| 1 |
On Friday 17 Apr 2015 23:33:40 Andreas K. Huettel wrote: |
| 2 |
> Am Freitag, 17. April 2015, 06:18:47 schrieb Alan Grimes: |
| 3 |
|
| 4 |
> > #8,#16; INTERNAL COMPILER ERROR!!!!!! (recently RMA'd some RAM, new ram |
| 5 |
> > had a 1-bit intermittent failure in 32GB, jacked voltage and hoped was |
| 6 |
> > good...) |
| 7 |
> |
| 8 |
> Well, seems it isn't. These are the two memory-intensive packages in the |
| 9 |
> list. |
| 10 |
|
| 11 |
As Alan says, if you are using bad RAM you're playing a rather deterministic |
| 12 |
game. When, not if, the hole in your memory is met by a running process the |
| 13 |
output will go sideways. |
| 14 |
|
| 15 |
If you don't want to return the memory back and get a new module, then you |
| 16 |
could try the badmem or badram patches. I would recommend you get new memory, |
| 17 |
but either of these patches may bring a quicker result and at least confirm |
| 18 |
that the faulty memory caused the emerge failures. |
| 19 |
|
| 20 |
-- |
| 21 |
Regards, |
| 22 |
Mick |
Archives