Core i7-965, Rampage 2 Extreme & 4870 X2 CrossFireX

[s e t h]

what memory is in the faster cpu systems...i saw like 200mhz difference just switching memory.
maybe there are some imc things going on

[Sampsa]

Looks like the one we tested earlier in Stockholm was:

3830A417 (worst)

and this one is: 3830A432 (bad)

[Tony]

what is this i am hearing the RE2 is not clocking as well as the P6T?

Specifically with memory

[giorgioprimo]

I have to agree e8600 + Rampage Extreme is very exciting combo, still good for most of 3DMarks and I have a lot to do in hwbot

+1

[k|ngp|n]

I have to agree e8600 + Rampage Extreme is very exciting combo, still good for most of 3DMarks and I have a lot to do in hwbot

I am on my way to the post office right now for you and sf3d. More fire power on the way guys .

[EMok1d08]

For this whole system right here, how much would it cost together, if like you are just having it as a computer, not with the OC stuff?

[T_M]

Dont be so quick to dismiss chips as being bad.
Multipliers 30x and above can do funny things Also check that you aren't pushing too high on uncore and QPI speeds.

[dpa]

Dont be so quick to dismiss chips as being bad.
Multipliers 30x and above can do funny things Also check that you aren't pushing too high on uncore and QPI speeds.

correct; but some chips cant do more than 29x.. (like the one we tried in Sweden)

on the 965ES chip I got now doesnt like over 28x in vantage at all! therefore: test 27x!

sweden chip: 3830A417
"norway" chip.. lol :3831A457

[T_Flight]

I saw something in another thread about 0503 BIOS for RIIE. Just a heads up in case you guys don't have it.

[Rol-Co]

with new 0503 bios multiplier is not a issue, can do x42 133 easely and even boot at 45 x100, maybe its bclock related but can't test that aircooled.

3831a are a lot better like i said already, have a proper true cooler on it now and can run about everything stable @ 4200mhz 1.39v

like this

[T_Flight]

4.2 at just a tick below 1.4 is a beautiful thing. I hope I'm lucky enough to get a chip like that. I have water cooling I'll be putting on it too after I run and test it on stock air first. This 1.5v stuff is a little scary for me personally.

[SF3D]

I want to join this "bad ES chip club"

I just finished my 4 hour test session with my other 965 CPU. It can do 4.65GHz max on LN2. It was doing the same with dry ice

Light benches can be done at 4.7GHz, but that's it

Cold bug -70c
Batch: 3830A432

I wish, that my other chip is better.. the batch is the same so no high hopes here.

Kind of frustrating...

[Rol-Co]

what is this i am hearing the RE2 is not clocking as well as the P6T?

Specifically with memory

Tested both board here and no difference if you ask me.

sf3d
use that chip for dreamhack contest please holland take this one

[ryba]

Same situation here.

I test on P6T Deluxe. culd bug about -80.
But sweet point for max freq is around -65 - (-72).
Vcore higher than 1.5750 is not help me.


I have 2 batchs:
3830A408 - average
3830A417 - bad

I guess all ES CPUs are this two batches.

It's not good info I have the same batch as yours bad cpu, damm...

[dpa]

with new 0503 bios multiplier is not a issue, can do x42 133 easely and even boot at 45 x100, maybe its bclock related but can't test that aircooled.

3831a are a lot better like i said already, have a proper true cooler on it now and can run about everything stable @ 4200mhz 1.39v

like this

thats correct, I have an 1a also.. does almost 4.85Ghz 06 cpu test.

EDIT: ok, lets get it straight. there is two "steppings" for ES it seems like. we have the 3830A and we have the 3831A

the 3830A clocks like , while the 3831A clocks fairly good.

"intel sweden chip": 3830A stepping, did MAX 4.55Ghz vantage/06 cpu test. coldbug at about -80. didnt like multi over 29x
"usa chip": 3831A stepping, does 4.85Ghz 06 cpu test so far. coldbug at about -95. I can do 40x with this one....

summary: 3831A is the thing to get? stasj your 3830A in a daily machine on AIR!

[Rol-Co]

Finnish a 3d2005 run @ 4312mhz 1.39v true cooled. pretty neat if you ask me. room temp 22 degrees, 1 sanyo denky 35 mm fan on it.

[T_M]

my best 965 so far is a 3831a460 that does 4.9ghz 32M and CPU-Z @ 5270mhz

[Rol-Co]

my best 965 so far is a 3831a460 that does 4.9ghz 32M and CPU-Z @ 5270mhz

cooling and temp ??

[dpa]

my best 965 so far is a 3831a460 that does 4.9ghz 32M and CPU-Z @ 5270mhz

tested 06/vantage?

[dinos22]

outch so many crappy CPUs

looks like KP's got the best chip so far.........not sure how far charles can push his cherries in Vantage hehehe

[T_M]

Have been less interested in 3D performance testing so far, ever since i saw charles killing 32M
But certainly for a 4.9 32M chip the multithreaded tests will be far less (4.6-4.7?).

Using Ln2 @ -57C coldboot and running at -65C to -70C benching 32M (it could run same clocks at -50C too....)

[Sampsa]

I managed to add Kingpin's F1 Dragon EE mounting bracket to my dear rotary cascade's evaporator.



Testing contact between evaporator and IHS.



The result was VERY bad



Lapping IHS



Hmm... also evaporator was lapped and contact is now perfect.



Insulating around Asus Rampage 2 Extreme's LGA1366 socket with Armaflex.





Evaporator insulated and mounted on Core i7-965.



Unfortunately -110C is too cold to this 3830A batch CPU. Cold boot issues, BSOD and Vista not loading

Anyway, I'm ready for Core i7 so now I just need to get good chip

[Pt1t]

This one boot @ -100° but no as good as 3837A

and it needs to be lapped.

[k|ngp|n]

There are ALOT of issues with the chips preventing them from being run hard like yorkfield or wolfdale. The retail chips seem to be different from the few 3.2 ES's I got the chance to try, but not by much. Maybe with more bios revisions on some diff mobos, we can start to see whats holding what back. I get varying performance on the same chip depending on what mobo it is testd on atm. Some scale with volts, some don't. Some like cold, some don't. Some like high multi's, some don't...uncore, bclock etc etc.
Despite the seemingly low percentage of "good extreme overclocking" cpus, the nehalem arch. is capable of alot more than what anyone posting actively has posted, believe it. Fact: 5400mhz+ 32M run..5200mhz+ 2k6 .
Prolly needle in a haystack at this point with this revision, but still possible.

[T_Flight]

Testing contact between evaporator and IHS.



The result was VERY bad

Owweee. I don't wanna cast a cloud on Intel's moment of glory, but that really is bad. That is concavity right in the middle where the cores are.

I don't wanna be captain obvious here, but that is pretty bad. It would be nice if Intel would true these up before the laser ethcing process. Not even air can cool properly if it doesn;t have a proper TI.