1 |
On Wed, Jul 27, 2011 at 10:52 AM, Alan McKinnon <alan.mckinnon@×××××.com> wrote: |
2 |
> On Wednesday 27 July 2011 15:40:03 Neil Bothwick did opine thusly: |
3 |
>> On Wed, 27 Jul 2011 14:41:33 +0100, Peter Humphrey wrote: |
4 |
>> > Doesn't do that here. When tmpfs is full it starts being swapped |
5 |
>> > out to the swap partition. Perhaps you didn't have any swap at |
6 |
>> > the time. |
7 |
>> |
8 |
>> The default size for a tmpfs filesystem is half the physical RAM, |
9 |
>> unless you specify more as a mount option, it will never use |
10 |
>> significant amounts of swap. |
11 |
>> |
12 |
>> I wonder how effective tmpfs is for PORTAGE_TMPDIR as the builds |
13 |
>> that need a lot of disk space can often require a fair bit of |
14 |
>> memory too, and tmpfs is using it all. |
15 |
> |
16 |
> In this last week someone reported doing actually measurements and |
17 |
> found that using a tmpfs was actually slower. |
18 |
|
19 |
Hm. I wonder why that is; it seems counterintuitive to my |
20 |
understanding of how tmpfs is implemented wrt the kernel's caching. |
21 |
But I haven't red up on that in years. |
22 |
|
23 |
-- |
24 |
:wq |