1 |
On 7 January 2015 13:16:03 GMT+00:00, Rich Freeman <rich0@g.o> wrote: |
2 |
> On Wed, Jan 7, 2015 at 3:36 AM, Neil Bothwick <neil@××××××××××.uk> |
3 |
> wrote: |
4 |
> > |
5 |
> > Personally, I wouldn't bother, there is not that much of a gain when |
6 |
> > using tmpfs, so if you want to keep all the working files after |
7 |
> > compilation, the extra overhead and complexity of copying to hard |
8 |
> disk |
9 |
> > would make it not worthwhile. Just stick to a spinning disk. |
10 |
> |
11 |
> I've found that compiling on tmpfs has a fairly significant |
12 |
> performance gain, but you'd completely negate it of you copied |
13 |
> everything to a hard drive anyway. |
14 |
> |
15 |
> When you build on a hard drive the filesystem is going to treat all |
16 |
> those intermediate object files with great care and ensure that they |
17 |
> aren't lost in the event of a power failure, forcing them to be |
18 |
> committed to disk within 30 seconds (typically) and blocking IO when |
19 |
> that happens. Then a minute later it is going to go and have to |
20 |
> delete all those files it so carefully committed. |
21 |
> |
22 |
> When you build on tmpfs nothing gets written to your filesystem except |
23 |
> for the final output of the build. All those intermediate files are |
24 |
> deleted having never blocked your disk IO. However, if your goal is |
25 |
> to actually save them anyway, then you might as well just build on |
26 |
> disk - moving files doesn't take all that much IO on most filesystems. |
27 |
> |
28 |
> If your goal is to create tarballs of the saved builds then you're |
29 |
> probably still better off building on tmpfs and creating the tarball |
30 |
> via a hook or something. |
31 |
> |
32 |
> -- |
33 |
> Rich |
34 |
|
35 |
That's why I thought XFS may help. |
36 |
|
37 |
Reports of the speed gain from tmpfs are quite mixed, but I do use it myself. |
38 |
-- |
39 |
Sent from my Android phone with K-9 Mail. Please excuse my brevity. |