Why Underclock your Intel Processor (x14)?

[Eastcoasthandle]

Could someone explain why you would reduce your multiplier from x18 to x14 then overclock your CPU to 300 or 350 FSB??
Would overclocking your CPU at x14 multiplier yield any better results then at x18 multiplier?

[Twisted-net]

Ehm thats a hard one...... To get a HIGHER fsb maby !!!

lol doesnt seems to be that hard to understand

You can stay on the same speed as the X18 multi but your fsb is way higher when you use the X14 multi.

[Kunaak]

say your CPU maxs out at 4.2 ghz stable on air...

but your default multiplier is 18, like you said...

18 x 233 = 4.195 ghz.

but say you know your ram can do alot better then 233 at 1:1.

so you lower your multiplier to 14.

then 14 x 300 = 4.2 ghz...
then you have the options of using the 5:4 divider or 3:2 divider, or maybe some others depending on your board...
this gives you the option of hitting your CPU's limit (in this case, we will just say the CPU's limit was 4.2 ghz for easy math) and the option of being able to push your ram farther then 233 at 1:1.

the higher your ram goes, the more bandwidth you got.

more bandwidth is always nice

thats why.

[Eastcoasthandle]

say your CPU maxs out at 4.2 ghz stable on air...

but your default multiplier is 18, like you said...

18 x 233 = 4.195 ghz.

but say you know your ram can do alot better then 233 at 1:1.

so you lower your multiplier to 14.

then 14 x 300 = 4.2 ghz...
then you have the options of using the 5:4 divider or 3:2 divider, or maybe some others depending on your board...
this gives you the option of hitting your CPU's limit (in this case, we will just say the CPU's limit was 4.2 ghz for easy math) and the option of being able to push your ram farther then 233 at 1:1.

the higher your ram goes, the more bandwidth you got.

more bandwidth is always nice

thats why.

Thank you very much this is the explanation I was looking for.
This is a bit complicated because I have a P5GD1, 660, with a Corsair 3200C2PT (2-3-3-6 up to 225, 2.5-3-3-6 230 so far at 18x multi) and, it appears that I also have to factor in the PCI Express Frequency. If I leave it at 100 it will not post when I overclock. If I am doing this right, I start off with a PCI express Frequency base of 100 with a FSB base of 200 then work my way up by a factor of 2.2 (for every 2.2 MHz on the FSB you increase your PCI express frequency by 1...I think that's right...)
For example:
Your PCI Express Frequency has a base of 100. Your FSB has a base of 200.
If you increase your FSB to 230 then you increased your base by 30 MHz.
Divide that by 2.2 (30/2.2) and you get 13.63.
Round 13.63 up and you have 14.
I add 14 MHz to my PCI Express Frequency (100+14) and get 114
114 will be my new PCI express frequency at a FSB of 230. So far it's been working and pretty stable. Not sure if this is right or not though.
Also, not sure if this will also work at a 14x multi.

However, if I want to push my 3.6 to 4.2X (I believe thats at 285 FSB) using a 14x mulitplier what do I need to do? Any help is appreciated.

[n00b 0f l337]

Leaving your pci-e frequency at auto works just as well.

[FUGGER]

Leave PCIE alone, adjust that only for benching 3D.

I wonder how you came up with that calculation for PCIE speed??

[Eastcoasthandle]

Leave PCIE alone, adjust that only for benching 3D.

I wonder how you came up with that calculation for PCIE speed??

I found it one of the threads....do not recall where but there was no instruction as to how they derived at that number nor how to use it. So I did some experimenting and figured it out realizing it actaully worked. I knew you needed some sort of prime number to calculate in order to know how to increase the PCI express frequency along with your FSB but I didn't know what the ratio was. So I googled it and that's how I found it (would have taken me forever to figure that out) I will try placing it back to auto and see what happens (thanks n00b 0f l337). I believe I did this before and had CTD. BSOD and no longer had a boot drive...I forget the error but my MB no longer recognized my Sata HD . But I'll see what happens as this has all been trial and error for me.


But I still like to know how to underclock my CPU via 14x multi...any help is appreciated...

Edit: This is for a 915 Chipset, Asus P5GD1 MB

[saaya]

thats only for some 915 chipset based boards like the p4gd1, the 915 and 925 chipset have fcked up locks afaik. at least some boards.

and as to why we lower the multi, 10x200 is faster than 20x100 even though both result in 2ghz cpu speed. the reason? faster fsb = more bandwidth, lower latencies, both limiting factors for cpus

[Eastcoasthandle]

thats only for some 915 chipset based boards like the p4gd1, the 915 and 925 chipset have fcked up locks afaik. at least some boards.

and as to why we lower the multi, 10x200 is faster than 20x100 even though both result in 2ghz cpu speed. the reason? faster fsb = more bandwidth, lower latencies, both limiting factors for cpus

Does anyone know of any timing configs for PC3200C2PT @ 300 FSB?
It runs at 2-3-3-6 at 200 FSB what's the formula for finding the timing of the ram @ 300 FSB?

[targ]

there is no formula to work out timings,

you just have to try it and see what works.

if you know what type of ram you have (TCDD, BH-5 etc) you can search other ppls threads and settings.

but you just have to test.

targ

[ibby]

thats only for some 915 chipset based boards like the p4gd1, the 915 and 925 chipset have fcked up locks afaik. at least some boards.

and as to why we lower the multi, 10x200 is faster than 20x100 even though both result in 2ghz cpu speed. the reason? faster fsb = more bandwidth, lower latencies, both limiting factors for cpus

yes yes. i hate them timings.

[Salkcin]

Theres no way yo get to overclock a 915 based chipset above 278MHz FSB and that's in it's very best conditions. It have a unfixable lock wich means you have to make a lot of considerations if you wanna overclock on such chipset not to meet a limitation...
Also considering noone ever made it to DDR600 stable on a Pentium 4 platform (865/875 chipset is the limiting factor, 915 is un-overclockable and all other don't support DDR). Maybe a few got DDR600 CAS3-4-4-8, but it's few!

Read Anandtechs article about the lock here : http://www.anandtech.com/cpuchipsets...oc.aspx?i=2124

Anandtech reached 278MHz, but the best I myself ever seen anyone archive on 915/925X (925XE is fine though it still has the lock) is 254MHz FSB. 925XE is better because of the default FSB of 266MHz wich the lock can be manipulated to consider as default on some motherboards.
945/955/975 boards are without locks

[saaya]

Theres no way yo get to overclock a 915 based chipset above 278MHz FSB and that's in it's very best conditions. It have a unfixable lock wich means you have to make a lot of considerations if you wanna overclock on such chipset not to meet a limitation...
Also considering noone ever made it to DDR600 stable on a Pentium 4 platform (865/875 chipset is the limiting factor, 915 is un-overclockable and all other don't support DDR). Maybe a few got DDR600 CAS3-4-4-8, but it's few!

Read Anandtechs article about the lock here : http://www.anandtech.com/cpuchipsets...oc.aspx?i=2124

Anandtech reached 278MHz, but the best I myself ever seen anyone archive on 915/925X (925XE is fine though it still has the lock) is 254MHz FSB. 925XE is better because of the default FSB of 266MHz wich the lock can be manipulated to consider as default on some motherboards.
945/955/975 boards are without locks

thats was a long time ago, and they didnt update the article when we found out that the ocing lock was actually just a low vdd
and 915 cant reach over 274fsb?
then how did ojdr2001 and hipro5 reach almost 300fsb?
i think hipro even reached 315fsb?

but yeah, the 915 chipset is a horrrrrable ocer, just like the 925
wich is really weird considering the 865/875 oced so incredibly well... ive seen up to 400fsb with 865/875 boards and the next gen chipset has problems to reach 300fsb

[Eastcoasthandle]

OK, I get the hint I'm getting a ASUS P5WD2-E Premium. But there is one slight problem. There is a green version floating around and vendors could try to sell it as the real McCoy at a cheaper price. The green version (from what I understood...could be wrong) has something to do with EMI emission standards using less conductive metals due to RoHS in other countries. Different metals and components are used but in a nut shell is a crappy board. I was told that at this time sells for $250-$260. The real verisions are $300-$350.

Check this out
Computer HQ P5WD2-E PREMIUM GREEN 975X VS. NewEgg ASUS P5WD2-E Premium

Now check out the ZZF ASUS P5WD2-E Premium
Why are they out of stock on Asus P5WDG2-WS 975X with 8-Phase Power Design


Now, back at the ranch. What are some good 2 GIG DDR2 that's oc'able??

[SlackerXL]

OK, I get the hint I'm getting a ASUS P5WD2-E Premium. But there is one slight problem. There is a green version floating around and vendors could try to sell it as the real McCoy at a cheaper price. The green version (from what I understood...could be wrong) has something to do with EMI emission standards using less conductive metals due to RoHS in other countries. Different metals and components are used but in a nut shell is a crappy board. I was told that at this time sells for $250-$260. The real verisions are $300-$350.

Check this out
Computer HQ P5WD2-E PREMIUM GREEN 975X VS. NewEgg ASUS P5WD2-E Premium
i can't wait to see what the "non-green" version can do

Now check out the ZZF ASUS P5WD2-E Premium
Why are they out of stock on Asus P5WDG2-WS 975X with 8-Phase Power Design


Now, back at the ranch. What are some good 2 GIG DDR2 that's oc'able??

I sure hope this is just a speculation and not a real fact
as my newly arrived P5WDG2-WS has the "Green ROHS" logo on the box
if thats a crappy board,
i can't wait to see what the "non-green" version can accomplish

anyway, the price difference is normal
as the first link refers to P5WD2-E Premium
while the other is about the more expensive WS version

[Eastcoasthandle]

Rohs means that it remove some level of lead out of certain components. Sorry for the scare as the removal of lead from metal doesn't sound to me like a bad thing at all. Besides, all P5WD2-E Premiums in the States will come Green so there is no way of getting around it anyway.

[SlackerXL]

Rohs means that it remove some level of lead out of certain components. Sorry for the scare as the removal of lead from metal doesn't sound to me like a bad thing at all. Besides, all P5WD2-E Premiums in the States will come Green so there is no way of getting around it anyway.

looks like the same applies here in Greece also....

anyway
i hope Asus ecological concerns will not not affect the clocking potential of this mobo in any way...

[perkam]

Eastcoast,

The P5GD1 was one of the worst clocking 915 boards...they were pairing those up with X600XTs for like $100 because no one was buying them.

...not to mention most reviews of the board were done in the era when everyone was dealing with Intel's supposed oc lock (lack of pci lock on mobo to discourage oc) etc...

In any event, getting the fsb that high on your mobo would be unlikely...

Perkam

[JoeBar]

looks like the same applies here in Greece also....

anyway
i hope Asus ecological concerns will not not affect the clocking potential of this mobo in any way...

Indeed, i just got a P5WD2-WS green.
Really good oc'er!

[SlackerXL]

Indeed, i just got a P5WD2-WS green.
Really good oc'er!

Plaisio rulezzzz

(although i don't believe that crap about worldwide exclusive release etc..)

[JoeBar]

Plaisio rulezzzz

(although i don't believe that crap about worldwide exclusive release etc..)

However they got both asus 975 mobos really fast!
Not many ppl have them.