**Official e8400/e8500 Retail OC Thread

[flesheatinvirus]

They have a new type thermal issue as well that has a lot of peeps pissed, still awesome performance for the money.

[Ymir]

They have a new type thermal issue as well that has a lot of peeps pissed, still awesome performance for the money.

True that. There are some crazy OCs possible if you're lucky enough to score a good chip.

[rge]

RGE, that's good to hear that you haven't seen any major change. I really hope I can sort out a good reason for this other than damage but so far not yet.

I need an even seven extra notches consistently across all of the clocks I had recorded.

Instead of vcore being the culprit, could it be high VTT or high FSB? One chip died after running 600FSB for 6 hours, nothing else and at 1.4 vcore. Another degraded after running high FSB ~ 540. Did you do a lot of stress testing (simulate lot of time) at high FSB or high VTT?

[mrcape]

I always kept VTT, PLL and NB very low on this chip. Never needed them actually. Funny you mention FSB cuz I'm thinking somehow it could be that. The night I was benching (i'm positive the change occured that night) I had the FSB up at 540 - 550 for a bunch of runs of superPi 32m, 1m and some 3dmark, but no prime. I had figured I could get away with that heh.

I remember what you're talking about with the dying chip at 600. I think that was CN's chip.

[Mav451]

Not to bring it up again, the other night I lowered my VTT down from 1.15 to stock 1.10. Unsurprisingly, after reboot, my VID is now 1.0875 (whereas it was 1.0375 with the VTT @ 1.15v).

There's gotta be a connection to all this.

[Mungri]

I wouldn't expect to keep that stable for long, actually.

@bhavv: Very nice for 1.3V--quite nice in general, really Looks as though you got a good chip. BTW, as you say cas @ ~4 gets unstable above 910 Mhz... what about cas @ ~5 and hitting around 1000 or 1066?

I can get to 1150 Mhz with 5-5-5-15 with one pair only, but with both pairs they reach their limit at 1000 Mhz at lower FSB's and dont scale any higher. Above 450 Mhz FSB they are unstable at 933 Mhz+ 5-5-5-15 and cant pass Super PI 32m or Orthos.

Basically at 333-350 Mhz FSB all four gigs can do 4-4-4-12 or lower at 1000 Mhz, but they cant even scale to 1066 Mhz with 5-5-5-15.

Over 450 Mhz FSB, they start to burn T_T

I'm waiting for a 2x2 Gb set that can do 1200 Mhz now.

The Promos chips on G skill HK dont respond at all to higher volts and looser timings, but their Cas 4 scaling is brilliant with one pair anyway. Some people have gotten them to 1066 Mhz cas 4 with one pair.

When I go over 1.35v and 4.25 Ghz, my temps start to exceed 70 degrees, even though my exhaust air is cool. It could be the temperature problem, but at 1.3v it operates within its safety limit

[Metroid]

Where from?

Several places still have them priced at £122~£130.

Just sold my second one on ebay so i might pick up an E8500

Dabs

Metroid.

[unclewebb]

I got tired of trying to chop 0.001 seconds off my SuperPI time so I decided to actually put my computer to good use by developing a new temperature monitoring program. After a couple of days of work, here's how it's looking so far:



A screen with Prime running only on Core0 which is why it's running a little hotter.

Features include: single, dual and quad core support, proper TjMax=95C for the E8x00 series and TjMax=105C for the mobile 45nm chips (T8100-T9500), independent logging of minimum and maximum temperatures and time they occurred on all cores, visual display of current processor throttling or any history of throttling since last power on and a way to calibrate your idle temperatures so they better represent reality.

I have another feature or two to add on like hopefully some code tomorrow to test the on chip DTS sensors but in a day or two it should be ready for some beta testing. No problems so far and the displayed temperatures, after calibration, are very close to the temperatures I recorded with my IR thermometer from basement room temp (12C) to TjMax and beyond. Daughter says it looks ugly so it must be good!

[maxwolfie]

nice unclewebb

[emoners]

I got tired of trying to chop 0.001 seconds off my SuperPI time so I decided to actually put my computer to good use by developing a new temperature monitoring program. After a couple of days of work, here's how it's looking so far:

A screen with Prime running only on Core0 which is why it's running a little hotter.

Features include: single, dual and quad core support, proper TjMax=95C for the E8x00 series and TjMax=105C for the mobile 45nm chips (T8100-T9500), independent logging of minimum and maximum temperatures and time they occurred on all cores, visual display of current processor throttling or any history of throttling since last power on and a way to calibrate your idle temperatures so they better represent reality.

I have another feature or two to add on like hopefully some code tomorrow to test the on chip DTS sensors but in a day or two it should be ready for some beta testing. No problems so far and the displayed temperatures, after calibration, are very close to the temperatures I recorded with my IR thermometer from basement room temp (12C) to TjMax and beyond. Daughter says it looks ugly so it must be good!

Great job unclewebb! so when are we gonna get to try that too?
"Daughter says it looks ugly so it must be good" -> hmm you could put some pink into it

[hersounds]

E8400 Q746A500( AIR) TOP 100 -- 39 th ---



Validation

http://valid.x86-secret.com/show_oc.php?id=320543

ranking pic



http://www.hwbot.org/listResults.do?...true&limit=100

[Ron 61]

I got tired of trying to chop 0.001 seconds off my SuperPI time so I decided to actually put my computer to good use by developing a new temperature monitoring program. After a couple of days of work, here's how it's looking so far:



A screen with Prime running only on Core0 which is why it's running a little hotter.

Features include: single, dual and quad core support, proper TjMax=95C for the E8x00 series and TjMax=105C for the mobile 45nm chips (T8100-T9500), independent logging of minimum and maximum temperatures and time they occurred on all cores, visual display of current processor throttling or any history of throttling since last power on and a way to calibrate your idle temperatures so they better represent reality.

I have another feature or two to add on like hopefully some code tomorrow to test the on chip DTS sensors but in a day or two it should be ready for some beta testing. No problems so far and the displayed temperatures, after calibration, are very close to the temperatures I recorded with my IR thermometer from basement room temp (12C) to TjMax and beyond. Daughter says it looks ugly so it must be good!

Very good. I want to try and thanks for the effort.

[HDCHOPPER]

how bout pent D too unclewebb

[HDCHOPPER]

WTG Hersounds !!

[emoners]

great job hersounds!

[laragirl83]

I got tired of trying to chop 0.001 seconds off my SuperPI time so I decided to actually put my computer to good use by developing a new temperature monitoring program. After a couple of days of work, here's how it's looking so far:



A screen with Prime running only on Core0 which is why it's running a little hotter.

Features include: single, dual and quad core support, proper TjMax=95C for the E8x00 series and TjMax=105C for the mobile 45nm chips (T8100-T9500), independent logging of minimum and maximum temperatures and time they occurred on all cores, visual display of current processor throttling or any history of throttling since last power on and a way to calibrate your idle temperatures so they better represent reality.

I have another feature or two to add on like hopefully some code tomorrow to test the on chip DTS sensors but in a day or two it should be ready for some beta testing. No problems so far and the displayed temperatures, after calibration, are very close to the temperatures I recorded with my IR thermometer from basement room temp (12C) to TjMax and beyond. Daughter says it looks ugly so it must be good!

Is this gonna be more accurate with E8x00 series CPUs than coretemp...?

[Charles Wirth]

My contribution, messing around with an E8400.

I will try to get 5.6Ghz stable, should pull a nice pi 32m

-50c

[flesheatinvirus]

Nice speed fugger, what week is that one?

[Charles Wirth]

I will check when I break down, an awesome chip doing 3D at that speed.

[C-N]

I always kept VTT, PLL and NB very low on this chip. Never needed them actually. Funny you mention FSB cuz I'm thinking somehow it could be that. The night I was benching (i'm positive the change occured that night) I had the FSB up at 540 - 550 for a bunch of runs of superPi 32m, 1m and some 3dmark, but no prime. I had figured I could get away with that heh.

I remember what you're talking about with the dying chip at 600. I think that was CN's chip.

Not to bring it up again, the other night I lowered my VTT down from 1.15 to stock 1.10. Unsurprisingly, after reboot, my VID is now 1.0875 (whereas it was 1.0375 with the VTT @ 1.15v).

There's gotta be a connection to all this.

I am still about guys its just I've given up posting regarding my E8*00 problems as there are too many people that just try to shoot you down as an idiot for blowing up your chip rather than being constructive about it.

The vid thing is strange, sure I've seen it go up & down board to board, with or without C1E or EIST enabled but I have also noticed my DFI sets different vids & cpu voltage (according to CPU-Z) sometimes when you leave the vcore on auto. Sometimes I get a boot at 1.15v with a 1.050 vid & others I get a 1.25v boot with a 1.125 vid.

My 2nd 8400 that I initially thought died in the DFI actually boots & works fine at stock speed all be it with a ton of voltage in my ABIT IP35 Pro. I believe whatever I or the DFI did to the chip prevented it from booting bye lowering the vid to the point that the chip simply wouldn't boot at such low voltage.

I have an E8200 in my Commando & an E8500 in my Abit both have seen crazy voltage & both chips have been pounded hard & are still fresh as a daisy.

Sure it could be something I'm doing with the DFI but in my defence I am just tickling the DFI compared to my other boards & its only the DFI I am having issues with. Also I am back on a Q6600 with the DFI & she's running perfect, imo it just don't like running the 45nm's over 525ish fsb for any length of time.

CN

[flesheatinvirus]

Here is a run of 3DM06 at 4800 with a single 8800 Ultra. http://service.futuremark.com/compare?3dm06=5187654

[Big Lar]

@ C-N
I see you mentioned different Vid's on your DFI, kinda goes along with what I have here. Just for testing I flashed my bios to the original it shipped with, and afterwards my vid is 1.0375v whereas with the latest bios I was running my vid was 1.125v.

Reported cpu voltage lowered to corresponding vid as well. Very odd indeed.

Larry

[C-N]

@ C-N
I see you mentioned different Vid's on your DFI, kinda goes along with what I have here. Just for testing I flashed my bios to the original it shipped with, and afterwards my vid is 1.0375v whereas with the latest bios I was running my vid was 1.125v.

Reported cpu voltage lowered to corresponding vid as well. Very odd indeed.

Larry

Yes its a strange one that only appears to happen on my DFI, I dont get the vid changing on the Commando or the Abit.

I have even lowered the vcore down to 1.1v on the Commando (low as it will go) yet the vid remains 1.1125v with the e8200 it could just be a trait of the DFI but its strange the vid is constant with a new chip then somehow becomes dynamic.

Its interesting to note you have the P35 DK & see the same anomaly.

CN

[ixtapalapaquetl]

it could just be a trait of the DFI but its strange the vid is constant with a new chip then somehow becomes dynamic.

Digital PWM's? I'm not familiar with the other boards.

[eva2000]

nah VID changes with cpu ratio from my Asus Blitz Formula, Commando and DFI LT X38 testing for both my E8500s