Newer CPUs difficult to kill by heat?

[Gimpymoo]

I remember in the olden days, a CPU running too hot for too long, would simply die.

I built my 4770K rig a few years ago and left it with the stock fan as it was adequate for what I needed.

I installed a R9 390 just before Christmas. It does get a bit toasty and I forgot about the stock fan on the CPU.

I check a week or so ago and realise it has been running nigh on 90 degrees and throttling itself reguarly at 100 degress which it is TJ. max. It is also reducing voltage as well to help lower temps.

This got me thinking, compared to what used to happen, I found it all very clever.

Much better than a dead PC

 

[AiiR]

I think its near impossible to kill CPUs with heat alone these days, voltages are different though I think.

 

[joeyjojo]

True the throttling is very effective at preventing "sudden death".

When chronically overheated though they're probably more fragile due to the smaller lithography.

 

[beany_bot]

I watched a video on youtube of a guy trying to melt a cpu. The amount of hoops he had to go through to get rid of all the trottles was staggering. Even when he finally did. It didn't stop working till like 190 degrees!

The thermal cutoffs are in a very safe range.

Really makes you wonder why people water cool and cry when their CPU goes above 65 degrees lol.

Thermal throttling is bad because you are loosing performance. So keep it under that temp for that reason. but really I don't think anything under that temp is in anyway dangerous.

 

[Nasher]

Some modern components can take a LOT of heat. Some GPUs don't even have heatsinks on the vram these days because they can run at something like 150c and be fine.

I remember purposely frying an old Cyrix CPU at work for a laugh about 16 years ago, by simply turning it on without a heatsink. Smoke came from it almost instantly

 

[malachi]

Some modern components can take a LOT of heat. Some GPUs don't even have heatsinks on the vram these days because they can run at something like 150c and be fine.

I remember purposely frying an old Cyrix CPU at work for a laugh about 16 years ago, by simply turning it on without a heatsink. Smoke came from it almost instantly

That happened to me with an old K6 CPU.......but that wasnt an accident.

 

[Gimpymoo]

I watched a video on youtube of a guy trying to melt a cpu. The amount of hoops he had to go through to get rid of all the trottles was staggering. Even when he finally did. It didn't stop working till like 190 degrees!

The thermal cutoffs are in a very safe range.

Really makes you wonder why people water cool and cry when their CPU goes above 65 degrees lol.

Thermal throttling is bad because you are loosing performance. So keep it under that temp for that reason. but really I don't think anything under that temp is in anyway dangerous.

Seeing as thing has been running at 100 degrees almost daily since December I assume and is now running a 1Ghz/600Mhz Turbo overclock, the proof is in the pudding?

 

[NieNie]

I'm not sure what you call newer but back in the Pentium 4 days I built a system with a Prescott 3GHz HT processor. I was young and simply wanted a gaming machine so I cheaped out on just about everything to cool it. It ran with 50-60 degree case temps and 95 degree processor temps whilst gaming going up to the thermal throttling limit multiple times per day (100 degrees if I recall). The computer ran in those conditions for 7 years and if I had kept it would continue running to this day. I didn't change any settings, no overclock or overvolt (or under) on anything. I just neglected cooling on what was an already very hot processor (for the day).

 

[philo-sofa]

If it's of interest what's called a thermal cascade failure, which is where the chip tries its hardest to explode. Basically as chip's temps get hotter its resistance rises, which increases its resistance, which increases its temps, which... well usually things don't head too far borth (it's a Log progression for the most part) but at a certain temperature this really becomes an issue. I remmeber the Intel Core series had about a second to turn off once it reached its max temp (iirc 105 degrees(), otherwise you'd get a thermal cascade. However all chips made in the last decade have microcode to auto shut off once they get near that point. There's a great video of a French guy deliberately exploding (it popped, shattered and smoked) an older AMD chip and shouting "Oh myyy gooowd it's too extraaame?!?!" on youtube.

That aside, chips can take a lot of heat and I wouldn't expect CPU/GPU temps to affect life expectancy that much, although some other internal components are more sensitive.