1 |
Neil Bothwick wrote: |
2 |
> On Thu, 06 Sep 2012 16:09:12 -0500, Dale wrote: |
3 |
> |
4 |
>>> Reading the tarball has nothing to do with this, we are discussing |
5 |
>>> filesystems for PORTAGE_TMPDIR, not DISTDIR. It's where the source is |
6 |
>>> unpacked, the object files compiled to, the executables linked to and |
7 |
>>> the install image created that is relevant to TMPDIR. |
8 |
>> Well, on my system, when I run emerge, it has to go read the tarball |
9 |
>> from the drive before it can unpack and do all the rest that needs to be |
10 |
>> done. |
11 |
> Of course, but it is reading from a different filesystem that is |
12 |
> unaffected by your choice for $PORTAGE_TMPDIR. It has about as much |
13 |
> relevance as the brand of mouse you are using. |
14 |
> |
15 |
> |
16 |
|
17 |
|
18 |
But whether it is on tmpfs or just regular memory doesn't matter. Once |
19 |
emerge starts, everything is in ram including portages work directory |
20 |
which would be on tmpfs here. That's why it doesn't matter if portage |
21 |
is on tmpfs or not. Once emerge loads up the files, it's the same |
22 |
thing. That's why using tmpfs doesn't matter. I knew that the whole |
23 |
time. The amount of ram on a system doesn't matter either. If you have |
24 |
a system that doesn't have a lot of ram, then you can't really use tmpfs |
25 |
anyway. That is not something I would recommend to anyone. |
26 |
|
27 |
I just don't agree that one should *disable* cache to run the test since |
28 |
no one would disable cache on a normal system. It's not a memory speed |
29 |
test. It's a test to see if putting it on tmpfs makes it faster. The |
30 |
fact that emerge loads everything up in memory when it starts is not |
31 |
relevant for what I am testing. It does that on its own anyway. |
32 |
|
33 |
Since portage and the kernel does this in the most efficient way |
34 |
already, I still say putting portage's work directory on tmpfs is not |
35 |
needed UNLESS a person needs to save wear and tear on a drive, such as |
36 |
the SSD in this thread. I just don't want someone that is sort of new |
37 |
to Gentoo and compiling things to think that a package that takes 10 |
38 |
minutes when done on disk will take 3 minutes when on tmpfs. I see that |
39 |
thinking from time to time, usually on the forums. |
40 |
|
41 |
Dale |
42 |
|
43 |
:-) :-) |
44 |
|
45 |
-- |
46 |
I am only responsible for what I said ... Not for what you understood or how you interpreted my words! |