More P5B secrets uncovered

[lowfat]

hmmm, very interesting indeed.

[cadaveca]

No mention of strap/timings...

Well, that's what i meant by next-gen. Faster FSB speeds. And i think we need it. Sometimes ya gotta read between the lines, i guess. It wasn't that they wanted us not to have it because it was working already, more that it DIDN'T work. The chipset revision confirms this, and why in the first place I mentioned the chipsets "sucking".


Anyhoo...



That's a great write-up FCG.

[Ace-a-Rue]

More to read:
http://www.bleedinedge.com/forum/sho...63&postcount=3

-FCG

fantastic work FCG and Tony!

[bing]

Great value & finding !

Vote sticky !

[Gam3Ra]

I don't understand all.. you talking abaout is that, 7x400 and 8x400 is the same performance?

[HKPolice]

I understand the default multiplier issue, but what about on an X6800? I'm assuming that whatever multiplier you set becomes 'default' and no overclocking of the NB core clock happens?

[Kunaak]

I don't understand all.. you talking abaout is that, 7x400 and 8x400 is the same performance?

same or very similar performance, but very different levels of stability depending on the default multiplier of the CPU.

say you have a E6400 that has a default multiplier of 8.

well, you take the default multiplier, divide it by the multiplier you set, and then multiply it by the FSB you set, and then you get the Northbridge chipset clock.

so 8/8 x 400 = for 400 mhz or 1600 mhz on the chipset clock.

but then just lower the multiplier to 7, and stability should change cause your stressing the chipset even more.

8/7 x 400 = 457 mhz or 1829 mhz on the chipset clock.

even though all you did was change the multiplier.

so the default multiplier is the leas stressful multiplier to use, but it's also the most limiting in most cases, cause it's easier to do 7x 500 then it is to do 8 x 500.

upon learning all this, I think now its way easier to spot problems when overclocking conroes, so thanks FCG

[corvus_corax]

Wow, this info is great, thanks for share Tony and FCG

[v0dka]

same or very similar performance, but very different levels of stability depending on the default multiplier of the CPU.

say you have a E6400 that has a default multiplier of 8.

well, you take the default multiplier, divide it by the multiplier you set, and then multiply it by the FSB you set, and then you get the Northbridge chipset clock.

so 8/8 x 400 = for 400 mhz or 1600 mhz on the chipset clock.

but then just lower the multiplier to 7, and stability should change cause your stressing the chipset even more.

8/7 x 400 = 457 mhz or 1829 mhz on the chipset clock.

even though all you did was change the multiplier.

so the default multiplier is the leas stressful multiplier to use, but it's also the most limiting in most cases, cause it's easier to do 7x 500 then it is to do 8 x 500.

upon learning all this, I think now its way easier to spot problems when overclocking conroes, so thanks FCG

Nice explanation also.

Interesting stuff, I'll just keep default multi on from now on all 965 boards.

[corvus_corax]

Yup. So, in confs like this, the NB might be crying (that was without vmods and stock passive/pipe heatsink)
E6600 @ 6x500 = 3.0GHz
NBCC = (9/6)*500 = 750MHz (3000MHz data rate)

[Harshal]

Nice Info.. must try

[aggybong]

Tony, which BIOS did you use? 507 or something else?

[Kunaak]

Yup. So, in confs like this, the NB might be crying (that was without vmods and stock passive/pipe heatsink)
E6600 @ 6x500 =
NBCC = (9/6)*500 = 750MHz (3000MHz data rate)

well, I still have to wonder about things like this...

see, so far, I cannot even begin to guess how things like EIST, effect the chipset clock... cause EIST actively changes the Multiplier in windows.

so I assume, getting 6 x 500 wouldnt be too hard, if you did it from within windows, by adjusting the multiplier with a tool like Systool, or or Crystal CPUID, then clockgening your way up...

I think the NBCC would stay the same... atleast, theres no way for me to test this idea yet... Sisoftsandra... like always, does not detect changes within windows, and will not read anything other then its boot up speeds.

[mrc207]

so does this apply to all 965 chipsets or just the P5B so far?

[dinos22]

good findings Tony and FCG

[TuKo]

Is this true for 975 too ?

[tzitzos]

amazing news tony & FCG

but all this is only for the p5b or works also for the ds3,ds4,dq6 etc?

[rob725]

Tony and FCG: Thanks for all the hard work and incredible explanation.

[Tony]

Guys this is fact for all intel chipsets, 845 and up, they all work the same.
Chipset overclocking has probably caused no end of memory RMA's...I needed to ignite a spark in a few minds as i got testing this P5B, Kris was the first to add some info and come 4am my time we had enough for a second post.

Kris and I both have 6400's P5B Deluxe and 0507 or later bios files flashed, boards are unmodded and the rigs are not special in any way, infact Kris is testing on his multimedia rig so lets just hope all this chipset clocking has not damamged his OS instal

To get back on track, you have to realise with Intel platforms its not just set the multiplier push the fsb and go, there is a lot more to it than that. I also hope we have shown why Intel platforms are actually quite bad for memory reviews, the chipset strap and core clock has HUGE effects on stability as well as memory performance, infact some of the reviews out there showing dimms not doing rated speeds...im sorry but the chipset was crapping out and not the ram, you need to retest!!

We are looking to find the latency settings next, and see if we can force latency changes.Second to find is what effect forcing strap changes (like BSEL manipulation) has on the chipset clock etc, we all know a lower strap increases core clock but what else does it do. I am sure all this could be added into memset at some stage but what we need is it available from bios. I feel a huge begging email being sent to Oskar asking for a ton of options on his next Intel board...lets hope he comes thru

I will need to update my initial guide with the new info we found as we now know NB Core clock and how its found. If any of you guys what to add some to the article get posting, either here or on BE, i will add it in the mix and credit the authors as needed. We need to crack intel overclocking wide open so that everyone has a good idea of what is going on and why.


T

[Gam3Ra]

Sorry, but my english is very bad.. I read the topic. 400FSB 1:1 with the memory is faster that 401 1:1, but who is better.

[Icer]

Good info to know if the P5W DH bios ever allows us to change the multiplier.

[Tony]

Sorry, but my english is very bad.. I read the topic. 400FSB 1:1 with the memory is faster that 401 1:1, but who is better.

400 1:1 is faster than 401 1:1 due to an internal latency change happening at 401fsb.

[SIOUX]

Very good reading guys

Though I dont see any BW decrease from 400 until 485FSB 1:1 9 defult multi, only BW gets higher (as it should) all the way to 485. Infact i get the BW decrease after 485FSB. Then it changes dramaticly to REALY low numbers like 300mb/s in sandra. Only latency seem to improve after that. dont know why though .

[Kunaak]

I have always wonderd, why doesnt my DS3 boot up at 6 x 450, and struggles with 7 x 450, but does 8 x 450 with ease.

last night, after reading this, and learning the math on it... and testing alot, I find the bios seems to have a limit, of 499.
after 500 mhz exact on the NBCC, it will rarely if ever boot.
which I only found out last night from testing multipliers and FSB.
so it seems there may be actual allowable limits in BIOS, that hold various boards back abit.
like maybe gigabyte doesn't want this board at higher then 2 ghz on the NBCC...

its a very interesting thing to learn about, cause now I can more completly see where problems come up now, without blindly wondering if the board or ram are the problem.

however I have a interesting Question as well...

how does EIST effect the NBCC?

if you boot up at 8 x 400, then EIST takes the multiplier down to 6X during Idle, does the NBCC change, or stay the same?

if it stays the same, regaurdless of multiplier, then the key to overclocking really high, for say a E6700, would be to use clockgen in windows, while booting up with the default multiplier, and adjusting the multiplier from within windows...

assuming of course the NBCC stays the same, once actually in windows, and doesnt change with the multiplier.

[freecableguy]

Changing the multiplier in Windows also changes the NBCC on the fly. However, as far as I can tell EIST does not change the NBCC when temporarly throttling up and down on the CPU speed. Because the set multiplier really does change in the same sense as when it is manually selection via software/BIOS. That's about the best explaination I have so far. EIST is a completely different beast than manual multiplier selection and it shares about nothing in common with what we see in BIOS.

-FCG