Can software cause hardware damage through overheating? (Split from: Kerbal Space Program 1.10.1 is live!)

[Incarnation of Chaos]

32 minutes ago, steve_v said:

Yeah, pretty much the same CPUs but with a slightly different socket to make them incompatible with desktop parts. The usual Intel "make Xeon special" shenanigans.
Plus the coolers mount to the case rather than the board, so you have to break out the machine tools to fit them in a non-server case. I know, I've been there. LGA2011 is a far more sane proposition.

Mhm

Though if i recall there was a mod that allowed you to basically cover up some pins/capacitors and allow running 771 on 775, but that also requires modded BIOS or support in the BIOS. That machine in particular had a BTX board, and was a complete loss due to the fact none of it could actually fit standard components. It was well out of date at the time, so i wasn't too broken up about it.

And yeah, LGA2011 coolers aren't hard to find by any stretch which was another of my main considerations. Also DDR3 is just....so much better...

[kerbiloid]

Also we should recall the horror of 1980s-1990s, the virus(es) destroying the HDD by periodically referring to the sectors on its opposite sides at the resonant frequency.

(Hardly can imagine what is "resonance" for that heterogeneous metal scrap, but respectable people told it's real. Still idk if it's a truth or untruth.)

Edited September 19, 2020 by kerbiloid

[dave1904]

On 9/17/2020 at 6:27 PM, Boyster said:

Hmmm, you guys make a lot of sense, i aint gonna lie i feel my opinion on this matter weaken in every reply.

I kinda, i kinda still want to stand by my logic.

That is easier/more meaningful to control the software effectiveness and quality than try to improve the hardware.

So for me it makes more sense to make the software treat the hardware better than the other way.

I can only speak for my 7 year old 4770K. I have played a total of 5000 hours of KSP with mods, Total war with mods and Stellaris with mods. All these games are extremely CPU intensive. I have played on 4-4.2ghz overclocks, had water cooling pump fail and my cpu was running 100% most of the time playing these games. I do not clean my case as often as I should and right now 1 fan is not working. I smoke in my gaming room also. These are not good things. My 4770k is still perfectly fine. Benchmarks are the same as they were 7 years ago. The reason for this is that my CPU was always within its designed temperature except for my pump failing. Even then my PC shutdown to protect the CPU. Just make sure you have your fans installed properly and that the waste heat can escape and your hardware should last years. 

 

[dave1904]

On 9/17/2020 at 11:13 PM, Lisias said:

Music? This beast has some "music" for you!

You reminded me of this  

 

 

[Lisias]

4 hours ago, kerbiloid said:

Also we should recall the horror of 1980s-1990s, the virus(es) destroying the HDD by periodically referring to the sectors on its opposite sides at the resonant frequency.

(Hardly can imagine what is "resonance" for that heterogeneous metal scrap, but respectable people told it's real. Still idk if it's a truth or untruth.)

In the 80s, the read/write heads were way.bigger than nowadays. And so more prone to the inertia. At that time, we were still using MFM and RLL technologies, not too diferent from what was used on floppies by the way. With the software having direct access to the step motors and solenoids, yes, it's feasible to abuse the mechanism by stepping the motors faster than the design allowed, stressing the heads' mounting and leading them to crash into the platter.

Nowadays every HD has its own CPU for doing this - hard disks are essentially specialized computers atached to a host using some kind of network.

But at that time, the hardware was fully exposed to the userland.

The even older MSX computers had a design flaw that allowed the PPI (a peripheral control chip) to be programed in a way that it could short-circuit itself and burn, rendering the machine without access to joysticks and the keyboard. And yes, they wrote virus for it. And yes, that PPI chip was expensive.

Edited September 19, 2020 by Lisias
Kraken damned Autocompletes

[magnemoe]

On 9/18/2020 at 3:33 PM, Incarnation of Chaos said:

That's the sad part, i can find a dual xeon (v2 if I'm not mistaken, so 32nm) server with 16GB RAM, and PSU from a reliable "recycler" for 100 bucks shipped on ebay....

I literally cannot beat that price unless i go for something like a raspberry pi o.o

But yeah, for a NAS i came to the same conclusion. I'll have to rig a ATX box up for what I'm after. 

Unless you have an basement or garage to put it. 
Problem is that most of the 1U systems are not ATX compatible, and their footprint is huge.
On the other hand you are not an real nerd unless you have an server rack in you basement. 

Good PC cases 7 years was perfect for file servers as they tended to come with room for +6x3.5" now that has gone down a quite a bit. 

 

[magnemoe]

1 hour ago, Lisias said:

In the 80s, the read/write heads were way.bigger than nowadays. And so more prone to the inertia. At that time, we were still using MFM and RLL technologies, not too diferent from what was used on floppies by the way. With the software having direct access to the step motors and solenoids, yes, it's feasible to abuse the mechanism by stepping the motors faster than the design allowed, stressing the heads' mounting and leading them to crash into the platter.

Nowadays every HD has its own CPU for doing this - hard disks are essentially specialized computers atached to a host using some kind of network.

But at that time, the hardware was fully exposed to the userland.

The even older MSX computers had a design flaw that allowed the PPI (a peripheral control chip) to be programed in a way that it could short-circuit itself and burn, rendering the machine without access to joysticks and the keyboard. And yes, they wrote virus for it. And yes, that PPI chip was expensive.

And back before 2000 nothing had overheat protection, today cpu's are thermally tootled. The turbo frequency is because running an single core cpu generate much less heat, its still frequency limited. 
As other say if system is stupid designed as some laptops you can get problems, but that is an bad design derision  or as we call dangerous overclocking. 
Or not dangerous if you don't have enough cooling but you override the thermal limits.  

Games are among the few software common outside device drivers who can can cause blue screens on windows xp and newer, this is because games has an mode who is unsafe. 
The NT core and the Vista core we use now and the 386 and later cpu uses an memory mode who don't let you access other resources. 
This is why old dos and windows 98 games don't work on newer systems as they typically used direct access to hardware on cd rom to use low level functions to check for bugs on the disc, later OS, hey that are you doing get out. 

However since xp games has some freedoms other has not, assume its stuff like writing into graphic memory. 
Its to increase performancy and its assume you don't play games on computers doing critical stuff anyway so if you allow some chance that an game crash can cause an blue screen. 

And for fun replace the windows 10 login screen with an blue scene image. 

 

[Incarnation of Chaos]

9 hours ago, magnemoe said:

Unless you have an basement or garage to put it. 
Problem is that most of the 1U systems are not ATX compatible, and their footprint is huge.
On the other hand you are not an real nerd unless you have an server rack in you basement. 

Good PC cases 7 years was perfect for file servers as they tended to come with room for +6x3.5" now that has gone down a quite a bit. 

 

Note that the NAS would be a seperate machine, not built from the same server i mentioned prior. It would just be connected to it via a switch and Ethernet, because the server would be doing routing.  Also i'm planning on either a garage or closet with acoustic paneling

And yeah, the irony of it all. Bulk storage has never been cheaper, but cases are designed for SSD/NVME now instead of spinning disks.

14 hours ago, kerbiloid said:

Also we should recall the horror of 1980s-1990s, the virus(es) destroying the HDD by periodically referring to the sectors on its opposite sides at the resonant frequency.

(Hardly can imagine what is "resonance" for that heterogeneous metal scrap, but respectable people told it's real. Still idk if it's a truth or untruth.)

Likely has something to do with the rotational rate of the platter, which could easily be retrieved from the target system and then the resonance solved for with a bit of math. Though in the 80's and 90's that would actually be pretty taxing....i'd imagine they'd use a lookup table or something similar instead of calculating directly.

For a similar, but distinct modern example i'd advise you to look into StuxNet.

Edited September 20, 2020 by Incarnation of Chaos