Problems since 5.3-RELEASE-p15

[Alex Zbyslaw]

Tuc at T-B-O-H wrote:

>>*If* there are intermittent memory errors, then it could be that SETI 
>>always happens to get them in the data it is dealing with, in which case 
>>it might run perfectly happily but just produce the wrong results.  
>>Memtest is dull, and stops you using your PC, but like Windows virus 
>>scans, it seems like a necessary evil in this case.
>>
>>    
>>
>	I did the standard tests for 24 hours (21 passes) and it didn't fail.
>I ran the 90 minute fading one for 24 hours (8 runs) and it too didn't fail.
>  
>
It's almost a shame when a test like this passes.  It's much easier to 
just replace some memory and get on with things than be stuck with an 
intermittent problem.

Could something be overheating?  Are you still getting random 
segfaults?  Is there *any* pattern you can see?

--Alex