A general thread to throw your current clocks and such in … and for us all to help each other out etc
As john is firing up the bios to battle with the i5 750 along with me … this may help each other out a bit
vcore - 1.2250
vtt(IMC) -1.162
cpu pll 1.8
ram at specs and 1440mhz
speedstep and turbo etc all disabled
llc enabled
screeny on completion of 10 IBT runs
EDIT for correct volts
After recent tinkering … thought to go back into it and see whats whats with a fresh approach and more data on the net …
this is very interesting.
After that run, have gone back to my 3.8 and looked at refining it … dropped it 2 notches on vcore while giving imc a boost ( but only to 1.2 compared to 4ghz 1.23).
Yielded a temp drop of 4 degrees on hottest core to 68 in linx … and trying to further that improvement now … i think i may have finally found my imc sweet spot
thats the SS of the above post … tried to trim it by 1 increment in vcore … and it failed after 9/10 linx runs … dammit lol. Was 66 degrees max core too.
Still a great result. IMC (QPI for GB boards) can be a funny thing, I was even able to lower my QPI/IMC voltage lower by increasing the QPI PLL voltage a bit which had my OC stable at a lower overall votlage across the entire board.
It is a balancing act, there’s always something else you can tinker with that will get you a different result.
Temps have dropped around 10 degrees which i would guage to be on the money so im confident i have a good block mount … might do it again just for curiousity at some point.
4ghz i a happy 24/7 … wonder how much further i can manage heh … only 1 way to find out
Nice work, finally cracked the magic number.
Your temps look good for the system you have used. Looking at the previous screenshot, jumping up to 1.304V from 1.224V whilst keeping similar temps is a satisfying result.
The real test will be to see what it does during summer.
Well i’m feeling left out so i did some voltage tweaking and hit the big 4.0 temps are below 80c with a room temp of approx 19c which isn’t too bad…
been reading and thinking about the gap in my core temps … core 4 being the prick that climbs a good 5 odd degrees above all others.
So i reseated the cpu twice … and tried different paste applications ( a grain of rice, smear the base and my own of a grain and a half) while making sure i evenly tightened the cpu mount studs …
After it was all done and a 10 run linX pass for each ( last with my own method didnt get another reseat as it did fuck all lol) i couldnt correct that temp …so its just the way my cpu/socket works i guess. Did find that my own appication is pretty damn good… as it yielded the best temps by the 1 degree over grain and 2 degrees over the smear. Others may find their method is best … but im finally 100% on my application and mounting ( which was the same as i ended up doing with the d14 too) … no excess tim, but good/ even coverage. So, while i didnt cure the temp wandering, at least i managed to confirm something … and got the tinker lol… been a while since tinkering didnt cost $$$ heh.
On another note for my future tinkering, really eyeballing a x3450 ( 750 with HT). While you never find a definitive post saying “xeons are higher binned retails” …that seems the general consensus … so that means a chip that handles less volts theoretically … but im not fussed… the ht is the last thing to grab so its looking the goer. Going for a 3460/70 seems pointless as its just a matter of bios fighting to get the desired / max clock. Of course the multis is what your paying for … but the 3450 comes with a native X20 so no sweat there.
Also been really getting keen to sleeve the cabling. Will bring the innards up to a nice freshness and look 10 times tidier. Still figuring out the amount i need , but from reading its looking like a good 100 foot covers it … but i dont think id do each wire but do pairs to limit time. Still an expensive thing to do just for looks … but it will be fun and at the end i reckon i’ll be chuffed with the results.
These of course wont happen for a bit as i like havin nuts and the wife would surely cut them off if i were to drop another few hundred on the pc just yet lol
The biggest problem with the grain of rice method is that we all have different pictures of how large a grain of rice is. I always use a bit more, my application is probably closer to the size of half a pea than a grain of rice. There is a youtube video that shows the application of thermal paste on a CPU, using a flat piece of perspex to show how it spreads. It’s very interesting. I’ll see if I can find it again when I have a bit more time.
I’ve done the same thing, I reseated my U12P about 20 times trying to get consistent temps but I never could. I tried every method that claimed to be better than the last. This is why people lap the surface of the CPU and HS, to get the flattest contact possible. It is something that I would only ever consider doing with an old chip as you would hate to botch it, haha.
Sleeving would be wicked fun, I also want to do it one day.
Great results there mac, the CM V8 does a sweet job. Just make sure to run Linx with all available RAM as it will be a better indication of whether the system is stable.
i’m sure with all Cpu’s I’ve had over the years there’s always been a difference in temps from core to core (q6600 etc)…nothing to worry about just one of those things.
As far as fitting coolers etc i’ve always gone with the add a blob to the cpu and spread with my finger covered in glad wrap always done the trick for me…
been doing some minor tinkering while juggling the kids … got 215x19 for 4.09ghz … havent benched it … this w/e when the cars arent goin round ill tinker more … see what my true bclock wall is for this board and see how high it goes lol
heh … nah … did it on 1.212 … 2 notches up from 1.200 which i was running.
I read that you need to give it maybe .1 to gain high bclock settings … and i am thinking that its much like vcore … relevant to what your base is … mine was 1.16 … so i thought …its like vcore … and just needs that similar amount ( not the value) … and its seeming to work.
Those with 1.3+ vtt … whats your base ? … ie 1.2 … 1.25 ?
one way to check is to run everest ultimate ( lastest version) and it should pick it up … mine does… but my asus board shows the live values in my bios anyway.
Heh … its funny how things go.
Here i am trying to squeeze the life out of my cpu … and yet ive already got 4ghz from a 2.67ghz cpu… massive. Funny how “4ghz” is the morn for a cpu oc these days lol.
I will say though … i really think that asus boards must differ from gigas with voltages etc … as near on all forums are revolving around gigas/evga etc… little asus floating around that talk about their clocks … yet i seen on ocau that ausus have a overclocking comp on atm… and fuck me are there some big clocks going on … as in highest bclock, lowest frequency lol… madness… i just need to find out how their imc, pch,etc tec all relate
Depends on the board too, my EX58 does not show up QPI/VTT settings at all under any monitoring program, including everest.
Not that it’s entirely relevant due to the different chipset, I think I’m running around 1.26V - 1.28V on the QPI(IMC). I can’t remember exactly what my base vcore was, but it was around 1.298V with HT on and Turbo enabled. With HT off, I think the base automatically drops to around 1.260V.
Haha, yeah it’s pretty funny. 4.0GHz seems to be the minimum to aim for these days.
Interesting observation in regards to the motherboards. Max clocks certainly can differ from board to board, but there is far too much variance in the CPUs to know if it is the board doing the magic or whether it is a good CPU.
For example, the system I built for my friend with the same motherboard that you are using (with a i7 860) was not able to hit the same clock speed with 1.20V on the IMC, needed 1.3V right off the bat. In saying that I only had a hour or so to get it relatively stable (purely LinX stressed, no Prime95 due to time constraints). I’d still put it down to the CPU being the limiting factor before the motherboard at a standard 4.0GHz OC.
There are some stand outs which I have witnessed, the Classified boards on the 1366 socket are heavy hitters and consistently hit +200MHz base clocks which is where my board starts to suffer. Problem is for most clocks needing more than 200MHz as a base, the CPU will need far more than water cooling to keep it at reasonable temps.
so while i havent done any major over clocking i couldn’t find anywhere else to put this post
yesterday i decided i’d have a good at using evga’s e-leet in OS OC tuner and for the most part it works i got it to 3.6 ghz without restart, going into the bios or raising the vcore (stock 1.1875v) and it was stable after running intelburntest.
after trying to get to 4 and failing i reset my bios to default to have a clean slate to start with and thats where i ran into my problems
i turn on my PC this morning and check the temps and its running at 40 degrees idle, i was like WTF considering its been running about 25 idle. it turns out that the MB had turned voltage contol back to auto and was giving the processor 1.257v which is clearly higher than stock and not needed what so ever considering hte processor isnt OC’d yet.
so after poking around the BIOS and changing setting ive got this
The thing im rly happy with is the voltage that its running at (and stable) it gives me a good chance of reaching the magical 4 mark on decent voltages and temps :yahoo: