The 860 price made me buy a Xeon 3440. Although I don't think 4 GHz is in sight of air cooling, a man can dream, right?
The 860 price made me buy a Xeon 3440. Although I don't think 4 GHz is in sight of air cooling, a man can dream, right?
Sometimes a good slap in the face is all you need
Bios my arss.....
I can fix this problem with a hardware mod....Hipro5
"Overclock till death. Overclocking is life." Hipro5
You had to say itOriginally Posted by Aerou
Looks like a dead mobo
Tried two different PSUs but same issue spinup but no boot (if you get my meaning) even after a CMOS reset
Wait till I upload the pic of the protection mat!
You'll have to report in and let us know what you do get out of it. I'm real curious about that chip now.
Sorry to hear that Emu.You had to say it
Looks like a dead mobo
Tried two different PSUs but same issue spinup but no boot (if you get my meaning) even after a CMOS reset
Wait till I upload the pic of the protection mat!It was in the name of science though.....Much like benching is.
I'll be interested in your pics. Did it burn a pin like the one in other thread?
Good to hear you have another on the way though. The test must go on.
Bob
HI all, this has been a VERY interesting thread to read, lots of great info. Bob/EMU you have spent a lot of time and energy and it's greatly appreciated.
I purchased an i7 860/Gigabyte UD4P a few weeks ago and have been loving it so far. It's mainly a GPU Folding/Surfing rig so I don't really need to push her too far. I did a few benches at up to 4.6 GHz @ 1.44v and then settled back to 4.0 GHz @ 1.30v (UD4P overvolts a little under load with LLC enabled, so CPUZ states 1.31v under load) but now just run it at stock.
Read the burned socket thread and freaked, so I pulled my processor out and looked at it and the socket. TG no damage at all
OK, the reason for my post, just wanted to share my 4.0 GHz voltage (prime stable, but DID NOT do a 24 hour prime run with it though, still a little paranoid of the burned socket thing still) but here's a screen shot.
Oh, almost forgot, this is on AIR = Mega Shadow.
[IMG][/IMG]
Thx for posting RJR. Good to have a bit more data. I guess I got a bit of a dog CPU. Runs rock solid at 3.7 though.
Oh, and welcome to the forum. I see that's your first post.
Bob
could you try running sciencemark2 clock for clock? it would be a nice benchmark to represent crunching and you wont have to deal with different WU's.
Don't have it. Got a link for a 64 bit Vista capable version? All I see is 32 bit.
32bit version works fine on 64bit. i can run it on vista x64.
http://www.techpowerup.com/downloads/170/
Yeah me too!You'll have to report in and let us know what you do get out of it. I'm real curious about that chip now.
Yeah bummer hey. I wasnt even pushing it that much harder than before so not sure what happenedSorry to hear that Emu.
I'll be interested in your pics. Did it burn a pin like the one in other thread?
Good to hear you have another on the way though. The test must go on.
Bob
Took the CPU out and gave it the once over and nothing likewise for the socket. I cannot see any damage on the board at all and the "sniff" test didnt turn up any burnt out parts on the mobo.
the bit that did catch my eye was the insulator (from the box) that I had the board sitting on. It melted away under the PWM
Oh and I dont actually have another board atm just means I will have to go and get another one(maybe a Giga). Will try taking this one back to the store.
Chum, OK it's works better when the operator pushes the dang start button....
Here you go.
920
860
Not sure what this benchmark means, but there you have it.
The more critical one is how it runs BOINC. I put my stock in that since it's the reason we have crunching rigs.
Regards,
Bob
Where is it man? I want to see the burn and I am not talking about the sun.....Wait till I upload the pic of the protection mat!
Why is the molecular dynamics test score higher on the 860?
Well, the 1156 socket is physically much smaller than the 1366 socket. Therefore, the 860's molecules have to be closer together than the 920. Since they are closer, the molecules are more dynamic, much like cramming 10 people in a VW bug. At least that's my story and I'm sticking with it.....
In other words, I have no clue.This was a 32 bit benchmark program from 2005.
Bob
I think it was more of a melt, but you get the idea!
hmmm maybe I need a benching table?
Some more stats for you all! (well those of you still following this thread!)
I managed to raise the QPI on the 860 as per the screen below but am still struggling with the 920
So Total run time of 98:58:20 (for 24 WUs) which is an avg of 4:07:26 per WU. Which is 2:14 faster on average per WU than the other 860 case with a slower QPI - 3438MHz but the same core speed 4GHz. Now remember this is the same core speed ie 4GHz but with a dropped multi to increase the QPI.
Food for thought for those crunching on all i7s I would think!
Might have to see how high I can get the QPI on the i7!
Currently I am running them both at stock speeds with the same WUs I know they wont be the same core speed but it gives a baseline for each CPU which might help with energy cost estimates as well.
Last edited by emuexport; 10-09-2009 at 06:55 AM.
Good info Emu. This helps quantify the effect of QPI. When I raised QPI on my 860, I knew it was faster but not how much since we were using the old test method.
Question, I notice CPUz reports what I'm assuming is idle volts at 1.117? Did you get that clock with a LOT lower voltage this time? If so, WOW!
Good to hear you got them to exchange your old board.
Bob
Great work guys!
This thread reminds me so much on the Mythbusters show.
I can literally hear the two guys (Adam and Jamie) say:
"Result: Myth confirmed
What can do we do next ...
... we gonna blow it up ... yeaaaah! ... ha ha ha ..."
Emu, try booting the fried board and cpu with a heatsink just sitting on top, but not tightened or anything. JUst a bit of paste, slightly press and check if it boots.
Sometimes a good slap in the face is all you need
Bios my arss.....
I can fix this problem with a hardware mod....
"Overclock till death. Overclocking is life." Hipro5
I thought it would be a good test seeing as jcool pointed out that it would be a differentiator. After these stock core tests finish I think I will try adjusting the QPI with stock core speeds if possible and see what difference that makes.Good info Emu. This helps quantify the effect of QPI. When I raised QPI on my 860, I knew it was faster but not how much since we were using the old test method.
Question, I notice CPUz reports what I'm assuming is idle volts at 1.117? Did you get that clock with a LOT lower voltage this time? If so, WOW!
Good to hear you got them to exchange your old board.
Bob
Sorry mate Idle CPU-Z the 920 @ stock was at 0.95VCore last night when idleI never knew it was that low stock as it went straight from stock to 4GHz as soon as the OS was installed!
We gonna burn the PWM area of the foam below!Great work guys!
This thread reminds me so much on the Mythbusters show.
I can literally hear the two guys (Adam and Jamie) say:
"Result: Myth confirmed
What can do we do next ...
... we gonna blow it up ... yeaaaah! ... ha ha ha ..."
Next test Adam!
Aerou the store replaced the dead board for me no worries so now I have a replacement. I did try what you suggested above but no go. So cheers for the help anyway.
The new board is crunching away on these test WUs with no problems.
Currently they are both running stock BIOS settings and the 920 is getting its arse whooped by the 860
They both have the same core speed ie 2800MHz (due to turbo) and the QPI link (in CPU-Z) is 2400MHz on both of them. However the NB frequency is 2133 on the 920 and 2400 on the 860. The 920 is spending 10 hours on WUs that the 860 finishes in less than half the time
I'm gonna stop these WUs on the 920 as this is just weird.
I'll post up a screen dump if someone wants to throw in their two cents!
I got an update for you guys. I got my Prolimatech Megahalem the other day, but not time till last night to get it installed
I had it running the X3440@4GHZ@ 1.36v
I turned off the PSU, turned it back on and watched the mobo; I saw the led's come on for a split second and shut off. Tried to power back up and nothing. The led's only light up for that fraction of a second if I unplugg the psu for a few seconds then plug it back in; so obviously something is shorting out or fried.
I do not have money or spare parts to try and figure it out. I really do not know what to do; I am guessing mobo first. Might have to go to local store to get Raped, then send this one in for RMA
Hope you have better luck than my UD2, unless my problem is in the cpuYou two guys MADE me purchase an i860 with a Gigabyte mobo. Arrives today, so come Saturday we will have a pool of 3.
I know what you mean, you and Bob are the ones that got me to buy mine
I think it will be easy to do 4ghz on air with X3440, unless this is what caused my problems
I hope mine is just a dead mobo as well. I tried a different psu too, but no luck. I can't even get anything to power up on mineYou had to say it
Looks like a dead mobo
Tried two different PSUs but same issue spinup but no boot (if you get my meaning) even after a CMOS reset
I will keep you posted on my findings; I am not sure what caused this, but if it happens again, I will sell what I have.
It sounds similar to what happened to mine. What voltages were you running?
When mine died I was trying for 4.2GHz after this board sat at 4GHz for a little less than a month. The power would only come on for a second and then go straight off.
Take it back to the store and see if they will swap it out for ya.
I was only running@1.36v.
It is definately the mobo. If anyone is familiar with the Gigabyte UD2; In the upper left hand corner just below the cpu fan header there is a cap. Just to the right of that I can clearly see one of those little chips is burnt up around it. I will try to take a decent pic for everyone to see. There is also a little resister fried by this little chip.
The bad thing is I got it from ewiz; will have to try to rma. If not will have to RMA through gigabyte. I plan to go to my local store to see what they have for P55 mobo's. I think they carry the UD2, but I do not know if I want one anymore.
EMU: wasn't yours an eVGA? That would mean its something with the p55 chipset/ or design in general. But others have had great luck with much higher voltages. Did we miss an important note sent out
Then the question would be, what are safe voltage levels to stick with, 1.36 is not too terribly much; even Anandtech went higher than this I believe on their review; can't recall.
Such a sad day; can't wait to tell the Wife I already broke her new puter
Yeah mine was EVGA but I had a lot more juice going through mine! Mine is the EVGA P55 SLI one so it is not a budget board and should be built of better components than the cheaper models you would think. Also seeing as they pitch it as a OC board it should be able to do the volts I had through it originally. I should have checked the revision on the board as maybe they have changed some stuff already.I was only running@1.36v.
It is definately the mobo. If anyone is familiar with the Gigabyte UD2; In the upper left hand corner just below the cpu fan header there is a cap. Just to the right of that I can clearly see one of those little chips is burnt up around it. I will try to take a decent pic for everyone to see. There is also a little resister fried by this little chip.
The bad thing is I got it from ewiz; will have to try to rma. If not will have to RMA through gigabyte. I plan to go to my local store to see what they have for P55 mobo's. I think they carry the UD2, but I do not know if I want one anymore.
EMU: wasn't yours an eVGA? That would mean its something with the p55 chipset/ or design in general. But others have had great luck with much higher voltages. Did we miss an important note sent out
Then the question would be, what are safe voltage levels to stick with, 1.36 is not too terribly much; even Anandtech went higher than this I believe on their review; can't recall.
Such a sad day; can't wait to tell the Wife I already broke her new puter
One question did you have airflow down onto the mobo?
Yeah I would like to know what safe voltages are. I think Bob posted a pic of Intel max voltages on this thread or my LGA1156 thread.
Hope the store helps you out with it.
As you can clearly see the damage. Also notice the discolored area , more damage than I though. Obviously something got a little warm and/or just fried. What does everyone think I should do. If EMU had a similar problem on his evga, makes me thing its in the p55 board design or something. I might get it rma'd and sell the mobo when it comes back and sell me X3440
hold on...pics were too big...k think I might have it now. Its easier to see all the detail in the full pic,but I do not know how to get it so these zoom out
Last edited by INFRNL; 10-09-2009 at 07:15 PM.
Is that one of the PWMs?
Mine didnt even smell mate. so not sure mine was a burnout etc
Posting Permissions
Reply With Quote
Bookmarks