**Official e8400/e8500 Retail OC Thread

[ixtapalapaquetl]

I have concluded that VID is not an intrinsic property of the CPU that is directly coded onto the hardware, but rather is a value that is inferred by software which calculates VID based upon another dependent hardware reading. unclewebb has provided technical details before; he may be the one to provide conclusive proof.

I have seen VIDs of 1.0675, 1.0375, and 0.9xxx (can't remember the exact value, but it was less than 1). Variables which I have noted causing change are booting into Vista vs. XP, booting at default settings after resetting CMOS, and core voltage.

CN's results aside, I am presently not concerned with this discrepancy.

[Zucker2k]

what bios you running Zucker?

0907.

[DStealth]

http://www.xtremesystems.org/forums/...d.php?t=173847
This one:



4420mhz 1.528V

[Grnfinger]

I played with these features and could not get my vid. to change whatsoever.

I have tested the 2 in my bios as well.
My VID has never changed from 1.125, I have tested bios 903 and 907.
Under load or idle, C1E & EIST enabled or disabled the VID stays the same.

I have not seen any "degrading" , right out of the box I have ran this chip at 4.3GHz or higher @ 1.38v. I have primed for 24 hours blend and 15 hours small ffts. Ran OCCT for 12 hours and many sessions of 3DMark, chip is still rock solid and have yet to find any issues with it

[cadaveca]

Interesting...nice to see a few of us with the same board getting different results...easier to pinpoint problems this way, IMHO.


Just to confirm tho...multi cannot be set manually or VID change does not happen.

I thought it was weird to see the VID change while coretemp wqas running...we are not talking one boot, one vid, another boot, another vid, my cpu will change vid based on..well..i dunno..but while transitioning from idle to load, and back.

can you other guys confirm sensor data(stuck at one value, and which core, which value it's stuck at) and difference between core0 and core 1 temps? FPO/Batch, and pack date? my work week is alsmot over, so I'll be spending some time on this over the next few days, and will be trying my other E8400's out as well.

[rge]

My chip's VID is 1.125 regardless as well.

Since the first thing I did when got this cpu, is find where I am stable at 4, 4.05, 4.1, 4.15, 4.2....by running orthos every night, and every day while at work....my combined total was 142 hours of orthos runs (recorded... actual may be little higher), 8 hours of linpack, and numerous 3dmark 6 and pcmark5 runs. All of that was at 1.3625 v and less (35hrs was at 1.3625v). Additionally, I ran one 15 mins bench at 1.53v to get sub10spi, and 4 hours of orthos at 1.39v for 4.2mhz.

I just ran 14 hours last night to make sure run at 1.53 for 15 mins did not change anything, and it did not.

I am done with orthos until month from now, I will check couple runs again then. But so far my chip seems rock stable, doing what I want, when keeping chip at/below max VID of 1.3625 (except couple short transgressions).

BTW...wish we could get a single page of data on specifics of what was done to each chip including mobo, vcore and ~time at vcore, batch#, dates, etc...where people saw degradation or death, and what specifically they mean by degradation, ie more volts for ?.

[jcniest5]

Hmm, my Q748A115 was not so good. Did you thrash your current one, the Q745 with high voltage? Don't want you to be bummed on the second one.

Sorry this was reply to jcniest5.

Okay, what about Q746???? then? Pack date is: 1/15/08. Any good?

[Spyrus]

It will go well as most of 746 (like mine)

[laragirl83]

Here goes my BIOS screen for 4GHz

[Damien]

Okay, what about Q746???? then? Pack date is: 1/15/08. Any good?

same pack date and stepping as mine. i'd tell you how it does but i need a cpu to flash the bios :[

[ThatGuy16]

I know the core readings are most likely inaccurate, but on load they read a 7*c difference. Is this normal? The "CPU" (bios reading) says 45 on load while the cores are 52/59 on load.

I'm probably just paranoid

[laragirl83]

I know the core readings are most likely inaccurate, but on load they read a 7*c difference. Is this normal? The "CPU" (bios reading) says 45 on load while the cores are 52/59 on load.

I'm probably just paranoid

No worries, it happens!

[Supertim0r]

cold air

bugged P5K Premium pci-e @ 1x over 400fsb (boot at 399fsb and use setfsb for 3D)

Also limited by the ram 4gb gskill PC8000 (last set of crucial 8500 is dead)

[JacKz5o]

1/15/08 should be better.

[turtletrax]

Interesting...nice to see a few of us with the same board getting different results...easier to pinpoint problems this way, IMHO.


Just to confirm tho...multi cannot be set manually or VID change does not happen.

I thought it was weird to see the VID change while coretemp wqas running...we are not talking one boot, one vid, another boot, another vid, my cpu will change vid based on..well..i dunno..but while transitioning from idle to load, and back.

can you other guys confirm sensor data(stuck at one value, and which core, which value it's stuck at) and difference between core0 and core 1 temps? FPO/Batch, and pack date? my work week is alsmot over, so I'll be spending some time on this over the next few days, and will be trying my other E8400's out as well.

Don't have all the info right now as I am trying to get the chip OC stable on a different system including different mobo, but I can toss some info out.

E8400 Q743A728 pack date 01/04/08 and the vid has changed on the fly in coretemp from 1.100v to 0.9750v.

I also have stuck core temps and if I remember it was 43c on Core "0". I will edit my post when I can confirm, but that was on water and I dont think it goes that low now as it is on air, and reports rediculously high temps.

I have the chip under a Zipang and everest reports high 50c idle @4.0Ghz with 1.2625v and load in low 70c. I also have about a 10c differential between cores at idle and lowers to about 4c differential at load.

Thanks for your help Cadeveca How are the little ones doing?

Edit: Nice work Supertim0r

[cliffy]

3.8GHz piece if cake on 1.22V but need 1.35 for 4GHz so not bad but i have seen alot better.

Any ways heres my highest Super PI to date, Pretty sure my mobo is holding be back so hopefully will be getting new bios as soon as Chinese New Year ends



And now some 3DMarks results with my 3870, could probably bench at higher Ghz but not to keen on frying me 8400 just yet

[WiTcHkInG]

My vid has changed from 1.1125 to 1.050. Was testing earlier trying to find some stable settings when it changed.

Anyone else used speedfan to check the coretemps, mine are showing 5c lower than coretemp.
Speedfan idle at stock is about 25c with fans clocked down to about 3/5 speed.
Coretemp shows 31c.

[Supertim0r]

my vid on both E8400 I have changed from 1.125v to 1.05v

this is P5K Premium to DFI Dark P35

[rge]

Speedfan is still reading tjunction as 100, coretemp 105, if want to read the same, set speedfan>configure, advanced, chip/intel core and set to +5 on both cores.

[Big Lar]

Okay, what about Q746???? then? Pack date is: 1/15/08. Any good?

I have a Q745 packing date of 01/15/08, not the greatest chip, does 4.0 at 1.36v

I've read the entire thread, as well as many more and thought I would add that my Vid in Core Temp has never changed,(1.125v). I do have weirdness as to temps tho. I water cool with water temps at 50 to 53f at all times. Bios shows 2C, SG shows 254c, and core temp has Core0 stuck at 40c, whereas Core1 seems to work with idle of 30C and 100% Load of 45C.

I did up my temp by 15 in bios, which gave me 15c in SG. I guess my only real value is Core Temp Core1/shrug. Tried all 3 bios for the board as well, so no change there. Just an info thread.

Larry

[rge]

Well in searching about 6 forums, so far I have found subjective reports of the following degraded/dead 45nm chips...many link back to here

Most deaths occurred from running above 1.5v and/or benching at ?1.6+, and one died from running fsb 600 for 6 hours 1.4v max. Edit - now one degraded to point of no longer being stable at 4 at any volts, and his was at 1.41v for 3-4 weeks and never saw higher than 1.44 volts and 1.44 for only short time.

Degradation (needing more volts for booting or stability at same mhz), occurred so far at volts 1.4 and higher, many noticed after stress testing orthos/prime/occt at 1.45 and higher, but 2 so far degraded at 1.41 without other significant risk factors. Another report of degradation occurred ? after primed fsb 545 10 hrs and had 511x9 with vcore 1.45 for 1 hour.

I think I am going to keep 24/7 and stress testing at/below VID limit of 1.3625, and for any long stress testing I am beginning to wonder if 1.3625 is safe. Also would not be priming real high fsb, though i would not do that anyways, since I only prime what I would run 24/7.


http://www.csd.dficlub.org/forum/showthread.php?t=5556
E8400 primed 17 hrs at 8.5x500 at vcore 1.35
511x9 vcore 1.45 benching 1 hour
primed fsb 7.5x545 x 10 hours 1.4 vcore.
then 8.5x500 at vcore 1.35 was then only 1 hour prime stable


http://www.xtremesystems.org/forums/...173999&page=45
E8400 needed additional +.075V to be stable at 4ghz (needs 1.45v now)
Known – benching at 8.5x500 at 1.46v and benching 4500mhz at had to be higher volts.


http://www.xtremesystems.org/forums/...21#post2759421
post 1165 E8500 died
Known - benching 1.65v on air, running at 4300 on 1.5V.

http://www.xtremesystems.org/forums/...d.php?t=176287
E8400 Just says using 1.45 bios, 1.41cpuz ? load for 1 week, then became unstable,

http://www.thelab.gr/showthread.php?...0&pagenumber=5
5 E8400's and 2 E8500's post 82 and scanning through other posts
greek forum, 7 people most/all running orthos at 1.5-1.55 volts on water saw degradation within hours or days, and ? all degraded after orthos runs or benching? One saw degradation while running at 1.41 for few days, but now found he had benched at 1.8v in bios, 1.71measured, and now states his chip is burned? translation? dead? another 5 for total of 12 but dont have specifics on other 5 yet (large forum and they have had theirs longer than in US)

http://www.xtremesystems.org/forums/...postcount=3427
E8400 running it at 600 fsb for 6 hrs, no higher than 1.4V

http://www.overclock.net/intel-cpus/...thread-33.html
post 321 ES 9350 died at 1.55vcore, 1.7PLL, benching spi

http://www.overclock.net/intel-cpus/...d-out-yay.html
edit feb 12,2008: new one just went south after running at 1.41 x 3weeks, never went higher 1.44...cant do 4 stable anymore.

Several Qx9650 and some E8400 benching at high volts 1.6+ with dead cpus, I did not bother linking to, I think everyone understands the risks now at high volts.

[RedBull78]

CoreTemp is not reading the temps correctly on these new 45nm cpu's.

I'm using TT Ultra 120-Xtreme and while in bios the temps read 38c at 1.36v but in windows Coretemp read 61c for both cores and while running orthos it showing 2-3c higher. And that is impossible.

Right now while im typing this "Everst utility" reads my
"Cpu-29c
"Cpu Core # 60c
"Cpu Core # 60c


While running orthos, the only one that shows higher temp during load is the first one that shows 29c, up to 40c load running orthos while #1 and #2 only go up by 2-3c which again is impossible.

So I would suggest you all use Everest utility and only monitor the one that reads as "Cpu" only which 100% reflects the temp from the bios.

[Capt Proton]

I assume when voltage values are given regarding the degradation of the chips that it would be the reported value, not the value set in the BIOS?

[ixtapalapaquetl]

Don't have all the info right now as I am trying to get the chip OC stable on a different system including different mobo, but I can toss some info out.

E8400 Q743A728 pack date 01/04/08 and the vid has changed on the fly in coretemp from 1.100v to 0.9750v.

I also have stuck core temps and if I remember it was 43c on Core "0". I will edit my post when I can confirm, but that was on water and I dont think it goes that low now as it is on air, and reports rediculously high temps.

I have the chip under a Zipang and everest reports high 50c idle @4.0Ghz with 1.2625v and load in low 70c. I also have about a 10c differential between cores at idle and lowers to about 4c differential at load.

Thanks for your help Cadeveca How are the little ones doing?

Edit: Nice work Supertim0r

Interesting. I have a very similar batch (Q743A748), same pack date. I too get wicked high temps (but do not have the core differential). And I have also seen "on the fly" VID changes. Are we getting closer to knowing something?

[gfourth]

Well in searching about 6 forums, so far I have found subjective reports of the following degraded/dead 45nm chips...many link back to here

Most deaths occurred from running above 1.5v and/or benching at ?1.6+, and one died from running fsb 600 for 6 hours 1.4v max.

Degradation (needing more volts for booting or stability at same mhz), most occurred at volts 1.45 and higher, many noticed after stress testing orthos/prime/occt at 1.45 and higher. One report of degradation occurred ? after primed fsb 545 10 hrs and had 511x9 with vcore 1.45 for unknown amount of time. One report on greek forum occurred after several days of 1.41 vcore, unkown if person went higher for benching earlier, unknown stress testing.

I think I am going to keep stress testing and 24/7 at/below VID limit of 1.3625 for now, or certainly no higher than 1.4 for 24/7. Also would not be priming real high fsb, though i would not do that anyways, since I only prime what I would run 24/7.

If anyone finds any that degraded at 1.4 and less, please post, especially if have specifics surrounding degradation.


http://www.csd.dficlub.org/forum/showthread.php?t=5556
E8400 primed 17 hrs at 8.5x500 at vcore 1.35
511x9 vcore 1.45 unkown time
primed fsb 545 x 10 hours unknown mhz, voltage.
then 8.5x500 at vcore 1.35 was then only 1 hour prime stable


http://www.xtremesystems.org/forums/...173999&page=45
E8400 needed additional +.075V to be stable at 4ghz (needs 1.45v now)
Known – benching at 4500mhz ? ? how high volts+ ?length of time.
? additional

http://www.xtremesystems.org/forums/...21#post2759421
post 1165
E8400 died
Known - benching 1.65v on air, running at 4300 on 1.5V.

http://www.xtremesystems.org/forums/...d.php?t=176287
Just says using 1.45 bios, 1.41cpuz ? load for 1 week, then became unstable

http://www.thelab.gr/showthread.php?...0&pagenumber=5
post 82 and scanning through other posts
greek forum, 6 or 7 people most/all running orthos at 1.5-1.55 volts on water, and ? all degraded after orthos runs or benching? And ? one at 1.41v seeing degradation ? orthos run

http://www.xtremesystems.org/forums/...postcount=3427
running it at 600 fsb for 6 hrs, no higher than 1.4V

Several Qx9650 and some E8400 benching at high volts 1.6+ with dead cpus, I did not bother linking to, I think everyone understands the risks now at high volts.

Well that first case would be me.
I will try to fill in the blanks:

511x9 @ 1.45 was ran for maybe an hour, if not less, for benching. The highest clock I stressed with P95 was 500x8.5, anything I did above that was for short periods of time.

When I ran that 545fsb I was using 7.5 multi for 4090mhz. Voltage was, IIRC, 1.4.

And just as above, my 500x8.5 1.35 OC lost its stability. I have got it stable again at vcore 1.376 loaded and slightly different gtl's.

Now to throw a twist into the story: I figured out that if I set vcore to 1.42ish or higher, NO MATTER what the core clock, it will not boot. This mobo has a vcore setting and a special add to fine tune vcore, so I set vcore to 1.35 and used the special add to get to 1.45 and then I was able to boot up at 500x9 again. Mobo or CPU? But still, my 500x8.5 clock requires slightly more vcore to be stable...