1 |
On Thu, Jan 27, 2011 at 08:18:34PM +0100, J. Roeleveld wrote: |
2 |
> On Thursday 27 January 2011 19:56:23 Allan Gottlieb wrote: |
3 |
> > On Thu, Jan 27 2011, Neil Bothwick wrote: |
4 |
> > > On Thu, 27 Jan 2011 17:09:27 +0200, Nikos Chantziaras wrote: |
5 |
> > >> > So on a 20 package world update, only 19 are faster while the 20th |
6 |
> > >> > runs at the same speed? Where's the loss there? Even if the last were |
7 |
> > >> > slower, it would be worth it. |
8 |
> > >> |
9 |
> > >> Given the amount of time unpack/configure/install of most packages |
10 |
> > >> needs (very short), my observation is that it would not be worth it. |
11 |
> > > |
12 |
> > > Even if that were true, how much time would you have to save to justify |
13 |
> > > adding -j 2 to EMERGE_DEFAULTS in make.conf? |
14 |
> > > |
15 |
> > > But it's not true, large packages spend a lot of time on these phases of |
16 |
> > > the install. |
17 |
> > |
18 |
> > OK I'm convinced since I know that those phases do take noticeable time. |
19 |
> > |
20 |
> > I have a "4" processor i7 model 620 (2 cores, doubled for |
21 |
> > hyperthreading) and have set MAKEOPTS="-j5". |
22 |
> > If I add -jobs=2 to EMERGE_DEFAULT_OPTS, should I lower |
23 |
> > MAKEOPTS to 3 (to 4)? |
24 |
> > |
25 |
> > thanks, |
26 |
> > allan |
27 |
> |
28 |
> You could, as if you leave it at -j5, you can end up with 2 * 5 = 10 |
29 |
> processed, eg: similar as if running with MAKEOPTS="-j10" |
30 |
> |
31 |
> I think the option that YoYo came with is a good compromise: |
32 |
> # MAKEOPTS="-j -l5" emerge -j --load-average=5 |
33 |
> |
34 |
> Next time I am doing a big upgrade, I'm going to test that to see how it |
35 |
> behaves. |
36 |
|
37 |
I was just building the whole system for my notebook in a chroot on my |
38 |
desktop machine ( I use FEATURES=buildpkg to build binary packages in |
39 |
the chroot on a fast desktop machine and then upgrade the notebook with |
40 |
the binary packages) and I used exactly that (-j -l5 for bot make and |
41 |
emerge). Can't say if it really is better or not ;) but most of the time |
42 |
all four cores were busy, though sometimes I saw even 6 or 7 gcc-s |
43 |
simultaneously in top ;) |
44 |
emerge was running 3 to 4 jobs most of the time, sometimes dropping to 1 |
45 |
and once I saw it emerging about 10 parallel packages ;) (mostly small |
46 |
things, which I guess were doing a lot of |
47 |
unpacking/configuring/installing but almost nothing of compiling ;) |
48 |
|
49 |
|
50 |
btw, just now I got this error from dev-lang/v8: |
51 |
|
52 |
SCons error: option -j: invalid integer value: '-l4' |
53 |
|
54 |
seems scons honors MAKEOPTS, but doesn't understand the "loadaverage" |
55 |
version (-j -l4) |
56 |
|
57 |
|
58 |
yoyo |