| 1 |
Duncan wrote: |
| 2 |
> So, does it sound like fun to anyone? For all I know someone's already |
| 3 |
> tested it. I haven't been following ext4 development close enough to |
| 4 |
> know for sure, if I didn't simply happen across it. |
| 5 |
> |
| 6 |
|
| 7 |
I'd certainly be curious as to the results of such a benchmark. Maybe |
| 8 |
I'll try to mess around and come up with some (good excuse to try out ext4). |
| 9 |
|
| 10 |
However, I agree with the crux of your argument. In theory tmpfs+swap |
| 11 |
should be able to outperform any filesystem out there. In practice it |
| 12 |
might not - kind of like seeing a really-well-written bubblesort |
| 13 |
outperform a really-badly-written quicksort. I still see tmpfs as being |
| 14 |
the better design, but if it is hampered by its implementation maybe |
| 15 |
there is something to doing it differently. |
Archives