Overclock.net › Forums › Intel › Intel CPUs › Overclocking a C2D E6300 Conroe
New Posts  All Forums:Forum Nav:

Overclocking a C2D E6300 Conroe - Page 2

post #11 of 23
@ the OP - yes that was on air cooling.. and with cheap Geil PC8500 Black Dragons. wink.gif

*clocking the 965p is weird.. you must understand where the NB latency resets are... otherwise you will be continuously fighting for lost performance as you increase FSB.

ie- there is a NB latency reset that occurs at 401mhz FSB... To see it.. run a latency test at 400mhz FSB.. record results.. and raise FSB in BIOS +1mhz.. re-run test. You will see with only +1mhz you will have lost around 12-15ns of latency. eek.gif

My advice - if you have RAM that will not run +1000mhz.. either run 400mhz FSB with RAM 1:1 and tight timings.. or you will need to jump all the way to ~465-480mhz FSB [this time pushing RAM speeds, still on 1:1 divider]. Your overall latency will be about the same at both settings... only CPU speed can scale higher obviously due to FSB increase.
post #12 of 23
Thread Starter 
Quote:
Originally Posted by CL3P20 View Post
*clocking the 965p is weird.. you must understand where the NB latency resets are... otherwise you will be continuously fighting for lost performance as you increase FSB.

ie- there is a NB latency reset that occurs at 401mhz FSB... To see it.. run a latency test at 400mhz FSB.. record results.. and raise FSB in BIOS +1mhz.. re-run test. You will see with only +1mhz you will have lost around 12-15ns of latency.
Yeah, I've read the the P5B has a couple of those points - something to do with "straps", maybe? I read about it here, though I'll need to re-read and apply it before I understand it fully.
post #13 of 23
Thread Starter 
Runs stable at:
400MHz FSB
RAM at 800MHz 5-5-5-15 stock
vcore @ 1.3250
PCI-e @ 100, PCI @ 33.3 (we'll just assume this from now on)
Everything else on auto

When I bumped the FSB to 420, I got rounding errors on the first pass in Prime95. Increasing the vcore caused it to hang during boot.

Would tightening up my RAM timings help? The SPD reported 4-4-4-13 as an option. Memtest86+ reported no errors. Is there a way to unlink the RAM from the FSB, so I can lock it at 800MHz?
post #14 of 23
Your most likely going to need more NBv [or max setting at 1.65v] to run 4GB's of RAM at that FSB on this mobo. And no.. Nvidia mobo's are the only ones with unlinked RAM dividers. Your stuck at 1:1 unless you get better RAM.

*try using the same FSB just with a lower CPU multi.. so you can eliminate your CPU from being unstable.
post #15 of 23
Thread Starter 
Also, should I be concerned if I change a particular value in the BIOS (vcore for example) and the selected value turns purple or red? Some of the settings recommended in this thread have caused this.
post #16 of 23
no.. your overclocking

1.45v FSBv for +500mhz is safe on air for 65nm
+1.45v vcore on air is safe for 65nm
MCHv depends on chipset and how much RAM your using.. in your case.. [older chipset+4GB's RAM] = 1.5-1.7v MCH/NBv depending on latency and FSB... this requires active cooling [ie- a fan] on the NB
post #17 of 23
Thread Starter 
I tried changing the multiplier to 6. Prime failed.
I changed the multiplier back to 7.
I tried bumping the vcore and NB up one tick. Prime failed.
I tried bumping the vcore and NB up another tick. Prime failed.

This is what I'm seeing (both cores eventually fail, I just grabbed the screenshot as soon as one did, so the readings would be accurate):

post #18 of 23
Your running.. what 1.35 or 1.4v FSBv?
post #19 of 23
Thread Starter 
1.3250

Is it normal for the core voltage reading in CPU-Z to be off?
post #20 of 23
no its not.. are you using the latest version? I have owned many P5B's and they never had this issue..

have you tried to read voltage using Asus PC Probe?
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › Overclocking a C2D E6300 Conroe