| 1 |
Neil Bothwick wrote: |
| 2 |
> On Wed, 27 Jul 2011 14:41:33 +0100, Peter Humphrey wrote: |
| 3 |
> |
| 4 |
> |
| 5 |
>> Doesn't do that here. When tmpfs is full it starts being swapped out to |
| 6 |
>> the swap partition. Perhaps you didn't have any swap at the time. |
| 7 |
>> |
| 8 |
> The default size for a tmpfs filesystem is half the physical RAM, unless |
| 9 |
> you specify more as a mount option, it will never use significant |
| 10 |
> amounts of swap. |
| 11 |
> |
| 12 |
> I wonder how effective tmpfs is for PORTAGE_TMPDIR as the builds that |
| 13 |
> need a lot of disk space can often require a fair bit of memory too, and |
| 14 |
> tmpfs is using it all. |
| 15 |
> |
| 16 |
> |
| 17 |
> |
| 18 |
|
| 19 |
|
| 20 |
I posted this somewhere but anyway. I tested this with my 16Gbs and |
| 21 |
some fairly good size packages. Most of them took a little longer to |
| 22 |
compile on tmpfs than it did on the hard drive. So, putting portage's |
| 23 |
work directory on tmpfs doesn't help a bit. |
| 24 |
|
| 25 |
Dale |
| 26 |
|
| 27 |
:-) :-) |
Archives