v mod 3/4 for p4b533

[Rectalfier]

Wow, Great News, I'm going to try it on my board.

 

[THUGSROOK]

 

[THUGSROOK]

ok i have confirmed (on a p4b533) that this is no longer a hack but rather the discovery of an undocumented dipswitch setting!

*dipswitch 6 on = 3:4 ratio in jumperless mode

NOTE: the ratio is misrepresented in the bios - its still says 1:1 and 4:3. but 1:1 is actually 3:4 - and 4:3 is actually 1:1.

the P4B533-VM board gets a 4:5 ratio - not a 3:4 and is dipswitch 5

use Sandra, CPU-z, or WCPUid to confirm.

 

[oldfart]

Very nice. The P4B533 now has my stamp of approval (for whatever that is worth).

Thugs, DDR444. Very cool. That setup must scream. Is that Corsair actually STABLE @ 444? How much voltage does it need to get there?

 

[THUGSROOK]

 

[oldfart]

Anything close to DDR440 is pretty darn good. Maybe a small fan blowing across the DIMM would help?

 

[THUGSROOK]

 

[SupermanCK]

P4B266-C is a no go...i repeat...It's a no go...
We have to do it the hard way

 

[THUGSROOK]

SupermanCK ~ thx for the update! sorry it didnt work out for ya - i tried

 

[THUGSROOK]

 

[THUGSROOK]

 

[NanoMem]

Thugs, can you please verify on your P4B533 if this anomaly where memory is tricked to the next high ratio above its BIOS entry for the following DIP switch settings:
OFF ON OFF OFF ON OFF
ON ON OFF OFF ON OFF

Note that although switch 6 is activated, as you suggest in this thread, the bus frequencies are below 133. Trying to confirm this correlation between CPU:MEMORY ratio settings and the last switch setting (jumper free mode!!!). As for your PM, the same trick under P4B266 has already been played out when switch 9 is flipped. Switch 10 really has no purpose for the goal of this thread but switch 9 can throw you on either side of the 133MHZ fence!

I am also guessing that the AGP/PCI are getting clocked as if jumper mode was on regardless of their BIOS settings

Peace

 

[ThreeEagles]

Wow! I was happy when my 1.6A oc'ed to 2.4 w/ stock hsf at 1.5v vcore, but now I'm running DDR400, and I would say I'm ecstatic!

Thugs- I had to go to 2.5/3/3/7 to get my machine to post also(I was running 2/2/2/5 before). I am using Corsair XMS3200. Is there a methodology you recommend to trying more aggressive memory settings? Or does order not matter and I can just do them one at a time?

 

[Duvie]

So suming it up....

The corsair I am getting will be hard pressed to run at 442mhz ddr and if it does(hand tied with such low settings)....I only need to really run at 171-172fsb with 342-344ddr to equal it in perfromance....

I think the memory is not worth the sacrifice of cpu power which in a lot of other apps not memory bandwidth happy like Q3a and 3dmark2k1 really wants every extra mhz of juice....


I will shoot for 175fsb 1.71v and 350ddr at cas 2,2,2,5....I obviously could not get memory to run high enough to equal that...


Once again Thugs the research was tremendous and helps many of us spend our money wisely and to see things we can't currently produce for ourselves...

 

[dbwillis]

Nice, thanks for this info...!

 

[Mikki]

Thugsrook, great scores! Thanks for all the testing and info.

Dinglehoser, thanks for testing the -E board!

Oldfart, I was curious about your reaction to all this, glad you approve...

Had a BUNCH of problems with my system over the weekend, that and having to do the weekend family stuff equaled ZERO time to play with this! I'll hit it tomorrow after work (not that it really matters since Dinglehoser already confirmed it).

Thanks again BobbyEyes!

 

[THUGSROOK]

NanoMem ~ you are confused - we are in jumper free mode this time
all you have to do enable the correct dipswitch on your board to get the 3:4 (or 4:5) ratio in jumper free mode. there are ZERO limitations this time - all bios functions work perfectly.

ThreeEagles ~ i was not able to tweak the Cas and remain stable too - CAS 2.5/3/3/7 is all shes got.

dbwillis / Duvie ~ anytime!

BTW: this INFO has made front page @ HOCP!
(i have a mirrored thread over there)

a big thx to everyone that participated in this project

 

[oldfart]

Duvie is 100% correct. 400+ddr is not worth that much due to CAS latency.

That would depend on what you have for a CPU. If you have one that tops out ~ 150 FSB like I and many others do, it means a lot. I'll take DDR400 cas2.5 over DDR300 cas2 any day.

 

[THUGSROOK]

 

[Duvie]

Oldfart, You are also 100% correct!!!

If you can do 400mhz or close to it at the speed your cpu tops out or very near to tops out that is the best performance....


What we need is true cas 2 400mhz ddr....

If I couldn't or can't get anymore then I would or will take the 440mhz ddr if I can attain it!!!


That is the unfortunate thing about good ocing 1.6a's....

 

[THUGSROOK]

 

[SupermanCK]

Wait...I am a bit confused now...
so is it dipsw 10 or 9???

 

[THUGSROOK]

 

[THUGSROOK]

 

[oldfart]

Thugs

Duvie is 100% correct. 400+ddr is not worth that much due to CAS latency.
its a big comprimize that each OCer will have to figure out for themselves. just check my benchmarks.

Sorry, the misquote was not intentional. Its just that I found statement above a little too vague. The "is not worth that much due to cas latency" needs a lot more said about it. Even though you say to check the benchmarks, the way you worded it plays down the importance of a great new option people have who run 133-150 FSB speeds. There are a heck of a lot more people with that setup than ones running 165 - 170+ FSB. For that majority, it is worth it, and it is not a compromise. For the people like you and Duvie with "Titanium Platinum" CPU's, it is.

Anyway, very nice work Dinglehoser for finding this trick, and Thugsrook for all the time benching/testing it out.

Peace