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