920 D0 #3909A163 4000MHz@stock vcore

[saaya]

on water, but i doubt that makes a big diference as we all know i7 doesnt scale that well with lower temps

bclock 190
uncore multi 12x
mem multi 6x
ht is on btw and vcore is 1.25v

...thats it
ahhhhh i7 ocing is so uncomplicated compared to 775

ive heard of some chips doing this with 1.1V
is there a new bios that allows lower than default vcore yet?

[saaya]

4100 failed :,(
bumped up vcore to +40mv and it passed
lol, tried again with +20mv and it passed... so 0.02v is all it needed to pass at 4100
lets hope it continues to scale like this! 100mhz per 0.02v

4100 +40mv:


4100 +20mv:


EDIT: daymmm multi dropped to 20x?
didnt see that... temp related? hmmm

EDIT2: nope, not temps... bah i forgot im on P03 on this board, need to update the bios to get the current feddback override option

EDIT3: alright, stable!
but temps went up to 83... and thats with all fans at 12V and an water ambient delta close to zero (36C water)
thats too much... damn this cpu is leaky... it must be great on ln2... but i dont give a damn, i want a good air/water chip! gah!

[SoF]

maaan you need to make it cold seems to be a quite good chip. well bring it with you when in germany, we (finally) make bbq and grill that chip too

[saaya]

dammit... looks like a high leakage chip... above 1.25v the tdp explodes... either that or the apogee gt cant handle more than the tdp i get at 1.25v...
ill try another block and check... sigh... all i wanted is 4.6 on water, is that asking for too much? :P hehehe

sof, i dont care about ln2! :P
maybe i can find somebody with a bad ln2 clocking chip that does well on water and we can swap ^^
i think victor has a 4.6 water chip that only does 4.8 on ln2... this one probably does 5.2G at least on ln2, so maybe we can swap em haha

[The-Fox]

lol, this time the users at XS will give you support for foxconn products

Have you flashed to G25 Bios ?
it's the best IMO (unless you have Perc 5/i Raid controller )

Make sure you enable Current override limit in BIOS so that the multy will not drop
oahh, and set the turbo to "allways on" too.
oahh, and get a hold of unclewebb's i7 turbo tool to see if the Turbo is dropping, it's the best app ever to display the multy in i7.
It uses Internal high performance timers (inside the CPU) to read the multy.

BTW, nice CPU.

[Mescalamba]

Funny that I saw 4,6 on air. My chip is probably ok.. air cooling is able to make it survive 1.6V .. tough, I'll not know max without better cooler or water.. Still C0.. so I don't expect much..

Btw. better block can help.. tough I would think probably only about HK 3.0 LT ..

Yea, maybe they don't scale with temperature, but I think that lower temperature helps achieving higher OC without adding much voltage.. but, just theory..

[p0opstlnksal0t]

had my d0 to 4.2ghz on an OK air cooler. 1.36v but it failed prime after about 6 hours temps got up to 96! lolz i let it run overnight

[saaya]

lol, this time the users at XS will give you support for foxconn products

Have you flashed to G25 Bios ?
it's the best IMO (unless you have Perc 5/i Raid controller )

Make sure you enable Current override limit in BIOS so that the multy will not drop
oahh, and set the turbo to "allways on" too.
oahh, and get a hold of unclewebb's i7 turbo tool to see if the Turbo is dropping, it's the best app ever to display the multy in i7.
It uses Internal high performance timers (inside the CPU) to read the multy.

BTW, nice CPU.

hehe, yeah funny idea isnt it?
but it has always been a give and take, without a bunch of really nice people here giving me feedback i couldnt have done any propper support

already flashed to G25 and enabled current override, thats the last 4.1 with +20mv screenshot i posted in my post above
but like i said, temps go crazy...

unclewebbs i7 turbo tool? im using cpuz and thats fine... and yeah it reads the internal hpet inside the cpu as well, doesnt everest etc do that too?
wheres that tool? can you link me?

Funny that I saw 4,6 on air. My chip is probably ok.. air cooling is able to make it survive 1.6V .. tough, I'll not know max without better cooler or water.. Still C0.. so I don't expect much..

Btw. better block can help.. tough I would think probably only about HK 3.0 LT ..

Yea, maybe they don't scale with temperature, but I think that lower temperature helps achieving higher OC without adding much voltage.. but, just theory..

you have a good c0... i always thought d0 are better and pull less power, but they arent at all... i asked around and a lot of people told me a lot of d0 chips actually run hotter than c0 chips and overclock worse on air and water.

yes, you can get higher clocks at lower temps with the same vcore... but you need notably lower temps for that, it doesnt nearly scale as well with improved temps as a64 or phenom2 or core2.

had my d0 to 4.2ghz on an OK air cooler. 1.36v but it failed prime after about 6 hours temps got up to 96! lolz i let it run overnight

that poor poor cpu
i think my chip is about the same...

[p0opstlnksal0t]

you have a good c0... i always thought d0 are better and pull less power, but they arent at all... i asked around and a lot of people told me a lot of d0 chips actually run hotter than c0 chips and overclock worse on air and water.

yes, you can get higher clocks at lower temps with the same vcore... but you need notably lower temps for that, it doesnt nearly scale as well with improved temps as a64 or phenom2 or core2.

I completly agree. Ive seen many c0 chips run less volts at the same speed as my d0 on air... think its more luck of the draw here.

that poor poor cpu
i think my chip is about the same...

lol these chips are monsters. one night I set my bios to shutdown the system when cpu reached 100 (for failsafe)... so I took voltage up to 1.45 and OC'd to 4.4ghz ran prime and in about a split second temps were past 90 and climbing. it maxxed out to 100 in realtemp and stayed there for about 15 mins before my pc shutdown. I really have no idea what the cpu temp actually got too. I think it would be pretty tough to actually ruin one of these chips

[Sch0cK]

Up to 4100 i am on stock 1,248V / 1,255V measured, only +120 VTT @ 66°-70°C @ Prime, HT + Turbo on.
With 4,6GHz i hit over 80°c on Benching.

No, atm there is no Bios with lower V than VID.

[zalbard]

EDIT3: alright, stable!
but temps went up to 83... and thats with all fans at 12V and an water ambient delta close to zero (36C water)
thats too much... damn this cpu is leaky... it must be great on ln2... but i dont give a damn, i want a good air/water chip! gah!

Did you try using lower PLL? Some people run PLL @ 1.3v or so instead of 1.8v and get better stability and lower temps.

Edit: just realised it might be not possible to undervolt on Bloodrage... IF not, that really needs to be fixed in future BIOS revisions.
Source: EVGA forums
Some people also undervolt IOH but there is no point I think, is there?

[saaya]

lol these chips are monsters. one night I set my bios to shutdown the system when cpu reached 100 (for failsafe)... so I took voltage up to 1.45 and OC'd to 4.4ghz ran prime and in about a split second temps were past 90 and climbing. it maxxed out to 100 in realtemp and stayed there for about 15 mins before my pc shutdown. I really have no idea what the cpu temp actually got too. I think it would be pretty tough to actually ruin one of these chips

yeah i set it to 100C max too, but seeing above 80 load temps really pssed me off at 4.1 and i stopped testing... im kinda dissapointed... i put up quite an effort to cool this glowing piece of silicon and it seems to be futile....

Up to 4100 i am on stock 1,248V / 1,255V measured, only +120 VTT @ 66°-70°C @ Prime, HT + Turbo on.
With 4,6GHz i hit over 80°c on Benching.

No, atm there is no Bios with lower V than VID.

nice
what settings did you use for 4.6? all cores plus ht?

no bios with lower vids, meh :/

Did you try using lower PLL? Some people run PLL @ 1.3v or so instead of 1.8v and get better stability and lower temps.

Edit: just realised it might be not possible to undervolt on Bloodrage... IF not, that really needs to be fixed in future BIOS revisions.
Source: EVGA forums
Some people also undervolt IOH but there is no point I think, is there?

lower pll voltage for lower temps? i find that very hard to believe...
but ok ill try...

undervolting IOH makes sense, especially if it runs very hot with the stock heatsink and all you want is stock speeds or slightly overclocked speeds...
since x58 runs very hot, id say around 30W stock, im not surprised that most cooling solutions cant keep it cool...

[zalbard]

lower pll voltage for lower temps? i find that very hard to believe...

Overclocking PCI Express bus to raise the bclk cap doesn't make much sense either imo.

[Mescalamba]

Overclocking PCI Express bus to raise the bclk cap doesn't make much sense either imo.

Eh, actually it does, cause it has something to do with less-then-needed dividers for PCIe clock.

Simply, it wasn't build to go over 220 BCLK, so there are no dividers for that speed.. but if you increase PCIe clocks, you can probably use those dividers that are already in system.. maybe it will be fixed in some next chipset, maybe not..

That's what I heard and increasing PCIe clock really helps.. tough it can destroy HDD and such nasty things.

Btw. QPI multi will be on 16x in P55 chipset CPUs .. but theres question.. will they add more dividers for PCIe or not? If not.. no 250 BCLK.. (xcept L2N/DICE).

[saaya]

Eh, actually it does, cause it has something to do with less-then-needed dividers for PCIe clock.

Simply, it wasn't build to go over 220 BCLK, so there are no dividers for that speed.. but if you increase PCIe clocks, you can probably use those dividers that are already in system.. maybe it will be fixed in some next chipset, maybe not..

That's what I heard and increasing PCIe clock really helps.. tough it can destroy HDD and such nasty things.

Btw. QPI multi will be on 16x in P55 chipset CPUs .. but theres question.. will they add more dividers for PCIe or not? If not.. no 250 BCLK.. (xcept L2N/DICE).

i dont think there are dividers in the chipset like 2:1 etc... its more of a buffer thing i think...

data arrives on one end and since pciE and qpi dont run synced, there has to be a buffer to hold the data until the next bus is ready and the next package of data can be sent. this buffer is only so big, and since they didnt expect more data than 4ghz qpi aka 8GT/s on the qpi side, the buffer probably gets filled before data can be sent to the pciE bus...

something like that i think...
and its not new at all, increasing pci clocs and agp clocks helped to reach higher fsb speeds on older chipsets

[Sch0cK]

nice
what settings did you use for 4.6? all cores plus ht?

no bios with lower vids, meh :/

Jep, everything on

GTX 280 Phx off with HK GPU-X2 and 1x Trippel + 2x Singel-Radi

Vcore under VID would be really help full because for 24/7 @ 3,6GHz 1,255V is to much.

[saaya]

daym thats a nice cpu you got there!
thats a good 500mhz more than what my chip does...

[Sch0cK]

Jep, GPU too

Will try to get more than my 224 Bclk, but more than 4,6 is going to hot for me.

I`ve nerver hoped that i could bench with 4,6 on water .
So i am very pleased with my BR + D0 + Apogee GT

[jcool]

on water, but i doubt that makes a big diference as we all know i7 doesnt scale that well with lower temps

That's a joke right? Most i7's love cold. My 975 needs 0,05-0,06V more Vcore to be stable on water compared to SS at the same clocks. Not to mention I can't get 4,6 stable on water, more like 4,2...

[fatguy1992]

That's a joke right? Most i7's love cold. My 975 needs 0,05-0,06V more Vcore to be stable on water compared to SS at the same clocks. Not to mention I can't get 4,6 stable on water, more like 4,2...

Thats not that much less, my Q6600, now its a chip that loves the cold. My 4.0GHz settings on air, when put under SS are good enough for 4.5GHz Thats ~1.525vcore on air, on SS I need ~1.325vcore.

[saaya]

That's a joke right? Most i7's love cold. My 975 needs 0,05-0,06V more Vcore to be stable on water compared to SS at the same clocks. Not to mention I can't get 4,6 stable on water, more like 4,2...

50-60mv equals what temperature diference between water and ss?
water is 65C load i guess? and ss is what, -30C load?
thats almost 100C temperature diference and it equals 0.05v at the SAME clockspeeds... to me thats not good temperature scaling...