very-very good chip!:up: What FPO?
Printable View
OEM E8400 fun, Q801A290, 4050mhz on air w/ an Ultra 120 Extreme.
http://img510.imageshack.us/img510/7...4ghzcf6.th.jpg
The voltage reading lies, it's 1.1125V + 262.15mV = 1.374V which the bios reads as 1.38V. Tried +250mV and error'd after 2.5 hours.
Realtemp thinks that my on-chip thermal diode is misreporting by about +10 or +12'C.
Haven't tuned my ram at all, that :banana::banana::banana::banana:'s on full auto at 2.1V.
i have brand new one Q806A (google doesnt give much answers on this one), i am waiting till my mobo arrives then i will post results
:D Just ordered an Q751A*** - also nothing to find about this batch with GOOGLE........., anyone with this batch here?:cool:
these new batch numbers are nice.
check tankguys.com
Q805A165 for the E3110 --> is $187.99
Q801A290 for the E8400 --> is $199.00
tankguys.com is the site with best prices i think online right now...
^^^
I bet he had C1E enabled and the multi just reverted to x.6
who wouldnt want speedstep or C1E enabled? Do you like your energy bill? It doesnt make a difference in performance whatsoever anyway. I overclock to 3.6ghz.... due to the games I play I see no need for 4.5ghz, I did it once for a benchmark and was satisfied and want to keep this puppy so I overclock 1600mhz fsb and 1.225v in bios and 1.2 cpuid. I think that's pretty good... what do you all think?
:up:
Im pumpin voltage atm, at 4.1GHz, i also just totally burning my Video-cards up by pushing 2 8800GTS 640mb (G80) to 675/1600/900, which stock cooling at 100% fan (Blue wire pulled on both) haha, 1.42v underload, 1.44 idle. 1.46825v in bios, 1.4v CPU FSB (VTT) lawl!!
So, I just got my E8400 from MWave.
Batch: Q808A256
Pack date: 03/26/08
I came from an E6600 @ 3.4GHz
Here's what the E8400's doing after 50 or so minutes of blend Orthos:
ASUS P5E (0605) Vista x64
E8400 @ 4.1Ghz 456x9 @ (1.328V CPUZ)
4GB(2x2GB) Gskill PC8000 @ 1094 (2.1V)
Temps don't go above the mid 50's during Orthos run - this is with a TRUE and silent 120mm's sandwiched on it.
I got a 1M SuperPi of 11.373s compared to my 15.6+s with my E6600.
Vista actually feels a LOT snappier (wasn't bad to begin with) - must be 700Mhz more clock + 2MB more cache.
I'm very happy!! Best $200 I've ever spent on a processor. I'll post some screenies after I get at least a couple hours of Orthos done...
Oh, and no stuck sensors and both read within 2 degrees of each other.
Nice Brahmzy, mine was also from mwave.com I came from an E6600. I'm running mine at 4.15ghz (461*9) with C1E enabled. Load voltage is set to 1.395v. Batch is Q802A321, 3/27/08. If I wasn't a little concerned about degradation I would go for 4.25ghz @ 1.45v or so but it's the point of diminishing returns.
dunno if I ever posted in here
IP35 PRO inside Stacker 830 SE...inverted ATX..
http://img514.imageshack.us/img514/7...utesdl6.th.jpg
http://img339.imageshack.us/img339/6...0095as5.th.jpg
32 minutes:
http://img132.imageshack.us/img132/9...nutesq7.th.jpg
8 hrs on inverted ATX.....
http://img397.imageshack.us/img397/5...8hrsjl9.th.jpg
Now, I'll try regular ATX, or as someone says, "right side up."
http://img355.imageshack.us/img355/8...0097hp3.th.jpg
http://img139.imageshack.us/img139/3...utesoq0.th.jpg
after 2hrs...will finish at 12am 4/8/08
http://img241.imageshack.us/img241/2...2hrsqf9.th.jpg
:up:
yes I was told to run Blend instead...total n00b verified :shakes:
:D
My stupid ram giving me problems, can't bench 3dmark at 4.5Ghz :shakes: :rofl:
I may have been overvolting a little since i thought it was my CPU giving the problems, but it turns out a ram cooler and 2.3v to the memory later, it helped the issues
http://i151.photobucket.com/albums/s...yhm1/19420.jpg
http://i151.photobucket.com/albums/s...perpi45ghz.jpg
Where have the good batches been coming from? I noticed the Retail was only $210 at the egg and Tank Guys has OEM for $200.
So, a question to you all. I don't want to derail this thread, but there are so many people that swear by quads. I don't do any decoding or ripping or anything that really uses more than 2 cores. But, I am constantly hearing about so many people saying that even a lower clocked quad will make everything feel so much smoother - like the OS and certain games etc. Even though they aren't technically multithreaded.
I'm half tempted to pick up a Q9450, try to OC it to 3.4GHz or so and see if it feels any better than my E8400 at 4.1GHz. Should I just be happy and stick with the E8400? How much smoother can things feel with a lower clocked quad? Really?
^^ It already feels "smooth" to me. I cant imagine it being much better. Games are fine IMO... the only benefit would be for Crysis... but even then I doubt the performance gains. What like 1%? Get real.... not for the $$ spent. This sucker screams for the dollar value. Maybe next year I will consider a Quad. Only when core for core it's faster than this E8400. I'm talking all cores disabled except one which is faster? When the Quad is faster than I'll pick one up.
hmm... seems the q801s are no better than the q748, need to get a q802/q808...
the max voltage i want for 24/7 is around 1.36, i'm now pumping 1.41 (1.45 idle) through my POS e8400 to get 4ghz...
Is anyone having temp problems with these? At stock speed/volts mine idles at 50c. and 70c with 1.36v. I cant stress because the temps go over 105c and it just shuts down. I feel no heat what so ever comming from my zalman 9700. I think i got a lemon. Reeeeeehhh!1
It sounds like you have a bad heatsink mount. Re seat your heatsink, and put about a rice sized amount of thermal paste in the middle, let the heatsink spread it out.
Are these temps from the on-die thermal diode, or from the socket sensor? Try running realtemp, most of the on-die diodes are off by at least 10-15'C on recent 45nm chips.
I had to relax my shutdown/throttle temps in BIOS and boost OCCT's halt temp up to past 80'C because my sensors are so off.
I double checked it. It has good contact and there is enough as5 on there. When I touch the hs base it is totally cool. Not warm at all, which is kinda wierd.
I am getting the temp readings from Core temp/hwMonitor/everest/smart gaurdian/RealTemp. They all seem to give different readings. Smart Gaurdian seems to be the lowest. I disabled my shutoff temp in bios but it still shuts down at 105c. I had a E3110 the temps seemed a tad high but nothing to compared to this.
Thanks for the fast responses guys.
might be a bad chip, change em out if you can.
I wonder if you have bad die to IHS contact, that can happen.. You'd have to RMA the chip or cut the IHS off.
I wouldnt try to take the IHS off. It's soldered for one and if you screw it up or find out that that is not the problem there goes the warranty. Try RMA. newegg is pretty good about that stuff. I'ver returned a 400 dollar monitor once and they re-embursed shipping and everything.
I do think you have a bad chip though. Especially if you are positive you have a good mount.
Yeah, it's going back. I contacted egg and they waived the restocking fee and paid for return shipping. Phew. Thanks guys.
I just picked up myself an e8400, so far it can do 4.05ghz with no problem. But what is a safe 24/7 max vcore to run on these chips ? Im using a True with a tuniq fan for reference :) .
IMO it depends on how much heat you can dissipate - on air I wouldn't feel too bad about going close to 1.4V, that's about a 25% overvolt, as long as your long-duration load temps are reasonable. On water you could probably run 1.45V (or even 1.5V if you're a risk taker :D).
Well with a load vcore of 1.33 @ 4.0ghz i reach a high of 52C with my lapped true. And thats after about an hour of occt, but man are these e8400s little speed devils.
I'm jealous of the speeds you folks are getting. My 8400 will not go above 3.825. I can only reach that speed at 1.40 vcore in the bios. It's been pretty disappointing. Any suggestions would be appreciated.
What mobo are you running ? It may just be that its pushing all it can with reasonable vcore :confused: .
Well I'm new here.. so dont bash at me for posting something wrong..
After waiting for 1.5 months (Initially I had a Q745A but gave it to a friend who needs it in a hurry), I got myself a Q802A568.
First try at OCing this thing...
E8400 4.05 @ 1.375V in BIOS, 1.35V in CPUZ
+ P5K-E
+ 2x1gb Ballistix Tracer 1066 @ 2.2 (333 strap, 1:1)
Orthos stable 5 hours. RealTemp says 61C @ load, CoreTemp says 71~75C. Gonna try lower the voltage bit by bit...
Edit: Just changed to 1.35V in BIOS, 1.325 in CPU-Z... lets see how it goes.
Edit 2: 1.35V doesnt work, Orthos failed after 25 mins. Trying 1.365
Q807a241 What You Think?
i think my recent e8400 isn't good, need something better like 4ghz at 1.25v max :P
http://img160.imageshack.us/img160/6...3mediumyz7.jpg
http://img362.imageshack.us/img362/1213/cpu4250wi0.pnghttp://img507.imageshack.us/img507/1...ory4250vy1.png
http://img507.imageshack.us/img507/1...d72e0504c6.jpg
http://img386.imageshack.us/img386/5...5primedoy0.jpg
http://img101.imageshack.us/img101/9...pie8500oo6.jpghttp://img525.imageshack.us/img525/1480/cpu4322bf4.png
Mark made up MHz... i was running at 4.2GHz
first run with E8500 on 3DM06
http://img99.imageshack.us/img99/334...dmark06ry6.jpg
Initially I was passing for hours Prime95 at 450mhz or 4.05ghz with 1.375 in bios. About the same 1.35 you are getting in cpu-z. Then after a few weeks I tried orthos and it failed after 15 mins. No errors just orthos crashed. At 1.38 in failed at 45mins. I didnt want to keep going up so I'm good with my 3.6ghz at 1.225 in bios, 1.20 in cpuz. :up:
I better not lose my overclock. Argh. All this degradation talk makes me nervous....
I've only had my E8400 4 days....
I just picked up an E8400 off ebay batch #Q745A357
Should be here next wednesday, the guy that owned it said he was doing 3.6Ghz on 1.18V:yepp:
This could get interesting...:up:
anyone using a 680-chipset?
Just FYI, my two e8400's did exactly what yours did/are doing. I originally
thought it was degrading, but now I think (and almost 100% know) differently.
It's seems very apparent to me, and others now, that this behavior IS NOT
degradation. The 45nm's (e8400 included) need a little more vcore (+0.024 to +0.050)
to regain 100% stability after initial break-in (or burn-in if you will).
This break/burn-in period will take a week or two (some cases maybe more) to
complete, during this time your CPU will start to fail prime, and orths, etc..
however after the burn-in period, upping the vcore (+0.024 to +0.050) a few
notches will bring it back to 100% prime stability, and it will not fail any more,
or degrade any further.
My two e8400's are now rock solid stable!!! With no signs of slipping any further:up:
Whats some good overclocking ram? this corsair junk i have now craps out at 500FSB at 6-6-6-18 and 2.3 volts :ROTF:
I was looking at the Transcend aXeram 1066 2Gb kit. They use D9's, anyone know of something else in that price range?
Just curious to what you got your Q745 to do?Quote:
745 batch rocks
Bench result -
superpi 1m 9.98 seconds on air with 1.42 vcore @ 4.65ghz
Video (w/8800gts) with 24/7 clock -
3dm05 - 25600
3dm06 - 15600
24/7 -
8.5x500 4.25ghz 4-4-4-12 1.325v in bios
Quote:
Bench result -
superpi 1m 9.98 seconds on air with 1.42 vcore @ 4.65ghz
Video (w/8800gts) with 24/7 clock -
3dm05 - 25600
3dm06 - 15600
24/7 -
8.5x500 4.25ghz 4-4-4-12 1.325v in bios
Very nice...any AQ3 runs with it at 4.2Ghz by chance?
I haven't used aquamark with this chip. If I have some time over the weekend I can try it.
Here is what i got:
e8400 9x450 @ 1.41V (cpu-z load) ~1.46 in bios Gigabyte X38-DS4
FSB Overvoltage control: +0.15V
MCH Voltage: +0.125V
Crucial Ballistix 2x2GB 2.2V DDR2-900 4-4-4-12
I tested it for about 7 hours on orthos without any issues.
I can get it boot with 1.47V (cpu-z) @ 475x9 but it tanks after like 7 minutes of orthos. It doesn't lock up the system or anything, but orthos generates an error. Does anything think i need to bump my voltage on my cpu more, or does my FSB/MCH voltage look a little low? The memory definitely isnt an issue.
Thanks!
Not sure about your bios optiions, but I would start over at 4ghz and try tuning gtl and maybe apply manual memory settings. Try to get stable at 4ghz with as little voltage as possible and keep pll, NB and fsb term low. you should be able to get stable with lower vcore unless your chip is weak.
If it's the just chip holding you back then a bump more vcore will probably help but I can't advise going over 1.4v at all.
memory settings are manual... trust me memory is not an issue, i am only loading the cpu anyways. I dont see and pll or gtl voltage options on my board, i guess they aren't offered on the gigabyte x38-ds4.
What kind of NB and FSB voltages are people using?
Does anyone have an E8400 Q804 batch results? Just got mine in, at work, so I can't play with it quite yet. But have searched everywhere with no one mentioning this batch at all. Retail Box also.
No doubt. Just mentioning some things to try in case you hadn't. Setting the gtls manually can sometimes help you get higher clock with less vcore. I think it might help to ask some people on the same board/bios too to see their settings. Maybe there's a better bios for your board.
Just for general reference, neither of mine need more than 1.2v fsb term and 1.4v nb to bench and 3dmark at over 4.7ghz with 545fsb. So I would say you can try to keep those low at first.
Just for kicks, have you tried using 8.5 multi and higher fsb?
Q749A224 E8500 anyone?
What voltage would y'all recommend for the CPU FSBwith a E8400 on 790i at 3.6 (1600 linked/synced)? The CPU Core is on Auto (1.32v in CPUID).
I had all voltages (except RAM) set to Auto, but was experiencing lagging and freezing in Crysis. I manually increased the CPU FSB to 1.35v from Auto-1.30v and increased the SPP to 1.40v from Auto-1.30v, and it lags less, but was wondering if I should increase the FSB to 1.4v instead of 1.35. The only problem is that it is in red instead of green. Before I put the E8400 in I had a E6750 and it automatically set the CPU Core and the CPU FSB to 1.4 (green).
Well all the e8400's i know can do 3600 with default voltage, mine too. So i think you should try it this way first and stop using auto voltage.
Should increase the CPU FSB to 1.4v or leave it at 1.35v? Either way it is manually set since auto was 1.3v and was unstable.
Geez I don't get why you guys are needing so much vcore to get stable under 4ghz.
1.4ghz would keep my chip stable at about 4.4 - 4.5ghz. I know it's partly because it's a great chip, but even my crappy 748 chip is stable at 4ghz with 1.36vcore. I think you guys need to dig into the settings more.
@mrcape
You have a really nice chip there! :up:
btw, what are your GTL settings @ 4ghz?
I'm seeking to lower my voltages by a few notches.
I guess it depends on the FSB and vtt you use. If I use 8x500 fsb, with 1.15vtt which I prefer, the gtls are 59/59/59. If you are at 4.05ghz with 1.29vcore, like in your sig, that's doing pretty damn well too. For some reason I had better luck with 8x and 8.5x multis for that speed.
I think you should be already satisfied - Your chip does as well as my xeon
Q746A534 at same vcore ( + 4h orthos small ffts stable ) , stable at 3.6ghz for 1.15v load and 3.75 ghz with 1.20v load.
my other Q807A091 is stable at 4ghz for 1.20v but seems to be more unpredictable ( maybe IHS concave )
No dont increase it at all. At 400mhz you should be 100% stable with all voltages in auto (especially the cpu vcore considering auto set's it to 1.325... I set it below auto to 1.225 and I have mine @ 3.6ghz also) on both 780i/790i chipsets. First off are you Orthos stable for over 2 hours? IF you are YOU are not NB/SB unstable. TRUST me you will have errors if you are SPP/MCP was unstable. As far as going out of the "green" zone on your board. DONT DO IT! A lot say to set SPP to 1.5v for above 450mhz stable. I dont agree. I ran my E8400 @ 450mhz or 4.05ghz for months with low "green" voltages on all SPP/MCP/Vcore/FSB. I even kept spp<>mcp to auto. Not a bit unstable. :up:
If you are having Orthos stability issues than maybe (sine you are only 400mhz) RMA that chip or motherboard. Especially if you have to go outside "green" voltages.
The E8400 should be able to go to 400mhz or 3.6ghz without touching voltages (accept) memory, and even should be able to set voltage blow what auto does. If not I would recommend RMA.
This was my first OC so I was hoping all AUTO settings would work. However, I was experiencing lag and freezing while playing Crysis and was unable to run the Crysis Benchmark with these settings --> 1900x1200;All High; 4xAA. I tried the benchmark 7 times and it crashed everytime with a blue screen. Then I went into the BIOS and increased the FSB from 1.30 to 1.35 and the SPP from 1.3 to 1.4 and it ran the benchmark the first time. Also the lag in Crysis has been reduced.
How would Intel accept an RMA like this i.e. can't overclock to 3.6 using default voltages? Wouldn't they just reject the RMA straight away or test it with default voltages @ 3GHz spec (test Ok) and then return it. Has anyone got any experience with this sort of return?
I know my Q745A740 can't do 3.6GHz stable with what BIOS thinks is default (1.225), I have to boost it to 1.248v (load) to be stable. Not everyone has a good batch.:(
Trust me I understand Crysis being very picky and crashing. However you have to run Orthos or PRime95 in order to KNOW if you are having stability issues. Crysis benchmarking tool is not for testing stability. I too have had random crashes with that tool and it had nothing to do with my overclock. It's probably a fluke that it worked the second time around with increased FSB voltage. Use ORthos and PRime95 to find out if you are stable. Otherwise you are working backwards.
Good point. However I dont think he has a bad chip or mobo. I think he is not doing Prime and doesnt know for SURE if it's stable. Crysis benchmark is the stability ISSUE. I have had problems with it as I'm sure others have too.
However... You having to raise your auto voltage to 1.248 is not a problem. It's your motherboard just not giving it enough voltage at that clock speed. My mobo gives it 1.325 @ 3.6ghz. That's way too much. I was able to get it stable at 1.225 in bios. 1.20 in cpu-z. I dont think you have a bad batch also. It's just your mobo not give enough (isn't a bad thing) and requiring more to push it to 400mhz. But by all means 1.25 and below is great for 3.6ghz IMO:up:
So why does coretemp and realtemp temp readings differ so much with the e8400s ? Is it wrong to trust realtemp over coretemp ?
RealTemp assumes a Tj max of 95C (based on the program writer's testing results) compared to CoreTemp and other programs that assume 105C. This means that RealTemp will report 10C lower temps.
RealTemp also attempts to cater for the non-linear increase of temps and 'sticking' sensors at lower temps.
It's difficult to advise which to trust as Intel will not release Tj max values for it's processors, so all programs are having to make assumptions. I can only suggest you read the thread and decide for yourself. Also bear in mind that (I believe) Intel have stated that the DTS sensors are primarily designed to control thermal shutdown and are not designed to reliably report lower temps.
Personally, I don't really care what the absolute temp of my CPU is as long my system is stable and the temps are within generally accepted tolerances. I use CoreTemp to check as it reports higher temps, and as long as CoreTemp reports temps below 70C and my system is stable, I'm happy. As it is, my load temps are 53C (ambient 21C). I'm currently running at 4GHz (8x500) 18+ hours Prime stable, but currently have no screen shot. I'll try and remember to take one next time.
I'm pretty newbie what comes to overclocking. I had opteron 144 before @ 2.7GHz all stock voltages.
Now this is my first WC rig and I don't even get temperature readings out of e8400. It's stuck on 30°C but in reality it should not be very hot. Now to the question.
Do I really have the worst e8400 that anyone could get, or am I doing something wrong? Becourse I really can't OC this processor much with stock voltages (on all, chipset etc). I got like 3.2GHz, whopping 200MHz rise.
With ~1.32V on bios I got 3.69GHz MAX, which was not even 24/7 stable. Just enough to run superpi 1M. I have corsair memory, TWIN2x2048-6400 G. WTF! I just noticed that the shop where I ordered my parts messed up with my memory. I ordered TWIN2x2048-6400PRO, and paid for them and I apparently got some G (what ever it means) memorys and not PROs. Now I'am glad I started writing this post :)
But still I have only ran the memory with divider, about 800MHz. So it should not be the limiting factor of OC. Should I rise voltages of chipset or something else instead of just adding more V to cpu?
Sorry about my bad english, hope you guys understand even something of what I just wrote :)
@natech
It would help if you would also post your specs.
@natech
These processors are known to have 'sticky' DTS sensors and are not really designed to read low temps. I personally have 1 core stuck reading 42C until the temp rises above that value. Nothing to worry about really but I understand it can be annoying.
As the previous poster stated, your specs would be useful in order for us to be of further assistance.
Here is my final 24/7 OC. I think tis will be good to get me through even the hot summer months. RAM is on the 2.00A strap (1:1 with CPU) using 2.05v
System specs:
E8400@4.0ghz (445*9)...2x2gb G Skill DDR2-1000 @ 890mhz...Gigabyte X38-DQ6...Corsair HX620 PSU...MSI HD2600XT 512 Silent ...X Fi Platinum...2x500gb WD RAID0 - 2x320gb Maxtor RAID0 ...Zalman CNPS9700...Ultra Aluminus (black)...Hanns G 19" LCD 1440x900...Logitech Z5300e speaker
http://img378.imageshack.us/img378/8900/445x9xq1.jpg
And this one performs exactly the same as the previous one. Even the core temps are stuck at 45C, just like the first one. But this is a very very good batch. The previous one I had sold to someone on BE that is running some amazing scores on Dice or LN2 (forget which).
4.3GHz at 1.344vcore under load with no effort as far as tweaking goes.
pretty great chips. I need around 1.36v to be OCCT stable at 4.2ghz. :down:
Fireice, that's not so bad man. From what I've seen yours is probably still in the 85th percentile.
Yeah, thats not bad.. it seems the early batchs are the best overclockers with low volts. I'm running 4ghz with 1.248v on load, and haven't even tried seeing if its stable with even lower voltage, yet.
I need 1.38v for 4.05ghz. You guys that are stable with low voltage... that must be nice!
I'm thinking about going higher as well :up:
I need new ram for benching, my ram craps out at anything above 4.4ghz :rofl:. With the 3dmark in my sig, i was giving the cpu far more volts than it needed. I can make it into windows at 4.5ghz with 1.4volts (still overvolting i think, didn't try less) but my ram was extremely unstable.
testing a Q807A067 with a bit bad waffer (8A1642) with air (TRUE),25C room temp and closed case..
e8400 Q750A196 1.1125vid
for 4050(450x9) need 1.368v bad luck
http://www.imageshack.gr/files/1tp2x...e5t3_thumb.jpg
Very nice score. I managed 197** at 4.5ghz (I needed 1.6v though to even get into windows) with 2 8800GT overclocked. I have Vista... You?
This isnt bad luck... this is the norm. I need 1.38 for 450mhz. Anything lower than what you are talking is good luck. Anything above 1.4 is bad luck. :up: