Quick question on e8400 oc

[Frozen_Canuk]

Hey guys, just a quick question,

I finally got around to running my first OC, pushing from 3 to 3.6GHz on my e8400.

I've heard the whole "every chip is different", so when i say that i've seen someone with my exact setup run 3.6 on a vcore of 1.300, but mine gets killed on prime95 after 2.5 hours at 1.3125 volts, i should just bump up the vcore a tad and say "well, my chip just needs more juice than others" right?

i mean, i didnt think it would be completely different, i've seen guys run a vcore of 1.300 with a clock of over 4GHz.

guess i just have a not-so-OC friendly chip. it is the C0 revision, which i heard sucked compared to the E0.

 

[jeanluclariviere]

I'd think you could push the E8400 to atleast 1.3625 and still be safe (probably more but thats just what reads on intel spec finder)

Could aslo be the NB/SB voltage that are giving you unstable results.

 

[Frozen_Canuk]

Could be, i'm new to this stuff so i'm kinda flying by the seat of my pants as far as some of the specifics. I'll give the main settings that i have, maybe its something that you guys find obvious that i just dont see.

FSB: 400

Multiplyer: 9.0

vcore: 1.3125

NB: 1.40

SB: 1.05

Ram Timings: 5-5-5-12 @ 2.0v

speedstep off

I thought I was alright with the NB voltage, just going by a youtube vid of someone with the same mobo and cpu as me (E8400, ASUS P5K-E) but maybe things differ more than i thought.

Prime95 runs for 2.5 hours, errors, then BSOD. Should my first step be to just bump the vcore to 1.325 and try again? i have lots of room on temps (Prime95 = 51C, woot water cooling!!!)

 

[dechy]

Hey! I've got what seems to be a system similar to yourself.

I OC'ed my E8400 to 3.6GHz from the get-go (when the chip came out), so I've got the so called "OC unfriendly" stepping.

Mine is only needing a very small increase in vcore to run at 3.6GHz; 1.17v if I remember correctly, which shows up at 1.14v in cpu-z & other voltage/temp monitoring software.

I haven't touched my chipset voltages, only set FSB to 400, timings to 4-4-4-12 (I'm using Patriot Extreme, yours might be "normal" Patriot), 100MHz PCI-E & forced 2.2v for RAM.

Disabled as well speedstep/E1 or 1E Intel thing (more power management crap), disabled spread spectrum... basically followed the stickied faq on this site.

 

[hurbt]

Keep an eye on the temps, and don't exceed 1.36v. So long as you do that, and make small incremental increases of voltage, you should be fine.

I'm running my E8500 at core voltage of 1.34v and 3.8ghz without a hitch. I hit about 65C in the worst supreme commander battles. My 4850, however, has seen 98C... even at 75% fan speed... Maybe I need to clean it.

 

[Frozen_Canuk]

Sheesh, I guess every chip really is that much different from the next. My system wouldn't even POST untill it had 1.275 on vcore.

oh, and CPU-Z reads 1.208v while BIOS is 1.3125. Hooray for vdrop.

i'll bump up the vcore a tad and see what happens.

 

[dechy]

haha, indeed, I posted after I hit my small OC goal at how very little I needed to raise it, but it was apparent after that that a lot of people are hitting 3.6GHz on stock cooler WITHOUT raising the vcore.

Huge variance in vcore needs between chips.

That vdroop is MASSIVE, I don't remember seeing one THAT big :O

 

[aznguy0028]

Sheesh, I guess every chip really is that much different from the next. My system wouldn't even POST untill it had 1.275 on vcore.

oh, and CPU-Z reads 1.208v while BIOS is 1.3125. Hooray for vdrop.

i'll bump up the vcore a tad and see what happens.

your Vdroop is INSANE! no wonder you're having such a hard time being stable in prime and needing higher voltages than other people. one thought, enable load line calibration in your bios. if you ahve that option, turn it on, and it should reduce your vdroop, by how much depends on the board, but it should help.

 

[Frozen_Canuk]

It IS enabled. Did it as part of the OC guide that i read to prep for OCing...at least i think i did, i'm not at my pc at the moment so i can't check, will in a bit though. I'm pretty sure its enabled.

I guess I just ended up with a less-than-stellar board/chip. Althougth i know i'm not alone since i've seen someone on youtube (as I mentioned before) with the same board/chip/settings, and he was running with a vcore of 1.300, so its not that uncommon.

Thank god i have water cooling, otherwise i'd think my chip wouldnt be able to make it to 3.6.

any other ideas? i'm wondering if i shouldn't back down the NB voltage a notch and see what happens. Is temp the only thing that comes from raising the NB? i mean apart from helping reach higher FSB. i'm wondering if it might be contributing to the vdroop/lack of stability.

ideas?

 

[aznguy0028]

if i were you, i'd take the NB voltages down to stock. you don't really need to increase NB voltages going from 333 -> 400fsb. unless you know for sure you gotta increase the NB voltage, don't. i'm guessing you're watercooling ur NB as well?

it is a commonly overlooked problem but one nonetheless, a high temp NB could affect the overall quality and limit of your OC. but from what im guessing right now, either bad board or chip. what is your vid anyways? run CoreTemp and let us know

 

[dechy]

BTW, have you taken a "small steps" approach to your OC? I can't tell by your posts, I know I did... it took me about a FULL day of BIOS+reboots+10 minutes of prime95 and all over again until I hit stable+lowest possible settings.

I always raised my FSB by incremenents of 5MHz from 333 to 400, so took a while. Then ran prime, when it failed or system didn't boot I'd raise vcore a single notch until it booted/ran prime95 correctly for 10 minutes, then go back to FSB.

I never touched Northbridge/Southbridge voltages, as those can be finicky, especially since the P5K-E can reach 1600MHz FSB stock (well, from the picture on my mobo box) without actually supporting 1600MHz CPU (so they support OC to some degree).

 

[Frozen_Canuk]

I had no option to do "small steps", though from the sound of a reply on my last thread a while back, i was tol that 333 -> 400 isn't worth doing in small steps on this chip, which should have done it just fine.

my "small steps" came in the vcore increasing, as I started at stock and bumped it up by 0.025v untill it could POST, then i bumped it up slowly by 0.0125v untill windows becamse stable and i could actually run prime95.

i'll try backing down on the NB and bumping my vcore by another 0.0125v to 1.325v. just my luck to end up with a bad setup :S

as far as your question aznguy0028, (and forgive me if i misinterpret the lingo, still learning) i'm assuming by vid you mean my gpu: XFX 9800GTX

as far as temps, is CoreTemp any different than RealTemp? I use it to monitor during Prime95 runs.

 

[Frozen_Canuk]

alright, finally got back home to my rig. i bumped up the vcore to 1.325v in BIOS, and CPU-Z now shows it at 1.224v, so it dropped the vdroop a bit, always good.

i also bumped my NB back to 1.25v (stock), and yes, load line calibration is enabled, so i really do have that bad of a vdroop. I'm starting up Prime95, we'll see how this goes.

 

[doomsdaydave11]

I have the same situation as you op.

E8400 C0 @ 3.6Ghz 1.32v. Most people are able to run 3.6Ghz with stock voltage 🙁. I have the GA-EP35-DS3L. Just don't understand. :|

 

[aznguy0028]

my "small steps" came in the vcore increasing, as I started at stock and bumped it up by 0.025v untill it could POST, then i bumped it up slowly by 0.0125v untill windows becamse stable and i could actually run prime95.

i'll try backing down on the NB and bumping my vcore by another 0.0125v to 1.325v. just my luck to end up with a bad setup :S

as far as your question aznguy0028, (and forgive me if i misinterpret the lingo, still learning) i'm assuming by vid you mean my gpu: XFX 9800GTX

seems like you're doing everything right! what i meant by VID is that, VID is the voltage set by intel that your CPU should run at at stock speed for 100% stability. if you run coretemp, you would see "VID 1.215v" or something like that.

http://www.alcpu.com/CoreTemp/

take a look at the page and you'll see it in the screenshot. so every chip is diff b/c they have a diff. VID. the lower your VID, the less voltage you have to work with starting out and that means higher OC potential at the end. a 1.125v VID chip would be better than 1.25VID chip because you got more headroom to start with. So i was just wondering how high of a VID your 8400 is. just stay under 1.365v for chip, and you'll be fine. good luck

 

[Frozen_Canuk]

I think I've hit the sweet spot!

Haven't done a full 12 hour test yet, but i ran for 4 hours, no errors, perfectly stable. Oh, and I love water cooling (from RealTemp, idle: 31°C, Prime95 max: 53°C).

Final settings:

FSB: 400

Multi: 9.0

vcore: 1.325v

NB: 1.25v

SB: 1.05v

RAM: 5-5-5-12 @ 2.0v

Feels good to finally get this right. And as its my first successful OC, very satisfying. Thanks to everyone here who gave me advice and helped me along, much appreciated.

Interesting thing about my setup, it really does have a terrible vdroop. BIOS: 1.325v, CPU-Z: 1.224v idle, 1.216v load. kinda weird. i just have a crappy chip then.

Just a question, is the VID correlated to the current clock speed or stock? i ran coretemp, it says my vid is 1.225v, which i thought must be a coincidence because thats what cpu-z says my vcore is right now. also, i find it odd that, if it was related to stock speeds, that my bios said my vcore at stock was 1.12v. does this mean that the 8400 is underclocked by intel to 3.0GHz on purpose?

 

[aznguy0028]

VID is correlated with stock speed. do 12+ hours just to make sure it's stable. i've had primes when i was at 4-6 hours and one core failed.

one more thing, in all of my OC experiences with my comp and all my friends, i noticed one thing very interesting, i never read it in any forums but i guess it's just my own experience. my VID is 1.2125v and my Vdroop is around .4v so i up my Vcore in bios around .4v and then i run stress tests. during that time, i notice my Vcore changing in CoreTemp from what i set it in Bios (1.2600) or something around there due to vdroop under load (1.2 - 1.2125)

i noticed that whenever my vcore hits under 1.2125 under vdroop (reported in coretemp), i would have an unstable OC during the prime. so i would bump it up a notch in bios so when under vdroop, my Vcore reported in coretemp would stay above 1.2125 and it's always been 100% stable (5 months and counting) and all my friend's OC are stable as well. and surprisingly, that is the lowest stable vcore i can find. worked for me, just thought i'd share.

 

[Frozen_Canuk]

so your saying that by that theory, i should have my vcore under load at or above 1.225 (VID) as reported by cpu-z during a prime95 run?

Thats pretty interesting, I should give it a test. I'm having to pack up my pc to move it, but when i get settled back in a few weeks i'll run a full 12+ hour test to see if your theory holds. If that's the case with me as well, then it just made OC potentially a little bit easier to find the proper vcore.

Cheers!

 

[aznguy0028]

yea that's exactly what im saying. on my own OC, i had a vcore when under load would give me exactly 1.2125 and it failed after 6 hours or so, so i up-ed it up one notch and that is the vcore i am using now, and it's been stable for almost half a year.

i think we can see it in your situation to a certain degree already. your vid is 1.225 and your vcore now is at 1.325 so that takes into account your vdroop. now we just have to see it under load and see how it performs. let me know how it goes.