HT ON:
3839A568
Linpack temps :up:
http://i125.photobucket.com/albums/p...la/Linpack.jpg
uhh linpack requires more voltage then prime. U sure its not the other way around?Quote:
Originally Posted by JWMc
Printable View
HT ON:
3839A568
Linpack temps :up:
http://i125.photobucket.com/albums/p...la/Linpack.jpg
uhh linpack requires more voltage then prime. U sure its not the other way around?Quote:
Originally Posted by JWMc
Positive, check the LinX thread for lots of screens from lots of people with i7 and HT on ;)Quote:
Originally Posted by The-Fox
NaeKuh, FFS, read the thread! :p:
IBT does NOT stress the CPU properly with HT enabled.
@JWMc: I have tested with 3 different i7 rigs so far.. Linpack 64 definitely needs more Vcore than Prime blend, even though prime blend needed one notch more VTT in one case for me, equal in the other cases. Overall, LinX is the hardest tool to test i7 right now.
I was able to get into windows at 21x205 monday night. I'll get the batch number posted up when i get home. I leave turbo on, i think i might have hyperthreading off atm though.
Yep, every time on my chip. I wonder if it's special. The temps are higher with LinX every time too, by a good bit. Usually when I'm using that low voltage that works for LinX only I can Prime Blend for a while, sometimes 1.5-2 hours! But it won't make it 4+ hours until I raise voltage 2-3 steps depending on where I am.Quote:
Originally Posted by NaeKuh
Edit: I don't have enough evidence and I doubt I'll see this behavior again, if I ever even did. I think I was simply lucky for passing LinX in the first place.
FPO/Batch #: 3836A405
OC: Core i7 940, 4.2GHz (21x200MHz) with HT, 1.39375V BIOS (see screenshot as well for volts and temps), 1.35000V QPI/DRAM, 1.65681V DRAM, 1600MHz DRAM at 7-8-7-20-1T, cooled by air using TRUE 1366RT, Load Line Calibration on Auto, C-states active
Yes I have noticed Prime blend needing more VTT, a good bit more for 4ghz with HT vs 3.8ghz with HT. But I have also noticed it needing slightly more voltage to make it through long Prime blend runs vs just burning through a quick LinX x 20.Quote:
Originally Posted by jcool
Whenever I try a drastically different OC I start with LinX, watch it blow up really fast, keep upping voltage until it passes x 20. Then I go for blend and watch it work fine for a while.. and think it's stable. Then re-test and sigh... it isn't stable.
One thing that has been a warning sign that has held true for me every time is if I see three different voltages for Min/Max/Current in HW monitor after leaving Prime blend running for a while. It means my voltage is borderline and it's occasionally drooping too much. I'm jealous of these boards that put out more voltage than you ask for in the BIOS. I have to enable LLC just to keep it from sagging further (Gigabyte UD5).
You mean like the R2E :p:Quote:
Originally Posted by JWMc
Those oscillations over time usually mean that you are pushing your board beyound its limits and the built up stress and heat are causing things to slowly fail.
Well if you use low memory values in LinX a 10min pass doesn't mean it will pass 10h prime ofc. But if you run on at least 3GB of memory (ideally 5 with 6GB) in 64bit mode for about an hour, I can almost guarantuee you that prime won't fail within >10h (on i7, that is).Quote:
Originally Posted by JWMc
Or just isolate your memory OC first, find a stable one, then go and OC your CPU.Quote:
Originally Posted by jcool
Doing both at the same time is not very good as you find yourself not being able to determine where the problem is coming from...
Not sure how much this will help, but over the last 2 days I reread the i7OC thread and pulled all this out of it. I haven't added the new info from this thread or the Study thread that popped up yesterday either, but I could if this kind of list is useful to anyone.
EDIT:Upadted 5:12PM CST.
Notes:
Added a [STABLE] flag. I don't want to get in a pissing match about what is the definition of stable, but if the chip owner says their system is stable I will flag it as so. Thanks
Added all the posts from this thread and the i7 study thread. If I missed someone let me know via pm or post here.
Core I7 920 A Batch
-3836a381
Chri$ch-3800.1Mhz @ 1.232V HT On http://www.xtremesystems.org/forums/...&postcount=144
Chri$ch-4009.3Mhz @ 1.360V HT On http://www.xtremesystems.org/forums/...&postcount=151
Chri$ch-4200.2Mhz @ 1.440V HT On http://www.xtremesystems.org/forums/...&postcount=201
deepsilver-3599.9Mhz @ 1.248V HT On http://www.xtremesystems.org/forums/...&postcount=259
-3841A383
JWMc-3800.0Mhz @ 1.218V HT On http://www.xtremesystems.org/forums/...postcount=1902
JWMc-4000.0Mhz @ 1.300V HT Off http://www.xtremesystems.org/forums/...3&postcount=15
JWMc-4000.0Mhz @ 1.406V HT On http://www.xtremesystems.org/forums/...3&postcount=15
16floz470ml-3827.6Mhz @ 1.195V HT On http://www.xtremesystems.org/forums/...postcount=1928
Tot31-3500.0Mhz @ 1.056V HT Off http://www.xtremesystems.org/forums/...postcount=1933
[STABLE]Tot31-3799.8Mhz @ 1.144V HT Off http://www.xtremesystems.org/forums/...08&postcount=5
-3839A403
Chicken Patty-4303.5Mhz @ 1.435V HT Off http://www.xtremesystems.org/forums/...postcount=2000
Chicken Patty-4192.2Mhz @ 1.411V HT On http://www.xtremesystems.org/forums/...04&postcount=2
-3836A405
Dejanh-4187.9Mhz @ 1.384V HT On http://www.xtremesystems.org/forums/...1&postcount=30
- 3841A431
Icecold316-No Results Posted.
- 3839A459
FragMagnet-No Results Posted.
- 3838A505
Scaramonga-4000.2Mhz @ 1.424V HT Off http://www.xtremesystems.org/forums/...0&postcount=10
- 3837A507
Nordien-4217Mhz @ 1.400V HT Off http://www.xtremesystems.org/forums/...postcount=1464
Mikesrex-No Results Posted.
- 3836A514
Sonofander-No Results Posted.
- 3836A531
Mascaras-3800.1Mhz @ 1.200V HT On http://www.xtremesystems.org/forums/...&postcount=651
Mascaras-4000.2Mhz @ 1.344V HT On http://www.xtremesystems.org/forums/...&postcount=464
Mascaras-4312.0Mhz @ 1.376V HT On http://www.xtremesystems.org/forums/...&postcount=464
-3838A554
Loc.o-4000.0Mhz @ 1.325V HT On http://www.xtremesystems.org/forums/...&postcount=597
-3839A568
NaeKuh-4005.1Mhz @ 1.392V HT On http://www.xtremesystems.org/forums/...4&postcount=26
-3839A579
Tobit-No Results Posted.
-3838A646
Uberbob10200-3400.0Mhz @ 1.153V HT ? http://www.xtremesystems.org/forums/...&postcount=610
- 3838A685
Jcool-“My 3838A685 920 sucks”
- 3835A699
CodeRush-4053.3Mhz @ 1.344V HT Off http://www.xtremesystems.org/forums/...2&postcount=11
- 3837A714
Revenant-No Results Posted.
- 3837A715
Joger-3618.1Mhz @ 1.094V HT On http://www.xtremesystems.org/forums/...5&postcount=18
- 3837A727
Freshy- No Results Posted.
-3837A728
ShoNuff-3800.1Mhz @ 1.2248V HT On http://www.xtremesystems.org/forums/...&postcount=470
Chuckbam-No Results Posted.
Bobbylite-4200.0Mhz @ 1.344V HT Off http://www.xtremesystems.org/forums/...&postcount=616
Bobbylite-4010.0Mhz @ 1.376V HT On http://www.xtremesystems.org/forums/...34&postcount=6
- 3835A766
Beast200-4200Mhz @ 1.425V HT Off http://www.xtremesystems.org/forums/...7&postcount=12
The-Fox-4199.8Mhz @ 1.416V HT On http://www.xtremesystems.org/forums/...7&postcount=50
- 3837A768
Ghigo-3800.0Mhz @ 1.34V HT ? http://www.xtremesystems.org/forums/...30&postcount=9
- 3837a769
Nosboost300-No Results Posted.
[STABLE]Hot’n’Chilly-3800.0Mhz @ 1.320V HT On http://www.xtremesystems.org/forums/...71&postcount=1
- 3836a772
Gligy-3800.0Mhz @ 1.264V HT On http://www.xtremesystems.org/forums/...&postcount=227
-3836A866
Jupiter-3800.0Mhz @ 1.232V HT On http://www.xtremesystems.org/forums/...&postcount=467
-3836A867
Metroid-3717.8Mhz @ 1.160V HT Off http://www.xtremesystems.org/forums/...&postcount=552
Metroid-4009.2Mhz @ 1.248V HT Off http://www.xtremesystems.org/forums/...&postcount=560
Metroid-4009.0Mhz @ 1.272 HT On http://www.xtremesystems.org/forums/...&postcount=566
- 3837A909
Tjelaw-4200Mhz @ 1.362V HT On http://www.xtremesystems.org/forums/...postcount=1494
SmokeyTheBandit-4049.6Mhz @ 1.312V HT On http://www.xtremesystems.org/forums/...84&postcount=2
Bullu$-4226.4Mhz @ 1.448V HT On http://www.xtremesystems.org/forums/...7&postcount=16
Core I7 920 B Batch
-3837B072
Webwilli-No Results Posted.
Chri$ch- No Results Posted.
- 3837B076
Cky2k6-4005.4Mhz @ 1.344V HT On http://www.xtremesystems.org/forums/...17&postcount=1
-3838B079
Namegt-3780.0Mhz @ 1.168V HT On http://www.xtremesystems.org/forums/...&postcount=603
Namegt-4000.2Mhz @ 1.216V HT Off http://www.xtremesystems.org/forums/...&postcount=620
Chri$ch-4011.2Mhz @ 1.248V HT On http://www.xtremesystems.org/forums/...postcount=1465
Dengyong-4020.1Mhz @ 1.296V HT On http://www.xtremesystems.org/forums/...14&postcount=9 //Comment-Deng posted 1.296V but OCCT displays 1.10V, so not sure on that.
-3837B080
Dj883u2-3705.0Mhz @ 1.088V HT On http://www.xtremesystems.org/forums/...&postcount=684
Dj883u2-4010.0Mhz @ 1.152V HT On http://www.xtremesystems.org/forums/...&postcount=619
Dj883u2-4104.3Mhz @ 1.264V HT On http://www.xtremesystems.org/forums/...&postcount=650
Dj883u2-4515.0Mhz @ 1.376V HT On http://www.xtremesystems.org/forums/...&postcount=706
- 3837B194
Nilos-4200.1Mhz @ 1.392V HT On http://www.xtremesystems.org/forums/...51&postcount=3
[STABLE]ReverendMaynard-4199.9Mhz @ 1.360V HT On http://www.xtremesystems.org/forums/...22&postcount=5
-3837B215
DerekT-3183.5Mhz @ 1.136V HT On http://www.xtremesystems.org/forums/...postcount=1859
Core I7 940 A Batch
-3836A405
[STABLE]Rge-4200.0Mhz @ 1.376V HT On http://www.xtremesystems.org/forums/...9&postcount=24
-3835A730
Jcool-4400.0Mhz @ 1.504V HT On http://www.xtremesystems.org/forums/...postcount=1912
Core I7 940 B Batch
Core I7 965 A Batch
-3836A383
FuriousSalesman-No Results Posted.
-3835A538
Giorgos th.-5044.3Mhz @ 1.568V HT Off http://www.xtremesystems.org/forums/...&postcount=500
Giorgos th.-4805.9Mhz @ 1.536V HT Off http://www.xtremesystems.org/forums/...&postcount=492
Giorgos th.-4653.3Mhz @ 1.424V HT On http://www.xtremesystems.org/forums/...&postcount=369
Giorgos th.-4423.9Mhz @ 1.440V HT On http://www.xtremesystems.org/forums/...&postcount=305
Giorgos th.-4271.5Mhz @ 1.424V HT On http://www.xtremesystems.org/forums/...&postcount=294
Giorgos th.-4102.3Mhz @ 1.360V HT On http://www.xtremesystems.org/forums/...&postcount=294
Core I7 965 B Batch
-3837B209
Kazukun-4200.1Mhz @ 1.544V HT On http://www.xtremesystems.org/forums/...postcount=1134
from page one
http://www.xtremesystems.org/forums/...22&postcount=5
add that to the "B" batch
Well done! You could add my crappy 3838A 920 as well if you like.
One thing might be important here: Add a "stable" or "screen" or "benchable" tag. Obviously I could post CPU-Z at 4,7 or 4,8Ghz but that wont help anyone comparing numbers...
I guess I could go back through all the links and try to dig that info out. Just finding some of that info to begin with was like a needle in a haystack. The original i7OC thread is just loaded with so much unorganized info.Quote:
Originally Posted by jcool
The other thing that is just as necessary to fix as Vcore is CPU temp. By that, I mean that I see a lot of validations where someone has passed prime at a high clock and low Vcore, but has a low CPU temp. In that case, they may not have an especially good chip, they may just have a really good cooling system. For example, a rig that has passed prime at , say, 4.2Ghz at 1.38 Vcore, but at 65C CPU temp is much less impressive as a rig that has done the same VCore/CPU Clk at 82C.
In general, I have seen few if any impressive validations on chips later than week 39. Week 36 seems to be the best. I recently got a 920 that booted into windows and ran a 4552 Mhz pass at 1.456 Vcore on super pi on air, batch 3836A977. I'll post validation in a day or two, that rig is down right now.
The thing is my temps are great though because I backed off to 3.8ghz. I need a lot more voltage for 4ghz than 3.8ghz and the heat under LinX load at 4ghz goes to 83c on this TRUE 120 with two 2100 rpm fans.Quote:
Originally Posted by dejanh
I'm finally fully stable at 19x200 with HT: 11 hours of Prime blend at 1.185v load and 62c max temp. LinX max temps are of course higher at about 65-66c. Er, actually the voltage is still dropping to 1.168v under LinX so it does drop more there. For some reason it's still more stable with lower volts in LinX than Prime blend though.
New Edit: HEY! Guess what, now my machine wants to blue screen on LinX all of a sudden. I think it only started failing because I raised the RAM usage up a bit. I was using a problem size of 10000, not big enough for even just 4 gigs memory. I just raised my bios vCore setting to 1.22500v and now I'm finally getting a solid 1.20v idle and 1.184v load on LinX (which passed).
Quote:
Originally Posted by Avan
thanks for putting them altogether Avan. :up:
Can someone explaine how the batch nr works? Where for exsample, is the produktion week?
Some say early produktion week are best..!?
I can choose between 3838A826 and 3838A828 batch of i7 920...
which is best? equal?
Thx!
ByE!
I second this request!Quote:
Originally Posted by PRJ
The production date explnation is pretty simple:
Lets take a common batch number:
3836A867
First two digits: not sure.
Second two digits: Production week.
First Letter: Batch type, A/B - we are not sure what is exactly the differeance but the rumor has it that B types requires less voltage at the same clock.
Last 3 digits: probably unit number in the batch as we have seen different CPUs with the same production week but last 3 digits were different, e.g.: 3837A905 and 3837A902.
If anyone has better info on this, feel free to correct me.
Avan, I duplicated my result in this thread on this page, earlier on so you can add it to your list. Stable Core i7 940 at 4.2GHz with 1.39375V in BIOS. Load was showing roughly 1.38V. I did not enable LLC.Quote:
Originally Posted by Avan
iirc, the first three digits are the lot number. You'll start to see a progression of this number over the cycle of the cpu.Quote:
Originally Posted by The-Fox
First digit OR letter - where the cpu is manufactured (3 - Costa Rica), second digit - the year of manufacture (8 is, obviously, 2008)
BTW, google is your friend ;)
Quote:
Originally Posted by anzial
Good one with google, i tryed but apperently not enough :up:
I just don't think it goes for i7 CPU's, but still.
I just ordered a i7 920 3837B067 - hope it clocks as well as the one Dj883u2 has ( 3837B080 ) :D
Here is my i920, 24/7 Result:
Batch#: 3837A766
CPU: 200x21(Turbo) = 4200Mhz @ 1.425v, HT=ON.
CPU Diffrential Amplitude: 700mv, CPU/IOH Clock Skew: delay 100ps.
TM Function=Enabled, QPI voltage=Auto, Load Line Calibration=AUTO.
Memory: DDR 1600Mhz 7-7-7-18 1T @ 1.64v
http://img293.imageshack.us/img293/4...001425vna3.jpg
http://img293.imageshack.us/img293/l...jpg/1/w854.png