Game developer says Intel should recall its defective, crash-prone CPUs

[TempleOrion]

I'll stick to intel myself; my 13700KF still runs with no issues...as do my i7 10700KF , i9 10920X, i7 6850K, i7 5930K, i7 4770K, i5 3570K, i5 2500K, i7 980X, i7 960, i7 930, i7 920, Q9650, Q6600, QX6700, P4 2.8Ghz, and PIII 1Ghz.

Missing the point kinda heroically there champ

 

[loki1944]

Missing the point kinda heroically there champ

Nope; intel has been rock solid for me,; nothing like the post FX5 seies.

 

[HardReset]

It's pretty funny that according to some reports (may be false of course), high end Raptor Lake CPUs used on servers rot within few months.

If true, that means Intel must have skipped normal testing procedures and just hoped it will work. Looking how fast Intel got Raptor Lake out, there may be some truth on that theory. There was around 8 months between Alder and Raptor Lakes, that's Very short time to get new silicon out.

 

[emajin1]

This is the culmination of over a decade of stagnation at Intel. They got really lazy and efficency went out the window. I have to manually put the power TDP limits in the bios for my 13700kf and it still spikes to 80-95c, averaging 80c. Intel calls this completely normal and within the thermal limits.

They really let themselves go w/ this generation.

I thought I was the only one. I actually undervolted my 14th gen as low as I could for stability, with a -0.070v offset and latest bios. Even with Intel limits applied I was still pushing 86c avg in some games. I already swapped out AIOs. It improved min temps by 7c. But that's it. My 12th gen never had this issue.

 

[trparky]

Now there's talk about how the problem might be with the ring bus. The reasoning is that for whatever stupid reason, the ring bus is on the same voltage domain as the CPU cores are. So as the cores request higher voltages to hit higher and higher clocks, the ring bus is being fed the same amount of voltage and getting cooked.

 

[user556]

Of course they're on the same supply. L3 cache is same too. This is all Vcore logic.

The ring bus is relatively sparse. It's unlikely to be a hot spot the way the compute units would be. And reports indicate it is heavy compute that crashes the CPU.

On the other hand, the switching current spikes in the bus drivers will climb faster than most other transistors as the volts rise. It is possible for those few transistors to be overstressed earlier I guess.