1 |
On 9/29/05, Dave Nebinger <dnebinger@××××.com> wrote: |
2 |
> > Were I you, I would consider: |
3 |
> > |
4 |
> > - If keeping X, switching to the absolute most minimal wm possible |
5 |
> > (twm, ratpoison, ion), to see what effect that had. |
6 |
> > - If downstepping from X, investigating what programs run under |
7 |
> > DirectFB and seeing what effect that had. |
8 |
> > - If going cold-turkey off X, seeing how far you get with the |
9 |
> > command-line and ncurses programs. |
10 |
> |
11 |
> I would also add the following: remoting X. X is a hog, as Holly said, but |
12 |
> there's no reason the X server would need to run on the same box as the |
13 |
> ongoing recording session. |
14 |
> |
15 |
> Running two systems, one running X and handling the gui operations, and one |
16 |
> running your audio apps, might provide enough of the separation to reduce |
17 |
> the latency on the audio box. Of course the two cards should probably be |
18 |
> connected with at least a 100mb Ethernet connection (to eliminate the |
19 |
> overhead of dealing with the network conversations for X). |
20 |
|
21 |
This is an interesting idea actually. I currently run two boxes |
22 |
anyway. All audio is connected between them using ADAT optical (i.e. - |
23 |
red laser) or spdif so I've got 26 digital audio channels going |
24 |
across. Maybe running remotely could solve some of this. Thanks. |
25 |
|
26 |
> |
27 |
> Another possibility might be your choice of filesystems (assuming the |
28 |
> recordings are going to disk). Different filesystems have inherent latency |
29 |
> based upon their design - journaling adds overhead, btree maintenance in |
30 |
> reiser adds overhead, etc. Just using a simple ext2 filesystem for the |
31 |
> initial recording followed by backups to a modern filesystem may have a |
32 |
> measurable impact. |
33 |
|
34 |
In fact it does. I wrote a short online paper about that a few years |
35 |
ago. I use ext3. |
36 |
|
37 |
> |
38 |
> Going back to X, it is a hog both in cpu cycles and in memory; you mention |
39 |
> having an amd64 but no quotes on memory. My assumption is that such a |
40 |
> system has a big chunk of memory, but I've learned what happens when one |
41 |
> assumes. Obviously a lack of sufficient memory can cause you some swapping |
42 |
> issues whether you were aware of it or not. |
43 |
|
44 |
Thisis a good point also. The machine has 512MB. This has been more |
45 |
than enough on my previous 32-bit machines, but on this AMD64 running |
46 |
the Gentoo 64-bit kernel it seems that memory usage is significantly |
47 |
higher. On the 32-bit machine I seem to use about 300-350MB by the |
48 |
time I have Gnome up and maybe Firefox open. I alomost have never seen |
49 |
swapping. |
50 |
|
51 |
On the AMD64 I'm seeing 450-500MB and a small amount of swapping every day. |
52 |
|
53 |
I'm unclear about the 64-bit environment anyway. OK, it's 64-bit, but |
54 |
I also have a pile of emulation libraries emerged to take care of |
55 |
dependencies. I don't know when they are getting used, except for the |
56 |
32-bit Firefox I'm running so that I get more multimedia stuff. |
57 |
|
58 |
Anyway, more memory may well be a good thing to do. |
59 |
|
60 |
Thanks for the ideas. |
61 |
|
62 |
-- |
63 |
gentoo-user@g.o mailing list |