1 |
On Monday, August 18, 2014 08:09:00 PM thegeezer wrote: |
2 |
> On 18/08/14 15:31, J. Roeleveld wrote: |
3 |
> > <snip> |
4 |
> |
5 |
> valid points, and interesting to see the corrections of my |
6 |
> understanding, always welcome :) |
7 |
|
8 |
You're welcome :) |
9 |
|
10 |
> > Looks nice, but is not going to help with performance if the application |
11 |
> > is |
12 |
> > not designed for distributed processing. |
13 |
> > |
14 |
> > -- |
15 |
> > Joost |
16 |
> |
17 |
> this is the key point i would raise about clusters really -- it would be |
18 |
> nice to not need for example distcc configured and just have portage run |
19 |
> across all connected nodes without any further work, or to use a tablet |
20 |
> computer which is "borrowing" cycles from a GFX card across the network |
21 |
> without having to configure nvidia grid: specifically these two use |
22 |
> cases have wildly different characteristics and are a great example of |
23 |
> why clustering has to be designed first to fit the application and |
24 |
> viceversa. |
25 |
|
26 |
I had a better look at that site you linked to. It won't be as "hidden" as |
27 |
you'd like. The software you run on it needs to be designed to actually use |
28 |
the infrastructure. |
29 |
This means that for your ideal to work, the "industry" needs to decide on a |
30 |
single clustering technology for this. I wish you good luck on that venture. |
31 |
:) |
32 |
|
33 |
> /me continues to wonder if 10GigE is fast enough to page fault across |
34 |
> the network ... ;) |
35 |
|
36 |
Depends on how fast you want the environment to be. |
37 |
Old i386 time, probably. |
38 |
Expecting a performance equivalent to a modern system, no. |
39 |
|
40 |
Check the bus-speeds between the CPU and memory that is being employed these |
41 |
days. That is the minimum speed you need in the network link to be fast enough |
42 |
to actually work. And that is expecting a perfect link with no errors |
43 |
occurring in the wiring. |
44 |
|
45 |
-- |
46 |
Joost |