Great expectations for 2.5 not met

[goatzool]

The P4 2.5 seemed promising in some early reviews as an economical stepstone to the range beyond 3.0, but I see a pattern to the contrary developing.
Of the 6 experiences listed in the CPU database, only 2 have made it beyond 3.0 (which is an FSB of 120) and only with expensive cooling equipment. I got mine today and it is running steady at 3.0 and a notch above, but it will not make the jump to 133. At 122, the PCI bus becomes unstabe because my ASUS P4S533 does not have a PCI/AGP lock.
I had a 1.8 on this identical system this morning and it went to 147 fsb without complaint at 1.65 Vcore. I have tried this 2.5 at the max Vcore of 1.725 and it just won't boot. I may try to wrap the pins, I am water cooled and hopefully less likely to experience sudden death.
At least I am seeing what 3.0 looks like, and folding rapidly.

 

[markodude]

Are you Prime95 stable at 3.0ghz? I could only get 2.8ghz on both an Abit BG7 and Asus P4B-533, thing is the BG7 did stable with default voltage, and the P4B needs to be set to 1.65v (it drops to 1.47 and undervolts everything)
I cant get over 2.8ghz to be 100% Prime95 stable no matter what voltage I put through it on this cooling (up to 1.825v anyway) Im reluctant to pin mod it as I dont have water cooling.
Im getting an XP2400+ today so its on the back burner for now and running happily at 2.8 until I get water/prometia, or another P4 (2.26 C1 is appealing)

 

[goatzool]

Stability

Hi, Markodude,

Since I use my computer for my work, and I only installed the 2.5 yesterday, I don't have enough experience to say it's stable. I ran 3dMark2001 looped for 10 hours beginning last nightwithout an error, and Sandra burn in fora few hours with no problem, but I haven't tried prime95 for 24 hours yet.

I do notice that my 3dMark scores bounce around at an fsb of 120, ranging from 11,900 to 12,850 without anything else being changed except a few degrees of temperature. I think that what instability I have there is attributable to the PCI bus running at 38 (no PCI/AGP lock on this board) rather than the CPU itself. If I could get to 133 fsb, the divider would bring me back to 33.

I went back and read Mr. Natural's new vid pin mod technique, using conductive paint to bridge the pinholes and achieve a Vcore of 1.8 and above, depending upon pins bridged. I have an externl water cooling setup and so heat is not a problem, I never go above 38C. I'll be hitting the auto parts store today for that paint. Must get to 133 and beyond!!

 

[larva]

It's not the PCI

You would think out of spec PCI was responsible for any and all oc'ing limits to listen to people talk. I run the P4S533 at 37.5MHz PCI speeds (150fsb) and it's fine. Get a board with the lock, like the P4B533 and you will observe the same 3GHz limit for your chip even with the PCI locked at 33MHz. 37.5MHz is not responsible for your lack of further clock speed capability nor the variation in 3Dmark scores your are experiencing. 44MHz PCI speeds have proven workable for all but the most sensitive components (in systems where the cpu is whithin its limits at these speeds).