my system back to default and played some of my games at their current "high" settings to visually see if there was any noticeable difference...
OMG!!! Jerky, lag and slow to catch up in places with stops and starts... What a diff without the added throughput!!!

[Gizmo]

Generally speaking, the AMD Athlon and AthlonXP cores perform better when running 1:1. I believe this is due to the way the internal caching is implemented. The upshot is that lower FSB running 1:1 (synchronous) can net you better memory performance than higher FSB running 5:4 (asynchronous) or something like that.

With the Intel P4 family, the caching mechanism works differently, and is less sensitive to memory latency, preferring higher FSB clocks. This is directly opposite the case with the AMD Athlon and AthlonXP cores.

Intels Core family CPUs have continued this trend from what I understand, and thus continue to perform better with high FSB, even at some sacrifice of memory latency. However, with the advent of the AMD Athlon64 and its integrated memory controller, the AMD architecture is no longer so sensitive to memory latency, and thus tends to now also favor higher bus speeds (though in this case it would HT clocks, rather than FSB clocks).

Hope that helps.

[Samuknow]

I have tested with mem at stock and OC also. I saw no diff. It loves the high FSB as Gizmo stated.