Hmmm, double check the voltage that is being supplied to the CPU is correct according to the specs.

I do know that one particular fault mode of Athlon 64 CPUs is that they report the wrong voltage requirement to the motherboard and thus the motherboard pumps a bit too much voltage into them, making them run excessively hot.

It's getting 1.55v instead of 1.5v, so not a huge difference. I can't turn it down either which is annoying.

You also should check temps in windows. The CPU is not idling while in the BIOS. It is busy waiting for keyboard input, so the temps will always be much higher than Idle.

Venice chips are fairly cool running.. if you plan on sticking with the 939 system i would get one. It is a later board so bios compatibility shouldn't be an issue (but you can always use the clawhammer to update, should you need it)

Are the wide range of 939 CPUs generally well supported by all motherboards with a BIOS update?

939 was relatively short lived socket and practically everything was supported on any board.

but for some derpy reason my motherboard doesn't have four corner holes

Socket 939 never had four mounting holes. It was a feature of Socket AM2.

I've ordered a 3200+ Venice chip for £4. I'm thinking of pairing this with a 7950 or 8600 (passive) as my 6800 is dead!

And sooner or later your 7950 and 8600 will be dead too. Especially if passively cooled.

Why will a 7950 / 8600 die on this PC? Especially ones passive (by design such as this) There were loads of passive 8600s.

Because they are flawed by design chips, which will die eventually by loosing BGA contact between the crystal and its PCB. Increased heat will only fasten the process.

The Serpent Rider wrote:

Because they are flawed by design chips, which will die eventually by loosing BGA contact between the crystal and its PCB. Increased heat will only fasten the process.

only half correct. The BGA balls are not the issue here, the solder bumps between the die and the chip carrier are. The reason is that nvidia used an unsuitable undefill material that doesn't support the solder bumps properly when expanding/contracting due to heating/cooling cycles.

That is correct. However, that means a passively cooled card will likely last longer than an actively cooled one. It's not running hot that causes damage to the solder bumps, but recurring sudden temperature changes. Since the passive card will generally run hotter at idle and takes a long time to cool off after load, it will go through smaller and more gradual temperature changes.

Koltoroc wrote:

The Serpent Rider wrote:

Because they are flawed by design chips, which will die eventually by loosing BGA contact between the crystal and its PCB. Increased heat will only fasten the process.

only half correct. The BGA balls are not the issue here, the solder bumps between the die and the chip carrier are. The reason is that nvidia used an unsuitable undefill material that doesn't support the solder bumps properly when expanding/contracting due to heating/cooling cycles.

it isn't the whole flip chip bga issue? like on a series of macbooks that the gpu suddenly dies and you can't even reball it properly cause it is a flip chip bga ?

Since the passive card will generally run hotter at idle and takes a long time to cool off after load, it will go through smaller and more gradual temperature changes.

That's not how it works.

.legaCy wrote:

it isn't the whole flip chip bga issue? like on a series of macbooks that the gpu suddenly dies and you can't even reball it properly cause it is a flip chip bga ?

It is the same issue and it is misunderstood. It is NOT the BGA balls that are the issue. that is why reballing doesn't help. The "baking trick" wich is technically just reflowing solder joints will temporally solve the issue since it also reflows the bumps. However that will not last because the undefill is still the problem and once that material is sufficiently compromised (wich is inevitable) even that ceases to work. There is no solution to the problem, cards affected WILL die sooner or later.

frudi wrote:

That is correct. However, that means a passively cooled card will likely last longer than an actively cooled one. It's not running hot that causes damage to the solder bumps, but recurring sudden temperature changes. Since the passive card will generally run hotter at idle and takes a long time to cool off after load, it will go through smaller and more gradual temperature changes.

that is not how it works. ANY temperature change is causing damage in this case not merely sudden ones, which aren't really a thing anyway.

Koltoroc wrote:

It is the same issue and it is misunderstood. It is NOT the BGA balls that are the issue. that is why reballing doesn't help. Th […]
Show full quote

.legaCy wrote:

it isn't the whole flip chip bga issue? like on a series of macbooks that the gpu suddenly dies and you can't even reball it properly cause it is a flip chip bga ?

It is the same issue and it is misunderstood. It is NOT the BGA balls that are the issue. that is why reballing doesn't help. The "baking trick" wich is technically just reflowing solder joints will temporally solve the issue since it also reflows the bumps. However that will not last because the undefill is still the problem and once that material is sufficiently compromised (wich is inevitable) even that ceases to work. There is no solution to the problem, cards affected WILL die sooner or later.

frudi wrote:

That is correct. However, that means a passively cooled card will likely last longer than an actively cooled one. It's not running hot that causes damage to the solder bumps, but recurring sudden temperature changes. Since the passive card will generally run hotter at idle and takes a long time to cool off after load, it will go through smaller and more gradual temperature changes.

that is not how it works. ANY temperature change is causing damage in this case not merely sudden ones, which aren't really a thing anyway.

I know it, that's the issue with the macbooks, Louis Rossmann made a nice video telling about, i guess it was a reply to linus from linus tech tips, cause in one video linus baked a gpu back to working order, and like we know, it will fail again eventually.
For those who don't know what we are talking about.

So essentially it's a design flaw that effects a whole series of cards? I'll probably take my chances I think, it won't need to last long (a couple of months before I get bored of it probably) before I swap it for something else.

However I've got my eye on a 1950 pro at the moment.

Koltoroc wrote:

that is not how it works. ANY temperature change is causing damage in this case not merely sudden ones, which aren't really a thing anyway.

I worded it poorly, I didn't mean that slower temperature changes don't cause damage, they do. But the faster and bigger the temperature change, the higher the stresses on the material. That is exactly how thermal stresses work. Especially when it comes to cooling, which is when materials contract and cracks form. An actively cooled card will cool off faster, leading to faster failure, compared to a passively cooled card.

OP, did you check your temps in windows?

My Athlon X2 system shows 45C in BIOS as well, but fan is quietish

.legaCy wrote:

I know it, that's the issue with the macbooks, Louis Rossmann made a nice video telling about, i guess it was a reply to linus from linus tech tips, cause in one video linus baked a gpu back to working order, and like we know, it will fail again eventually.

Yeah, I've seen that video and you are correct, it was a reply to linus attempt.

This entire issue is where the "baking a card to fix it" myth comes from. It's honestly sad that that people still do this now, particularly since it never was a fix in the first place.

frudi wrote:

Koltoroc wrote:

that is not how it works. ANY temperature change is causing damage in this case not merely sudden ones, which aren't really a thing anyway.

I worded it poorly, I didn't mean that slower temperature changes don't cause damage, they do. But the faster and bigger the temperature change, the higher the stresses on the material. That is exactly how thermal stresses work. Especially when it comes to cooling, which is when materials contract and cracks form. An actively cooled card will cool off faster, leading to faster failure, compared to a passively cooled card.

not in this case. the bumps are so small that damage happens quicker (miniscule thermal mass). Underfill is there to minimize the mechanical stress of thermal expansion and providing additional structural integrity and in many cases additional heat transfer and the material used by nvidia in these cards is insufficient for the job.