Re: Computer Flake Outs (was: Re: On Overclocking - READ THIS!)

Edwin Foo (efoo@mit.edu)
Wed, 07 May 1997 02:26:14 -0400


Hi there,

At 01:41 AM 5/7/97 -0400, Nelson Minar wrote:
>Registers are memory too. Anyone know the random failure rate for
>registers? Or all the other parts of the CPU? Failure rates are likey
>to all be very small, and as someone said the probability of a failure
>at the most crucial moment is very very small. Still, it's something
>to think about.
>
>Btw, if someone understands how ECC RAM works with only one extra bit
>per byte could you send me (not the list) email? Where's the magic?
>

Actually, since so far the vast majority of DESCHALL clients (I think) are
running on multi-tasking operating systems, context switches will force the
registers used by a running process to be swapped out to memory to make
room for the next running process. So no matter how you look at it, the
client will use memory to store at least the stack frame and register
contents while it is not running. This happens hundreds of times a second,
so I imagine you'd might as well treat it as a constant memory storage
location.

At least, that's my two cents on this matter.

-Edwin

------------------------------------------------------------------------
The FooBunny | MIT Computer Science '98 - Systems and Architecture
efoo@mit.edu | DEC Cambridge Research Lab - Parallel Computing Group
(617) 225-8837 | Residential Computing Consultant (RCC) - New House

"Love must be sincere; Hate what is evil; cling to what is good."
- Romans 12:9 <><
------------------------------------------------------------------------