I FRIED MY CPU!!!

[Steeplerot]

Zebo is on the same wavelength here. I suspect a popped cap. Look for a bulge on the top of one of them.

 

[Zebo]

Originally posted by: SlitheryDee
No I suppose I don't KNOW it's the cpu. What do I look for on my MB to verify that it's not the culprit? I've seen a cpu die in an overclocking attempt before. In that case a new processor fixed the problem. In my own case I'm seeing behavior thats similar to that one. Like whan I try to start my computer, the cpu fan twitches but doesn't spin. Dammit now I don't know if ive made the right decision in getting a new processor. Can anyone give me their opinion? I need to know how likely it is that I destroyed my cpu instead of my mb. Thanks

Pull board, find some light..Look at board..both sides

any score or brun marks? Look at each capacitor..any popped or buldging? Look at mosfets etc for any abnormalities.

 

[Avalon]

I'd say try the new CPU out when you get it...but then I'd be afraid to stick it in that board if it were the board and not the old barton.

 

[Steeplerot]

Originally posted by: Zebo

Originally posted by: SlitheryDee
No I suppose I don't KNOW it's the cpu. What do I look for on my MB to verify that it's not the culprit? I've seen a cpu die in an overclocking attempt before. In that case a new processor fixed the problem. In my own case I'm seeing behavior thats similar to that one. Like whan I try to start my computer, the cpu fan twitches but doesn't spin. Dammit now I don't know if ive made the right decision in getting a new processor. Can anyone give me their opinion? I need to know how likely it is that I destroyed my cpu instead of my mb. Thanks

Pull board, find some light..Look at board..both sides

any score or brun marks? Look at each capacitor..any popped or buldging? Look at mosfets etc for any abnormalities.

jinx

 

[SlitheryDee]

I have to find a screwdriver to pull the board, but a thorough examination of the visible side reveals no obvious damage, burn marks, blown capacitors, etc. I'll check the back in a minute.

 

[VirtualLarry]

Originally posted by: Jeff7181
As someone else said... the video card is the most likely cause... if increasing CPU voltage didn't do anything for 3DMark stability, then the instability was most likely video card related.

That motherboard uses Via's KT600 chipset... Via's Socket A chipsets aren't really the first choice for overclockers... in fact, I don't think they're anybody's first choice for anything. They use dividers to keep the PCI/AGP buses as close to spec as possible rather than simply locking them at 33/66 MHz no matter what the FSB is... so your system might not be stable at 239 MHz FSB, but when you hit 240 MHz a new divider might take effect and bring the PCI and AGP bus back into spec and all of a sudden it's stable.

I'm pretty sure that the KT600 added the PCI 1/6 divisor that the KT400 lacked. But I don't think that any Socket-A KTxxx PCI 1/7 divisor exists.. meaning - the highest FSB that you can go with a KT600, and keep your PCI/AGP in spec, is 200MHz. (Not much for overclocking, is it? No wonder the "overclocker's choice" chipset is the NF2, with async PCI/AGP busses.) Similarly, the KT400 lacks even a PCI 1/6 divisor, so 166Mhz is the max FSB that you can run on it, and keep everything in spec. (I have a KT400 and an unlocked TBred-A, I know that all too painfully well.)

Oh, and for anyone else reading this - if someone tells you "such-and-such volts are OK, I've tried it", that really doesn't mean much, especially if that voltage is more than a tenth of a volt above nominal. Best to go to the CPU mfg's site, their processor "thermal and electrical guidelines" will tell you the max operating and non-operating temps and voltages and things. Do not go above the max DC voltage spec, which is I think 1.8v or 1.85v for a TBred-B. At least, not if you want to ensure that your CPU will live to overclock another day.

Raising vcore should be the last resort for overclocking stability, not the first. (So says someone that once pushed a K6-2 2.2v chip to 3.1v.. that sucker got HOT. It also raised temps, and actually made things more unstable because of it. Raising vcore is not the ultimate overclocking answer to "life the universe and everything". Cooling the core to ~42C, on the other hand, is. )

 

[VirtualLarry]

Yeah, it could have been the board that went too, probably the CPU's VRM section, if anything. More vcore = more juice = more power supplied by the mobo's CPU VRMs = more heat on the VRMs too = greater chance of going up in smoke, if it wasn't designed to properly handle that much power.

 

[Steeplerot]

bulging cap on a bad mobo, maybe that will help?

 

[Steeplerot]

Originally posted by: VirtualLarry

Originally posted by: Jeff7181
As someone else said... the video card is the most likely cause... if increasing CPU voltage didn't do anything for 3DMark stability, then the instability was most likely video card related.

That motherboard uses Via's KT600 chipset... Via's Socket A chipsets aren't really the first choice for overclockers... in fact, I don't think they're anybody's first choice for anything. They use dividers to keep the PCI/AGP buses as close to spec as possible rather than simply locking them at 33/66 MHz no matter what the FSB is... so your system might not be stable at 239 MHz FSB, but when you hit 240 MHz a new divider might take effect and bring the PCI and AGP bus back into spec and all of a sudden it's stable.

I'm pretty sure that the KT600 added the PCI 1/6 divisor that the KT400 lacked. But I don't think that any Socket-A KTxxx PCI 1/7 divisor exists.. meaning - the highest FSB that you can go with a KT600, and keep your PCI/AGP in spec, is 200MHz. (Not much for overclocking, is it? No wonder the "overclocker's choice" chipset is the NF2, with async PCI/AGP busses.) Similarly, the KT400 lacks even a PCI 1/6 divisor, so 166Mhz is the max FSB that you can run on it, and keep everything in spec. (I have a KT400 and an unlocked TBred-A, I know that all too painfully well.)

Oh, and for anyone else reading this - if someone tells you "such-and-such volts are OK, I've tried it", that really doesn't mean much, especially if that voltage is more than a tenth of a volt above nominal. Best to go to the CPU mfg's site, their processor "thermal and electrical guidelines" will tell you the max operating and non-operating temps and voltages and things. Do not go above the max DC voltage spec, which is I think 1.8v or 1.85v for a TBred-B. At least, not if you want to ensure that your CPU will live to overclock another day.

Raising vcore should be the last resort for overclocking stability, not the first. (So says someone that once pushed a K6-2 2.2v chip to 3.1v.. that sucker got HOT. It also raised temps, and actually made things more unstable because of it. Raising vcore is not the ultimate overclocking answer to "life the universe and everything". Cooling the core to ~42C, on the other hand, is. )

Good info there about that via chipset. :thumbsup:
I never said 1.95 was smart on air but as long as his heat wasen't too extreme I would suspect the mobo first. I wouldn't recommend 1.95 to anyone unless you have a water setup IMO. -A bit of a disclaimer there

 

[Zebo]

No offense to jeff or anyone by I ran 2.2Vcore on tbred for over a year day in day out.

2.1 on barton when benching.

Course I water cool so CPU stayed ~50C

That said. SAFE OCing IMO is 10% overvolt of stock vcore on air which is exactly what I say in my A64 OC guide.

(edit: mobiles are exception..you can use desktop vcore to calculate vmax so mobile barton = 1.65 x 1.10 = 1.815 max)

 

[Jeff7181]

Originally posted by: Zebo
No offense to jeff or anyone by I ran 2.2Vcore on tbred for over a year day in day out.

2.1 on barton when benching.

Course I water cool so CPU stayed ~50C

That said. SAFE OCing IMO is 10% overvolt of stock vcore on air which is exactly what I say in my A64 OC guide.

(edit: mobiles are exception..you can use desktop vcore to calculate vmax so mobile barton = 1.65 x 1.10 = 1.815 max)

But as I stated in that thread last year that talked about this... it's irresponsible to say that 2 volts is perfectly fine and won't hurt a thing. It may not... and putting a 200 HP shot of N2O on a stock engine may not hurt anything either... then again... it may.