[mdlug] ECS A770M-A motherboard SATA controller failure

[Raymond McLaughlin]

Aaron Kulkis wrote:

> Yes.  Overclocking is asking for trouble.
> 
> Most over-clockers don't seem to understand how the electronics
> industry works:
> 
> The do NOT set out to make a patch of parts to a specific
> set of electrical specifications.  Instead, they make a bunch
> of devices (resistors, capacitors, chips, whatever...) and
> then TEST whatever comes out of the manufacturing process,
> to see what specifications it meets.
> 
> In other words, if a CPU is in a 2.2 GHz box, that's because
> it FAILED to run flawlessly when they cranked up the clock
> to 2.33 GHz.

Not necessarily true. It is my understanding that they do not set out to 
make parts of each specific speed rating, They do set out to make parts 
that will fall within a specific range. A part marked as 2.2 GHz will 
have passed testing at that speed. It may, or may not have been test at 
higher speeds. If they already have as many 2.33GHz, or higher, chips as 
they expect to sell within a given time frame, then they just don't 
bother testing at higher speeds. If they are getting good yields you 
chip might, or might not, be able to function at a higher speed than it 
is marked.

I have long thought that, if they really cared about overclockers, they 
could use tiny fusible links or other irreversible indicators of the 
maximum speed a chip had been run at.