I just delidded my newly bought FX 8320 [PICS]

[IronBalls]

Hi,

I bought an FX 8320 the other day to replace my aging thuban. Well I just put it on fast to see if it worked and then I started the delid process.
It was way more difficult than with the 1090T. Had to try twice before the IHS came off. Some pics of the process :

Here the beast :










It is OCCT stable @ 5 GHZ 1.45 v for now. Will do some more testing and report

[IronBalls]

Seems to be working ok so far. Needs a little more than 1.45v so be stable ( 1.4625v). For CPU/NB @ 2600 I need 1.2v. Temps are really high tough, I wasnt aspecting that. Visheras do produce huge amounts of heat compared to thuban.

Two pics I got of the delid process. Saddly couldnt film it on my own.





Visheras are the hardest it can get when it comes to delid soldered CPU. High end intel chips like 4930k must be jokes compared

[Oese]

Really good Job! Have Fun with it

[FlanK3r]

wow... 1.45V and 5 GHz, nice...

[ViViD]

Nice To See Remove FX-8320 IHS Nice Work

I Should Try For My 8120

[IronBalls]

Well, the Vishera 8320 @ 5 ghz and @ 1.45v consumes 430 watts !!!
I might consider clocking it down a little as this seems a little overboard.

[vario]

Wow, nice dude.
430W ,sheesus, mine not delidded at 5ghz 1.55v was 395W (for the system though) although for like 5secs because it crashed next.From my testing power consumptions drops below 300W at around ~1.4v

[FlanK3r]

Maybe he has very powerfull GPU or CLI/CF???

[IronBalls]

Maybe he has very powerfull GPU or CLI/CF???

450w OCCT stressed

Here is my setup

FX 8320 @ 5ghz/1.4625v
2x4 DDR1600
Asus crosshair V formula
AMD 290 (1x)
Samsung SSD 830 128gb
HDD 1T
Blue-ray burner 209M Pioneer
Aquastream XT Ultra
Mora 3 with 4x 180 mm Fan

[IronBalls]

Idle consumption is 180 w

[IronBalls]

I testet voltage given to the CPU manually, I read 1.485 instead of the 1.4625 I put in the bios. Little overvoltage but this cant explain 410w system consumption

[FlanK3r]

Hm, you right. Look at this: http://www.xtremesystems.org/forums/...nr2-by-FlanK3r

[BeepBeep2]

Broke off a capacitor on the PCB...

[IronBalls]

Broke off a capacitor on the PCB...

Right, the capacitor was " glued " to the black silicon. When the IHS popped up, it came off as well. This was no error of mine, just bad luck. This is why I say delidding vishera is extremely difficult.
Nevertheless, the CPU is OCCT stable 2 hours @ 5 ghz and 1.4625v. I have also been playing a little watch dog, no instability. So I guess the capacitor wasnt essentiel.

[FlanK3r]

Hope, you show some benchmarks, it could be fun with 5 GHz+ benchamrks at this system.

[IronBalls]

Hope, you show some benchmarks, it could be fun with 5 GHz+ benchamrks at this system.

What kind of benchmarks do you want to see. Cinebench, 3Dmark ??

[BeepBeep2]

Right, the capacitor was " glued " to the black silicon. When the IHS popped up, it came off as well. This was no error of mine, just bad luck. This is why I say delidding vishera is extremely difficult.
Nevertheless, the CPU is OCCT stable 2 hours @ 5 ghz and 1.4625v. I have also been playing a little watch dog, no instability. So I guess the capacitor wasnt essentiel.

It's possible maybe that the internal die voltage is raised by this? I've heard stories of intel CPUs losing capacitors and overclocking better.

[BeepBeep2]

Right, the capacitor was " glued " to the black silicon. When the IHS popped up, it came off as well. This was no error of mine, just bad luck. This is why I say delidding vishera is extremely difficult.
Nevertheless, the CPU is OCCT stable 2 hours @ 5 ghz and 1.4625v. I have also been playing a little watch dog, no instability. So I guess the capacitor wasnt essentiel.

It's possible maybe that the internal die voltage is raised by this? I've heard stories of intel CPUs losing certain capacitors and overclocking better, like taking .04v less for same clocks.
I'm not the most literate when it comes to electrical engineering...

[andressergio]

Hi,

I bought an FX 8320 the other day to replace my aging thuban. Well I just put it on fast to see if it worked and then I started the delid process.
It was way more difficult than with the 1090T. Had to try twice before the IHS came off. Some pics of the process :

Here the beast :












It is OCCT stable @ 5 GHZ 1.45 v for now. Will do some more testing and report

Good work bro is it soldered ? or just as intel ?

[andressergio]

Broke off a capacitor on the PCB...

Yes Beep to the left i saw it did you IronBalls ?

[IronBalls]

It's possible maybe that the internal die voltage is raised by this? I've heard stories of intel CPUs losing certain capacitors and overclocking better, like taking .04v less for same clocks.
I'm not the most literate when it comes to electrical engineering...

I cant help you with that. Im not even sure that thing was a capacitor. If it was I would have had to make the bridge between the pads for the cpu to work. But I did nothing, it worked like that

[IronBalls]

Yes Beep to the left i saw it did you IronBalls ?

Yes I saw it, even thought about soldering it back as im pretty experienced with soldering. But some pad was completely broken...

[IronBalls]

Good work bro is it soldered ? or just as intel ?

Its soldered. I wouldnt advice you to try this, really difficult compared to thuban. Had to heat it for like nearly two min ...

[FlanK3r]

Yes, some CInebench, 3D Mark, you can try max clock for SUperpi 1M run .

[IronBalls]

Yes, some CInebench, 3D Mark, you can try max clock for SUperpi 1M run .

First 3dMark : Score 9902 with R290 @ 1280 mhz

Score 829

Superpi wont run above 5300 mhz, it crashes instantly.


The funny thing is, I can put the voltage @ 1.575, not only it wont give me better results but the computer turns off instantly.
I think my PSU (650W) isnt enough. When benching 3Dmark I hit 740 w