E8400 Hit the wall?

H

Hades2008

New Member
I was wondering if my E8400 hit the wall already. Right now I have it at 370x9. Anything higher then that, Windows loads and that, but when I get into Windows the screen goes black for 4-6 secs, shows my desktop, then black again for 2 secs. I start up P95, and it's fine for about an hour and stops. Both cores stop. When I set it back to 370x9, it's fine. I can run P95 for 30 hours and it's fine. Overheating isn't a problem, as it's watercooled. I'm running the DFI LP DK X48. I don't remember the model off hand, but it's the Plus model. I even put my CPU voltage at 1.5, even know I shouldn't. I only did that to see if it needed more voltage. Got into Windows, ran P95, and after an hour, both cores got an error. This was at 375x9 @ 1.5v. I'm thinking, I was an unlucky guy to get a E8400 that don't overclock well. But any suggests in to what might be causing this. Specs are in my sig. Ignore the P5Q Pro, I just haven't put my DFI in my sig yet.
 
Have you tried setting the memory multi's so your ram isnt running as fast? Or possibly have you tried raising the voltage on your northbridge and memory both?
 
Just so you know, you can still overheat on a water cooled system. ;)

You may also be running into an issue with your RAM. It may not be able to run at the settings you have tried for the FSB. Try turning the RAM multiplier back some as boberboysk mentioned.

You should easily be able to reach a 10% OC on that CPU. Just remember that when you increase the Front Side Bus speeds, you are not just increasing the speed of the CPU.
 
I did what you guys told me to do, and it's still not working. But I did manage to get it to run for 2 hours this time. My memory was at 400MHz *800MHz*. I also set my FSB at 400x9, to get a 1:1 divider. As I did some searching, and decided to try that.

My voltages are when this happened were CPU was at 1.35, memory 2.0, and northbridge 1.3.
 
Bump the cpu voltage up a notch then, see if it fails, if it does, keep bumping it a small notch until it doesnt fail;)
 
Hades2008 said:
I did what you guys told me to do, and it's still not working. But I did manage to get it to run for 2 hours this time.
Click to expand...

i'm sure some will disagree with me on this, but if you can boot into windows and run a stress test for two hours, the overclock is stable. the ones that say you need to prime95 for 24 hours are really just kidding themselves, as the vast majority of users will never do anything with their computers that requires anywhere near that level of CPU usage. really all the typical user should be concerned with is running a half hour to an hour of stress tests and then playing some games for a while. if an overclock can remain stable during something as CPU intensive as gaming, you should be in the clear for anything else you choose to do on your machine.
 
Some people have had to push the V-Core closer to that chip's max of 1.45V and more. Be warned that at that voltage you are over the recommended operating spec for that CPU. It will shorten your CPU's life the higher you go over 1.3625V.
 
DirtyD86 said:
i'm sure some will disagree with me on this, but if you can boot into windows and run a stress test for two hours, the overclock is stable. the ones that say you need to prime95 for 24 hours are really just kidding themselves, as the vast majority of users will never do anything with their computers that requires anywhere near that level of CPU usage. really all the typical user should be concerned with is running a half hour to an hour of stress tests and then playing some games for a while. if an overclock can remain stable during something as CPU intensive as gaming, you should be in the clear for anything else you choose to do on your machine.
Click to expand...

I'll be one of those dissenters. :)

The definition of a truly stable system is one that can run non-stop without issue. If you are crashing after two hours of a stress test, it may crash on you intermittently at lower usage. The good stress tests are designed to flex all parts of the CPU to their limits. When it crashes, usually only one of those parts failed. That part can still fail even under low-use situations. That is the reason why the 24 hour stress test is recommended.

Also, if you are running your system far enough above spec that it will crash on a two hour stress test, you are most likely running at settings that will reduce the lifespan of your CPU. Over time, that "two hour stability" that you achieved will shrink to mere minutes. The CPU may even fail to operate at all at those settings eventually.
 
Zatharus said:
I'll be one of those dissenters. :)

The definition of a truly stable system is one that can run non-stop without issue. If you are crashing after two hours of a stress test, it may crash on you intermittently at lower usage. The good stress tests are designed to flex all parts of the CPU to their limits. When it crashes, usually only one of those parts failed. That part can still fail even under low-use situations. That is the reason why the 24 hour stress test is recommended.

Also, if you are running your system far enough above spec that it will crash on a two hour stress test, you are most likely running at settings that will reduce the lifespan of your CPU. Over time, that "two hour stability" that you achieved will shrink to mere minutes. The CPU may even fail to operate at all at those settings eventually.
Click to expand...

this is in some ways true, but let me put it this way. I have had my E8400 at 4.0 since i bought it, with short periods of clocking back to 3.6 and even back to stock at one point. I had it running 4.0 24/7 for about 4 months. from the beginning of owning this chip, i would stress test 3.6 for as long as I wanted and never have an issue. system was stable under any use. at 4.0, after about 90 minutes of small FFTs/mixed orthos would stop. I didn't worry myself and continued to use my system for all the things I do, like gaming, video editing, flight simulation, web design etc. despite failing orthos at 90 mins in, I was fully able to render a 45 minute clip without so much as a hiccup. a 45 minute video clip takes MUCH longer than the clip itself to render. as you know, video editing is one of the most intense things a person can do to a CPU. I did this for four months.... not a single issue.

my point is this, yes you should do some level of stress testing to initially check for system stability, but running tests like rime95 and orthos isn't the say-all-absolute final word in system stability, because it isn't realistic. it puts an artificial strain on a CPU that will never be duplicated under any practical application. if you look into the source behind orthos and similar programs you will see that 100% with them is not the same as 100% load under video rendering.
 
DirtyD86 said:
my point is this, yes you should do some level of stress testing to initially check for system stability, but running tests like rime95 and orthos isn't the say-all-absolute final word in system stability, because it isn't realistic. it puts an artificial strain on a CPU that will never be duplicated under any practical application. if you look into the source behind orthos and similar programs you will see that 100% with them is not the same as 100% load under video rendering.
Click to expand...

I agree, Prime95 and Orthos will put an strain on the CPU - artificial or not. They are designed to. They are designed to stress ALL functions of the chip, as I said. Practical/everyday applications (even video encoding) are not designed to specifically use ALL the functionality of the chip. My point is that the everyday apps could fail at the settings where Prime95 and Orthos will fail. It doesn't mean that they will quickly or at all in some cases. Stable means eliminating the "could fail" factor. This is why CPUs are released at a specified speed - not a range.

I am not disagreeing with you regarding clocking up and down depending on the need or for short durations. This can work. It does work for some. It won't work for all. Stable will work for all. That is the distiction between our definitions.
 
we have gotten off topic. hades, i suggest resetting your bios to factory defaults and starting fresh because something is not right. from what i hear, that motherboard is a solid overclocker so you shouldn't be having the problems you are. i would go back to 3.0 and stock voltage, and go forward in the slowest increments possible, stress testing after each change you make.

since 1.5v should get you well over 4.0ghz, if you have checked all other variables in the overclock it's definitely possible you got a lemon CPU.

http://forums.extremeoverclocking.com/t299228.html
googling p5q e8400 returns a huge list of forum discussions. try browsing through some of them, you may get lucky and find someone in the same situation as you
 
an average overclock is about 3.6ghz on stock voltages all-around. for some reason, no matter what voltage you pump into your processor, your not getting stable results that are easily obtainable. depending on the processor stepping, (in cpu-z, it would show up as C0 or E0) getting to 4ghz could require 1.45-1.46 volts, or as low as 1.3 volts.

personally, i have a slow chip. it will only run at 4ghz if i use about 1.46V in the bios, about 1.41V in cpu-z.

what i suggest is going back to stock voltages, (reset your bios perhaps) and try to set it at 400x9 again.
 
After I posted about the results, I decided to do it again with the same settings. I left for work, and just got home now. Both cores running still. I stopped. So, 1st time I did it, failed, but 2nd time, it ran fine for 12 hours with no errors. Same settings. I don't know but I think I have a computer that had a mind of it's own.
 
Zatharus said:
:D Welcome to overclocking! :D:good:

Did you start over from scratch with a BIOS reset like DirtyD86 suggested?
Click to expand...

Nope. The last post I did before my recent one, was the pretty much the last time I popped on here, until I posted my last when I read that. And thats when everything worked out fine. I also ran P95 overnight, and was still running with the same settings.
 
Weird. It happens some times. I thought perhaps an errant setting may have been corrected.

Glad it is working though. :)
 
oh. there are some settings you want to turn off or change, for instance, pci-e speed should always be set at 100mhz. cpu and pci spread spectrum should be disabled as well.
 
lovely? said:
oh. there are some settings you want to turn off or change, for instance, pci-e speed should always be set at 100mhz. cpu and pci spread spectrum should be disabled as well.
Click to expand...

Actually all spread spectrums should be disabled, as well as speedstep and C1E.
 
Well, I hit 4GHz lastnight. Ran P95 overnight. No errors. Ran it again after that, and just stopped it again. No errors again. I forgot to screenshot the one from the morning, but it ran for 10 hours.
4ghz-1.jpg
This is the 2nd round. Sorry if the picture is a little big. I don't post pictures on forums very often.
 
Last edited:
You must log in or register to reply here.
Back
Top