I am envious of those voltages. I am stable at 4410 but my best vCore so far is 1.44 though my temps are about the same as yours.
I may try 4.5 after I get some additional cooling but much past 1.44v makes me nervous.
Good oc - congrats.
Printable View
I am envious of those voltages. I am stable at 4410 but my best vCore so far is 1.44 though my temps are about the same as yours.
I may try 4.5 after I get some additional cooling but much past 1.44v makes me nervous.
Good oc - congrats.
3839A595 i7 920
VCore 1.30
VTT 1.30
HT On
Turbo Off
I'm lazy to rerun Linx 10 Times again I forgot to screen it here it's with 3 Full Runs
http://img35.imageshack.us/img35/126/247intel.th.png
i'm getting a 3845B071 to test against my 3911A200, anyone have any info on these steppings? I think on paper the 3845B should be better?
All i've tested so far is the 3911A on stock air which does 21X175 @ 1.165v for prime with HT on, which i think is ok but nothing special?
From my understandings the 3845b is pretty much one of the best you can get.Quote:
Originally Posted by barrymaldon
OK, I have 2 i920's with identical batch numbers guys, one of them can do 4GHz with 1.264v Vcore, but the other one needs 1.315v.
I really don't know why people try to get CPU's from specific batches anymore, as it would appear that there are lemons in every batch.
I have seen people with chips from the same batch as my last one, and it was extremely poor, yet they are running them at 4.2GHz+ with nice voltages and temps... I just don't get this whole batch thing at all..
Well, I've seen 4 others with the same batchnr as me, and they all do about the same clocks at +- 0.02volt.Quote:
Originally Posted by BLunTSmoKeR
There are exceptions ofcourse. But a certian batchnr has proven to clock much better than others. I havn't seen many 3845b or 3849b doing badly.
Im glad you got it going faster, Im the guy who sold you your chip.Quote:
Originally Posted by genec57
That batch has very good temps compared to voltage, but since you got the cooling I wouldnt be afraid of voltage. 4.4 is a decent clock.
It is a big improvement. I didn't dare take my CO past 4.0
Did you bump that nb voltage? That really helps with bclk on the top end. PCI-E also helps get better high bclk clocks. Most boards didnt fully support that feature because they didnt think it was that important, my bloodrage only goes to 103 also,but on the newer boards like the classified and P6TD deluxe have it. I can do a hard mod to access higher pci-e mhz, but Ill be happy with 4.2 on air cooling 24/7. Anything higher is just for benchmarks anyways. Instead of turbo boost, I wish these cpus could overclock and have speed step work at the same time, AMD does this on the MSI 790fx. I think that this is the future feature I want to see, and if a company implements this Ill be there customer for sure, considering all the power these chips use. All we need is some competition on this front to make it standard practice. Im all for green power.
Im glad your happy!
4.4 mhz 24/7 will be fine with me. My heat killer should arrive today and that should give me a few degrees temp improvement and that will do me for a while (until something else comes along :)).
Just got my D0 920 :D 3939A645 any good?
920 with batch 3838B079 here.
http://img4.imageshack.us/img4/1772/...embenchob7.png
http://img3.imageshack.us/img3/3864/linx126250vgh0.png
Anyone with a "brother" ? :D
Ahem excellent thread and guyz please keep it relavent.
Heres what I have 24/7 :cool:
http://img32.imageshack.us/img32/614...ssssssssss.png
i have the same batch as yours imo.keep us posted as to you results.
i cant tell if mine is any good because i cant even get it to prime with stock settings
I have a pick from two core i7 965:
FPO / Batch
3841A757
vs
3837A945
Is there a recommendation on which batch number to go with??
I cannot find any information regarding 3841A757 - I saw that another member on this forum has the 3837A945.
Thanks,
Van
Here's some stock air results from my 3845B (not prime or bench stable)Quote:
Originally Posted by barrymaldon
21x215 @ 1.344v
http://img32.imageshack.us/img32/6214/21x215at1344v.png
21x195 @ 1.60v
http://img32.imageshack.us/img32/8590/195x21.jpg
I did manage to run prime for 5 mins (before it became too hot (95deg) @ 21x190 @ 1.165v) and with my water cooling i think that could be a 24/7 clock
All overclocks were done with auto voltage on everything apart from core voltage.
Get 920 d0 if you can.Quote:
Originally Posted by vannguyen0
Here is mine 4.4 GHz stable so far crunching, about 10 or so 3dmark runs. Tonight I'll leave it running prime.
Batch # 3849A866
http://img.techpowerup.org/090905/Capture239.jpg
Just noticed I had CPU-Z covered, here it is, same settings.
http://img.techpowerup.org/090905/Capture123.jpg
looks like my future Watercooled OC ! :eek:Quote:
Originally Posted by radaja
Certainly thers a lot more room in the chip...5Ghz now looks like a possibility with LN2 :shocked:
http://img246.imageshack.us/img246/1043/1x2.png
Anyone with 3924A268?
very nice imo:DQuote:
Originally Posted by imo
Thanks, what about you mate? Any good news? :DQuote:
Originally Posted by radaja
no,still having trouble running everything
at optimized defaults.fails prime instantly.
very depressing.
Well , Iam definitly going to be keeping an eye on this thread .
Just came back from MicroCenter (Tustin , Calif) and they gave me a 3910A639 .
Although the chip is not opened or installed yet , after reading a few comments here about a couple of 3910Axxxx ; Iam not sure I'll open it yet . Maybe I'll put off finishing the build till some more results come in .
Iam not looking for a super killer oc or anything , but a solid 4.2 on low volatge would be a-ok .
*edit* Seems to be some better oc results arriving on this batch at Overclock.net : http://www.overclock.net/intel-cpus/...0as-batch.html
This is my latest stable overclock.
Batch # 3849A866
http://img.techpowerup.org/090906/stability 4.2ghz.jpg
I have been playing around with my chip, and the highest I have had it so far is 4.61GHz.
Sub 9 seconds with Super PI 1m:
http://img136.imageshack.us/img136/4...rpisub9sec.png
CPUZ verification @ 4.6GHz:
http://valid.canardpc.com/cache/screenshot/685630.png
Far from being Prime or LinX stable, but at least I know the chip has potential, I will try to get it stable @ 4.4/4.5GHz once I get my custom loop setup in the next couple of weeks..
Batch number 3921A624.
^^good job man, if you disable HT, you might be able to get a bit more out of it and super pi runs improve significantly being a sigle threaded benchmark :)
My updated 24/7 settings 4200MHz @ 1.22v 4 core 8 thread
(less voltage, prime stable) :)
http://img35.imageshack.us/img35/6595/50505897.jpg
Batch #: 3910A500
OC: Core i7 920, 4.0GHz (20x200MHz) [All power saving features disabled]
Hyper Threading: Enabled
CPU: 1.312v
DRAM: 1.5v
QPI/Vtt: 1.215
IOH Core 1.1v
Cooler: Prolimatech Megahalems w/ 2x 70CFM Fans (Push/Pull)
http://i25.tinypic.com/1zo9glx.jpg
Good to see some later batch numbers, I was starting to get nervous. I've got a 920 batch#3849B040 and didn't see any in the spreadsheet. I'm still going to be a week or two before mine is up and running.
Hey. Im new one here at Xtremesystems. And I already have question. Will Xeon W3550 work on my Rampage II GENE? I'm buying it to replace my i7 920. My setup at the moment is:
Asus Rampage II GENE (0203 BIOS now)
Intel Core i7 920 @ 4,3Ghz HT On @ 1.400V (Watercooling, Apogee GT)
3 x 1Gb Corsair XMS3 @ 1229Mhz @ 7-7-7-18
XFX GTX295 (EK Waterblock)
1.5TB Barracuda + 250Gb Samsung
Coolermaster HAF932
Enermax Modu 82+ 625W
Windows XP Professional SP3
:welcome:Quote:
Originally Posted by Esajuppeli
Don't see why not.
why are you buying it again???
Have you tried x19 or x21(turbo) ?Quote:
Originally Posted by SilverSilence
My cpu needed over 1.3V with x20 multi but when going to an odd multi voltage requirement dropped and temperatures dropped too, should give it a try :)
Also with temperatures like that you might be able to push it a little big higher.
Thanks Chicken Patty, I will try disabling HT to see if I can push for more out of this chip.
I should have it under water in the next couple weeks, so I am kinda hoping that I can get 4.4/4.5GHz stable for 24/7 usage.
I was happy with it at 4GHz, but it seems that practically everyone is running 4GHz now, and I don't like being the same as everyone else!!!
I'm currently trying to get 4.2GHz stable on air.
Yeah, at least for super PI you are better off with HT disabled. What are your temps under load with the TRUE? ALso what clock/voltage.Quote:
Originally Posted by BLunTSmoKeR
I7 920 Batch 3850B177
4.2Ghz air run with ultra 120 Xtreme:
http://i263.photobucket.com/albums/i...m/42GhzRun.jpg
CPU 1.3125v
Ram 1.64v
ICH 1.21v
IOH 1.15v
QPI 1.48v
QPI is pretty high because the ram is running @ 2000 ;)
All volts are MM verified...
I had the CPU 4.2 prime stable @ 1.265v on my bench OS but nothing installed to take the SS :D
very good bro, temp are a bit high for comfort, but considering it's on air, very good. I know all rigs are different, but I can do 4.2 GHz stable at 1.24v. If you can get it that low, or lower, temps should improve considerably. :up:Quote:
Originally Posted by Biker
How right you are :)Quote:
Originally Posted by Chicken Patty
Basically my main OS install is over 4 years old (been keeping it alive somehow :D) and it is not much good for benching....
However my bench OS is a much better prospect ;)
I7 920 Batch 3850B177 air run True Ultra 120 Xtreme single push
http://i263.photobucket.com/albums/i...unlowvolts.jpg
Voltages much more under control now ;)
Vcore 1.265v
QPI 1.455v (QPI is high because ram is running @ 2000)
IOH 1.21v
ICH 1.11v
Ram 1.64v
LLC on
All checked with MM @ load
:cool:
told ya :D I didn't try any lower than that, but maybe you can go even lower at that clock. So your temps dropped at least over 10-12ºcQuote:
Originally Posted by Biker
Max stable BCLK achived so far on Air
http://img98.imageshack.us/img98/638/230bclk.jpg
What PCI-E frequency do you use?Quote:
Originally Posted by imo
Quote:
Originally Posted by Chicken Patty
*grin*
Yep.
QPI voltage is still a bit high for long term but everything else is sitting just fine for 24/7/365 :)
Might drop the ram down to 1600 for long term use though and the QPI is happy with 1.25v at that speed...
just save this setting in the BIOS, and save the one with the ram and qpi lower for daily usage. Whenever you are going to bench, load the one with the higher ram speed and bench ;)Quote:
Originally Posted by Biker
here was mine when I was running it at 4.2 GHz. I overclocked more with that same voltage, so I didn't try to go lower. But I think it could have been stable at a bit less voltage than that.Quote:
Originally Posted by Biker
http://img.techpowerup.org/090906/stability 4.2ghz.jpg
110 :DQuote:
Originally Posted by zalbard
wot about you? what batch and speed/volt?Quote:
Originally Posted by zalbard
I'd be careful using non-stock PCI-E frequency 24/7 myself. It may kill your HDDs / sound cards / NICs.Quote:
Originally Posted by imo
I'm in process of RIG upgrade atm, will let you know soon. :p:Quote:
Originally Posted by imo
HiQuote:
Originally Posted by Chicken Patty
nice, whats your batch? :D
oh i thought you asked my max.... 24/7 is 101 ;)Quote:
Originally Posted by zalbard
Batch # 3849A866Quote:
Originally Posted by imo
Quick OC and test, just for a moment using Lynx 5 loops only
i7 D0 batch 3925A260
vcore 1.2v BIOS
vtt 1.24v
vioh 1.16v
http://i29.tinypic.com/2w57pdz.jpg
[QUOTE=WeHa;4001514]Quick OC and test, just for a moment using Lynx 5 loops only
i7 D0 batch 3925A260
vcore 1.2v BIOS
vtt 1.24v
vioh 1.16v
thats a good way to test like benching stability or something, but even with about half hour of LinX, i've had unstable computers, i've resorted to Prime 95, never fails if you run it over night :D
I am now LinX stable (20 passes using max available RAM), but my Vcore seems a little high compared to yours Patty and biker, I need 1.3375v Vcore (1.312v in CPUZ) and 1.355v Vtt, my idle temps are around 40c, and loads get to about 82c which is a little bit to hot for my liking, but this is only with LinX.
I will have a play around with those voltages tomorrow to see if I can get them any lower..
Well, I know I won't get Vcore any lower as I have been playing with that all night, and my chip needs 1.3375 in the BIOS otherwise LinX fails after 2 passes.
I might get my Vtt down a little though.
I was told that if you lower CPU PLL this can help stabilize things, so I have got mine down to 1.300v from 1.800v, and it has been fine, however I have also had to raise QPI PLL to 1.200v and IOH Core is at 1.140v as well.
Need to get some sleep now as I am beat, but I will be back at it tomorrow again, I'm gonna see if I can tighten up my RAM timings, because when I was running 191x21 I was able to set RAM to 8-7-7-21-1t instead of 8-8-8-24-2t..
@WeHa, I don't really think that running LinX with anything less than max available RAM is worth it tbh, you have only used 772MB which is pointless, this won't tell you if your system is stable or not.
Also, which board have you got?
If I set my Vcore to 1.2v in the BIOS it would end up being something like 1.14v in CPUZ, but yours is higher than 1.2v in CPUZ screenshot in your post... I would like to know which board it is, as I always thought mobos suffer from Vdrop, but yours is clearly overvolting..
::edit::
Sorry, should have checked your sig!!! :dumbass:
Do all EVGA mobos overvolt like this?
Quote:
Originally Posted by BLunTSmoKeR
Try these settings man, they worked like a charm for me :D
200*21 = 4200 MHz
CPU VCore ( 1.2500v )
CPU VTT Voltage ( +50 )
CPU PLL VCore ( 1.800v )
DIMM Voltage ( 1.65 )
DIMM DQ Vref ( +0 )
QPI PLL VCore ( 1.3500v )
IOH VCore ( 1.2750v )
IOH/ICH I/O Voltage ( Auto )
ICH VCore ( Auto )
thanks mate. in my experiences, if i had an unstable oc, i couldn't run lynx even 1 loop only, in the other hand if i could pass 5 loops lynx, it would be likely rock stable. but in some uncommon cases, some setup still can't pass 20 loops lnyx while it passes 5 loops :)Quote:
Originally Posted by Chicken PattyQuote:
Originally Posted by WeHa
thanks for the reply bro. i will try to run lnyx @ 6GB memory to find out if it is really stable or not :)Quote:
Originally Posted by BLunTSmoKeR
i am using Foxconn Bloodrage P06 BIOS. maybe the sensor is incorrect or something wrong about the voltage sensor? anyway i just use "default" vcore to get 1.232v in CPU-Z and 1.2v is shown at the BIOS.
yeah, I get what you mean, usually for me too, but i've seen the case where it isn't so I try to run longer tests with Prime.Quote:
Originally Posted by WeHa
Quote:
Originally Posted by WeHa
My EVGA E758 A1 does that too. IT has vdroop control, so actually if my idle is 1.141, load will be like 1.164
What's the best actual D0 batch for OC ?
Thanks.
yeah i am going to run Prime95 tonight to ensure the stability of the oc. thanks a lot for the advice :)Quote:
Originally Posted by Chicken Patty
:yepp: Currently pushing it as far as it can go, so far I've managed 4.4Ghz, gonna keep going though :cool:Quote:
Originally Posted by Dahmer
Thanks for the suggestions!
3845B. Closer to 0 = goodQuote:
Originally Posted by jVIDIA
I wouldn't guide myself by batch #. All CPU's differ I've seen a supposed "great batch" have :banana::banana::banana::banana:ty overclockers. It's just a matter of luck my friend.Quote:
Originally Posted by jVIDIA
K, keep us posted ;)Quote:
Originally Posted by WeHa
I agree with Patty, only use batch numbers as a loose guide, as all chips vary even when they come from the same batch.
Both of my 920's have identical batch numbers, and they are totally different in regards to overclocking.
I am really frustrated with my chip, I got it to run 20 passes of LinX using all available RAM, so I started Prime95 Blend, and when I checked it this morning, it has crashed, and reset my system.
I am using 1.3375v Vcore (1.312v in CPUZ idle, and 1.291v under load), Vtt is at 1.355v, IOH 1.140v, QPI PLL 1.200v and the rest are at stock.
I don't know what I need more of to get it stable, whether it be Vtt, or Vcore, or even both.
It is driving me nuts to say the least.
How high would you guys say is the safest amount of Vtt and Vcore for 24/7 use?
It is really annoying, because my chip will do 200x20 with Vcore at 1.264v, and QPI @ 1.335v, and it is both Prime and LinX stable, and I have been playing games for hours on end without a single issue, so I would have thought that my system would have been stable with the Vcore and Vtt voltages I mentioned above, but it just doesn't want to happen.. :(
Turbo mode must demand quite a bit of extra voltage, as this is the only thing I am changing to get my chip up to 4.2GHz..
I will keep tinkering with it though, as I don't give up easily!
It is bound to run at 4.2GHz stable, it's just a matter of finding the sweet spot with my voltages in order to get it stable!!!
I has taked:
3911A315
isn't bad?
bump the 1.26 to 1.3 and see if 4.2 holds
3.8GHz @ default vcore
http://i32.tinypic.com/2lc4fbq.jpg
http://i32.tinypic.com/2ryhhch.jpg
my i7 920 D0 #3909A179 (little bit shabby) result :lsfight: :
edit :
+/- 3hr priming and still ongoing :D
http://img35.imageshack.us/img35/4682/80735384.th.jpg
vcore : 1.19375v
vpll : 1.80v
vqpi : 1.20v
vdram : 1.58v
vioh : 1.10v
C1E,CPU TM, EIST, TURBO, C-STATE : all enabled <-go green :up: & reduced my electricity bill :rofl:
in my quick IBT test ( 3x passes:D ) at above configuration my cpu can run @3.945ghz (188x21) but i favor 3.88ghz more ( it's mystical decision ) :rofl: :ROTF:
4GHz 20x200 @ default vcore to run windows 7
http://i30.tinypic.com/5p4cr7.jpg
after this i'm gonna run prime95 to find out if it is rock stable :)
Any info on 3924A390 or 3919A?
Quote:
Originally Posted by WeHa
5 minutes of Prime95??
4.2GHz LinX stable batch 3921A624
http://img30.imageshack.us/img30/2695/4200mhz.jpg
Same Pic :D
different chip :D
http://img232.imageshack.us/img232/9471/capture1b.gif
LoL!
I am trying to get my Vcore down, in that screenie I had Vcore set to 1.3500v in the BIOS, but my Vtt is @ 1.255v, CPU PLL is @ 1.300v instead of 1.800v, QPI PLL is at 1.100v and IOH Core @ 1.140v, I am just about to test with Vcore set to 1.2875v to see if I can get those temps down.
I am getting really high temps compared to others with TRUE's, I have 2 Xigmatek XLF-F1253 fans strapped to mine as well, I think it is my case that has poor airflow tbh, the TJ10 only really has one intake fan, and it blows directly over the GFX card, it has another fan that they say is an intake, but it is sandwiched between the 2 HDD cages, so it doesn't really put any airflow through the main part of the case.. I wondering it I were to turn my rear exhaust fan round so it was taking air into the case If I would get lower temps, as I have another 2 x 120mm fans on the roof of the case that are exhausting air from the case..
Does anyone think this would make a difference in my temps?
::edit::
imo which fans are you using on your TRUE mate?
Fill this out man, let's see how we can help you pass Prime :D. Fill out what you can, depending on the BIOS version you have, options may differ!Quote:
Originally Posted by BLunTSmoKeR
Advanced CPU Features:
CPU Clock Ratio ................................
Intel(R) Turbo Boost Tech ..................
CPU Cores Enabled ............................
CPU Multi Threading ..........................
CPU Enhanced Halt (C1E) ...................
C3/C6/C7 State Support ....................
CPU Thermal Monitor .........................
CPU EIST Function ............................
Virtualization Technology ...................
Bi-Directional PROCHOT .....................
Uncore & QPI Features:
QPI Link Speed ..............................
Uncore Frequency ..........................
Isonchronous Frequency .................
Standard Clock Control:
Base Clock (BCLK) Control
BCLK Frequency (MHz)
PCI Express Frequency (MHz)
C.I.A.2 [Disabled]
Advanced Clock Control:
CPU Clock Drive
PCI Express Clock Drive
CPU Clock Skew
IOH Clock Skew
Advanced DRAM Features:
Performance Enhance ...................... [Standard]
Extreme Memory Profile (X.M.P) ......... [Disabled]
System Memory Multiplier (SPD) ........ [AUTO]
DRAM Timing Selectable (SPD) .......... [Manual]
Channel A + B + C
Channel A Timing Settings:
##Channel A Standard Timing Control##
CAS Latency Time
tRCD
tRP
tRAS
##Channel A Advanced Timing Control##
tRC
tRRD
tWTR
tWR
tWTP
tWL
tRFC
tRTP
tFAW
Command Rate (CMD)
##Channel A Misc Timing Control##
Round Trip Latency ...................
Advanced Voltage Control:
CPU
Load Line Calibration
CPU Vcore
QPI/VTT Voltage
CPU PLL
MCH/ICH
PCIE
QPI PLL
IOH Core
ICH I/O
ICH Core
DRAM
DRAM Voltage
DRAM Termination
Ch-A Data VRef.
Ch-B Data VRef.
Ch-C Data VRef.
Ch-A Address VRef.
Ch-B Address VRef.
Ch-C Address VRef.
__________________
Good luck.
ya give it a try.
Gentle Typhoon 1850rpm. :D
Advanced CPU Features:
CPU Clock Ratio ................................ 20x
Intel(R) Turbo Boost Tech .................. Enabled
CPU Cores Enabled ............................ All
CPU Multi Threading .......................... Enabled
CPU Enhanced Halt (C1E) ................... Disabled
C3/C6/C7 State Support .................... Disabled
CPU Thermal Monitor ......................... Disabled
CPU EIST Function ............................ Disabled
Virtualization Technology ................... Enabled
Bi-Directional PROCHOT ..................... Enabled
Uncore & QPI Features:
QPI Link Speed .............................. 36x
Uncore Frequency .......................... 16x
Isonchronous Frequency ................. Enabled
Standard Clock Control:
Base Clock (BCLK) Control
BCLK Frequency (MHz) 200
PCI Express Frequency (MHz) 102
C.I.A.2 [Disabled]
Advanced Clock Control:
CPU Clock Drive 900mV
PCI Express Clock Drive 800mV
CPU Clock Skew 0ps (I think this is what it says for default?)
IOH Clock Skew 0ps (same as above)
Advanced DRAM Features:
Performance Enhance ...................... [Standard]
Extreme Memory Profile (X.M.P) ......... [Disabled]
System Memory Multiplier (SPD) ........ [AUTO] 8x
DRAM Timing Selectable (SPD) .......... [Manual]
Channel A + B + C
Channel A Timing Settings:
##Channel A Standard Timing Control##
CAS Latency Time 8
tRCD 8
tRP 8
tRAS 24
##Channel A Advanced Timing Control##
tRC 41
tRRD 6
tWTR 6
tWR 12
tWTP 24
tWL 8
tRFC 88
tRTP 6
tFAW 30
Command Rate (CMD) 2T
##Channel A Misc Timing Control##
Round Trip Latency ................... Auto
Advanced Voltage Control:
CPU
Load Line Calibration Enabled
CPU Vcore I am just about to test with 1.2875v it passed LinX with it set to 1.3500v in the screen above.
QPI/VTT Voltage 1.255v
CPU PLL 1.300v (I read that lowering this can help gain stability, I have tested it with 1.800 and 1.860v and it doesn't make any difference to my o/c)
MCH/ICH Auto
PCIE Auto
QPI PLL 1.100v (tested this using every increment from 1.100v to 1.240v made no difference)
IOH Core 1.140v (again played with this right up to 1.240v)
ICH I/O Auto
ICH Core Auto
DRAM
DRAM Voltage 1.66v
DRAM Termination Auto
Ch-A Data VRef. Auto
Ch-B Data VRef. Auto
Ch-C Data VRef. Auto
Ch-A Address VRef. Auto
Ch-B Address VRef. Auto
Ch-C Address VRef. Auto
__________________________________________________ _______________
I should think I could get 4.2GHz Prime stable, I don't understand why my chip will be rock solid @ 200x20 with only 1.248v Vcore under load, yet I struggle to find stable settings for 4.2GHz and need to crank Vcore in order for it to pass LinX???
Like I said, I am about to test again with Vcore at 1.2875v instead of 1.3500v..
If you guys have any suggestions I would be happy to try them out! :)
::edit::
I've got some Arctic Cooling fans 12025PWM fans coming on Thursday hopefully, these will be getting used on my triple rad once I setup my WC loop, but I will try them out on my TRUE until I get my HK 3, I imagine they will be better than these Xigmatek fans I have atm, as they have got high static pressure, which should make a difference I've been told..
see above, I put my response in BOLD UNDERLINED TEXTQuote:
Originally Posted by BLunTSmoKeR
your problem is probably the memory diviver, it gets real hard to keep the divider at 8x at 4ghz and above. Try 6x, probably your issue. :D keep me posted.
Try scythe kaze 3000rpm or Delta might blow your CPU away :DQuote:
Originally Posted by BLunTSmoKeR
No problem Patty, I will try out your suggestions to see how I get on.. I have never tried using anything more than 1.66v on the RAM because of the Intel warning thing, but I see lots of others pushing more VDimm than 1.65v, so I guess I should try going for more..
I will let you know how I get on though, but I am going to bed now, as it is midnight here, and I am shattered!!!
Take it easy guys.
Every chip hits a point where the next increment of performance takes a large amount of voltage to accomplish. That is a good sign that you are approaching the overclock limit for that chip.
It will be interesting to see your result. It is a big drop down from 1.35 to 1.2875. I will be surprised if it works unless you were significantly over volted at 1.35.
My DO chip requires 1.36 vCore at 4200 and 1.44 to get to 4400 and that is with HT off (on for 4200). Every chip is different.
Quote:
Originally Posted by BLunTSmoKeR
goodnight dude, keep us posted. and as far as Dram, it won't hurt.
This is very true, my C0 did 3.8 Ghz @ 1.292v, then for 4ghz it need 1.36v. 4.2 ghz needed 1.45v. at this point, its better to look for a balance of performance / voltages.Quote:
Originally Posted by genec57
quick test on 4.2GHz 21x200 prime95 5 minutes only
http://i27.tinypic.com/1pdtm8.jpg
you going to try and run it longer?Quote:
Originally Posted by WeHa
sure tonight i'll try because i gotta go to office right now my friend :DQuote:
Originally Posted by Chicken Patty
awesome, hope it's stable :DQuote:
Originally Posted by WeHa
Well, I finally got 4.2GHz Prime95 Blend Test stable!!
I have to set Vcore to 1.34v in the BIOS, but QPI is only 1.335v, CPU PLL is 1.300v, IOH Core 1.160v, QPI PLL 1.160v, VDimm on 1.66v and all other voltages are on auto.
Been running Prime95 for just over 6 hours now, and it is still going, so I would say it is stable, however, I am going to leave it running overnight just to be sure.
http://img11.imageshack.us/img11/2936/prime6hours.jpg
Temps are a little high for my liking, but I'm putting my chip under water in the next couple of weeks, I have got my 120.3 rad, fittings and fans being delivered tomorrow, so the only thing I need to get after that is my Heatkiller 3 copper, and I will be all set to start planning my loop layout, I don't know whether I should use both my 120.2 and 120.3 rad in the loop or not, I will need to have a think about it, and see where everything is going to go, as it might be tricky for me to fit both rads inside my current case, however, I will be getting the Silverstone TJ07 in November to replace my TJ10, so I know that I will have plenty of room in the TJ07 for both rads, and I won't need to hack it up in order to make things fit..
Nice Blunt, good results, you have LLC disabled?
Anyone have experience with this batches?
3912B351
3913B192
Quote:
Originally Posted by BLunTSmoKeR
congratulations on getting it stable bro :D, temps are ok for now, they don't get that high with nothing else. For now you'll be fine with the 120.3. When you get the TJ07 just add the other rad, and maybe at the time you might want to add your motherboards NB to the loop or something, with both rads that wouldn't be a problem.
Thanks guys, I just stopped it to double check the BIOS settings were what I posted in that last comment, and everything was right, apart from QPI PLL, it is set to 1.240v, not 1.160v.
I have dropped Vcore down to 1.315v in the BIOS and I will try running Prime again to see if it is stable there, I saved the previous profile just incase it isn't though.
I think the main thing with o/cing these i920's is finding the right balance between all the main voltages, because I have tried running Prime with 1.35v Vcore and 1.415v QPI before, and it would crash after an hour, but I have considerably less QPI this time, and Prime ran for hours without any issues, and I have also been browsing at the same time to..
@phsinc, I have LLC enabled mate, if I didn't I would have to compensate for .09v of Vdroop, so that would mean I would have to set Vcore to 1.39v or something ridiculous like that, I have a 4GHz stable profile with it disabled, but tbh, I don't see the point in having it disabled, lots of guys on another forum I frequent say that it can damage chips, but I asked people here about it, and they said unless I am using xtreme voltages then it isn't going to cause any damage at all...
I agree, these are very tricky CPU's bro, but after all it's worth it. Glad you acheived what you were after :up:.Quote:
Originally Posted by BLunTSmoKeR
Did you drop the mem divider or you kept it at 8x?
Kept it @ 8x, I didn't want to underclock my RAM by 400MHz!
I just tried it with 1.315v, and got a BSOD immediately.. so I have put it back to 1.3500v again, and I will leave it running Blend Test overnight just to be absolutely sure it is stable..
Once I get it under water I will probably push for more, I would love to have 4.5GHz for 24/7 use, I just hope I get decent temps once it's under water.
What kinda temps do you think I should get, any idea?
I will be using these components in my loop:
Aqua Computer airplex XT 360 rad
XSPC DDC Pro 10w Pump/Res
Watercool Heatkiller 3 Copper
EK FC-GTX280 block and backplate
With 7/16" ID 5/8" OD XSPC tubing and 1/2" DTek high flow barbs.
This is the first ever time I will have used water cooling, so I have been reading up like crazy trying to learn as much as possible before I set everything up, so I don't make any mistakes!!!
I just want to have temps that are below 80c, though I have seen guys with similar WC parts clocked to 4GHz and their load temps don't even reach 70c, and there are guys on OcUK with that Corsair H50 thing getting around 72c at 4GHz, and that thing only has a 120.1 rad, so I imagine using a 120.3 should give much better temps than the H50??
Well at 4.2 GHz @ 1.256v load is like mid 60's. I have a 120.4rad, check below link to see my components.Quote:
Originally Posted by BLunTSmoKeR
http://forums.techpowerup.com/showthread.php?t=82607
LoL, a quad rad just for the CPU!!! :D
I don't think my temps will be as good as this because I will have my GPU in the loop, and I need much higher Vcore to get 4.2GHz..
I was wondering, would a different mobo allow lower Vcore, or would it just be the same?
I'm thinking about getting the Classifed at some point in the near future.
In the past I have had chips on one mobo do 2.8GHz, but then when I got a new board, I wasn't able to get the clocks I could previously, so that tells me that the mobo does play a part in how high you can clock a chip, but I am unsure about the voltages needed.
yeah I went all out in the rad in case, this rad can handle the CPU and board perfectly fine, and when I get a HK block in the future, temps should improve by 4-5ºc, this block kinda sucks on the i7, but at the time it was one of the few choices, so I went with it, only to discover it's not great on i7's.Quote:
Originally Posted by BLunTSmoKeR
Well from what i've read and seen the UD5 is the most solid X58 board, the classified is nice, but I think your limiting factor here is your CPU.
I had some time yesteday so I reset my cooler with 2 lines thermal compound method which gave me 4C improvement and new 24/7 overclock.
3922A415 batch
CPU - 1.296V (1.328V load)
VTT - 1.26V
RAM - 1.65V (1.66V shows in Everest) 7-7-7-20 (8-8-8-24 stock - I love this RAM)
IOH - 1.1V (1.12V shows in Everest)
http://i27.tinypic.com/qrf3bt.jpg
here's my 1M SuperPi
http://i26.tinypic.com/ojkxdw.jpg
Nice Oc
Nice tight RAM timings, I can only get my Dominators to do 8-7-7-21-1T, can't get CAS7 at all, though I have only tested with 1.65v VDimm, I guess if I were to raise it to 1.70v they might just do CAS7?Quote:
Originally Posted by McDown
You say you have your Vcore set to 1.296v in the BIOS but it shows 1.328v in CPUZ?
My board does the total opposite of this, if I set Vcore to 1.275v in the BIOS, it will end up being like 1.263v idle in CPUZ and under load it would be 1.248v..
quick test on 3dmark06 cpu test
4.4GHz @ 1.36v
http://i25.tinypic.com/2nlxvv8.jpg
I'm very impressed too. I never had Kingston before and always thought they make cheap value stuff. I only got it because OCZ was out of stock and Kingston was on sale (original price $240, sale $149) and I'm so not sorry I did it. I can run it 1700mhz at 8-8-8-20 stable as rock.Quote:
Originally Posted by BLunTSmoKeR
No I set 1.315V in bios, it shows 1.296V idle and 1.328V load in Windows. I think that's a Bloodrage thing. I call it Vjump :DQuote:
You say you have your Vcore set to 1.296v in the BIOS but it shows 1.328v in CPUZ?
My board does the total opposite of this, if I set Vcore to 1.275v in the BIOS, it will end up being like 1.263v idle in CPUZ and under load it would be 1.248v..