1 |
Dnia czwartek, 28 wrze¶nia 2006 09:24, Petter Haggholm napisa³: |
2 |
|
3 |
|
4 |
> Without having tried it yet -- I need to finish some work first -- I'm |
5 |
> curious about the MAKEOPTS thing, especially -j?. The other reply (so |
6 |
> far) in this thread suggests -j1 and asserts that "[-j2] has no sense |
7 |
> anyway, if you have no multicore or don't use distcc...". This is in |
8 |
> contrast to all the recommended settings I have seen so far -- standard |
9 |
> procedure seems to be -j$(no. cores+1), I presume so that one make |
10 |
> process can do some CPU work while the other is I/O bound? I suppose one |
11 |
> could keep the system more responsive at the cost of longer compiles by |
12 |
> running only one make process, but I have never heard of anyone else |
13 |
> having problems with -j2 on a single-core system ... |
14 |
|
15 |
Well, problem is not CPU but memory. Having -j2 + kdeenablefinal + gcc-4 eats |
16 |
ENORMOUS amounts of memory. This is the reason it kills some running |
17 |
applications (as in your case). Put some memory usage monitor and observe, |
18 |
especially the VM bar. In my system it goes up as if it had at least a foot |
19 |
of scale to collect, not mere half an inch... |
20 |
|
21 |
Remember that 'kdeenablefinal' works this way: it prepares application as one |
22 |
huge source file and then compiles it. Without kdeenablefinal each .cc file |
23 |
of app is compiled separately and then linked together, which significantly |
24 |
decreases memory usage - it is easier to compile 100 1MB files than 1 100MB |
25 |
file... |
26 |
|
27 |
-- |
28 |
Pawel Kraszewski |
29 |
www.kraszewscy.net |
30 |
|
31 |
-- |
32 |
gentoo-amd64@g.o mailing list |