| 1 |
On Saturday 19 December 2009 12:19:05 Stefan G. Weichinger wrote: |
| 2 |
> Am 19.12.2009 09:08, schrieb Stroller: |
| 3 |
> >> Could anyone comment? |
| 4 |
> > |
| 5 |
> > Might it help if you said WHY you're unimpressed? |
| 6 |
> |
| 7 |
> might be ;-) |
| 8 |
> |
| 9 |
> It seems as if my system doesn't benefit that much: |
| 10 |
> |
| 11 |
> With 8 gigs of RAM, suspend-to-ram, preload and only a handful of rather |
| 12 |
> lightweight binaries in regular use this system is already pretty fast |
| 13 |
> with hdds only. |
| 14 |
> |
| 15 |
> I expected more WOW in terms of overall speed ... |
| 16 |
|
| 17 |
SSDs are not a magic bullet, it's unlikely they will give you a killer |
| 18 |
performance improvement that makes you go "WOW!!!" |
| 19 |
|
| 20 |
SSDs suck at random writes. |
| 21 |
Typical usage scenario on a workstation is lots of random writes compared to |
| 22 |
relatively few random reads - reads tend not to be all that random as you |
| 23 |
re-read the same thing often and it gets cached. |
| 24 |
|
| 25 |
Intel SSDs are far superior at random writes than any other SSD out there but |
| 26 |
it's still nowhere near as optimised as spinning drives, and kernels by and |
| 27 |
large are still optimised for spinning drives too. |
| 28 |
|
| 29 |
This may account for your overall feeling of under-whelmedness why still |
| 30 |
seeing a significant boot-time speed up. You also have enough RAM so that |
| 31 |
almost an entire typical workstation session could fit in RAM and seldom touch |
| 32 |
the disk especially with a large interval between disk syncs |
| 33 |
|
| 34 |
-- |
| 35 |
alan dot mckinnon at gmail dot com |
Archives