1 |
Walter Dnes wrote: |
2 |
> On Sat, Oct 14, 2017 at 06:25:35AM -0500, Dale wrote |
3 |
> |
4 |
>> Some of my make.conf entries. You may not need all of these so edit out |
5 |
>> what you don't want or change values if you need to. I have a four core |
6 |
>> CPU. |
7 |
>> |
8 |
>> FEATURES="-usersync -userpriv -usersandbox buildpkg sandbox parallel-fetch" |
9 |
>> |
10 |
>> MAKEOPTS="-j5" |
11 |
> There is some controversy over setting MAKEOPTS=${number of threads} |
12 |
> possibly being better than MAKEOPTS=${number of threads} + 1 |
13 |
> https://blogs.gentoo.org/ago/2013/01/14/makeopts-jcore-1-is-not-the-best-optimization/ |
14 |
> |
15 |
|
16 |
|
17 |
I tried different settings before that one and settled on the one that |
18 |
works best. It may not work on someone else's system but it does fine |
19 |
here. One thing about most of my settings, they've been tested pretty |
20 |
well and work well on this system at least. |
21 |
|
22 |
Since I have a four core system, I usually set everything to number of |
23 |
cores plus one which should lead to a always busy CPU. So far, it has |
24 |
worked out that way for the most part. If I had enough memory, I might |
25 |
could up that but it could lead to another problem I had before lowering |
26 |
it to that number. Compiles that fail. It's been a while but at one |
27 |
point, some packages had to have -j 1 to compile. That is rare nowadays |
28 |
I guess but since the settings I have works for me, I'll likely leave |
29 |
them like they are. |
30 |
|
31 |
The one thing I did have to change recently, not compiling some large |
32 |
packages on tmpfs. If it was just one package I have enough memory. |
33 |
However, sometimes it would be two and sometimes even three that were |
34 |
large. The main culprits were Seamonkey, Firefox and Libreoffice. I |
35 |
think I added a couple other large ones in just to be sure but other |
36 |
than that, these settings have been around for a good long while. |
37 |
|
38 |
I might add, I've read that blog before. I also read the comments where |
39 |
others had different results and pointed out some issues with the blog |
40 |
points. The blogger used kdelibs as a test case. Thing is, I compile |
41 |
more than just kdelibs here. Most of the time I start my emerges before |
42 |
I go to sleep. I put Gkrellm where I can see it and I watch what the |
43 |
CPUs are doing. Generally speaking, the CPU cores stay busy the whole |
44 |
time. There are a few times that it is not at 100% but generally it |
45 |
is. If anything, since it does have times where it isn't at 100%, I may |
46 |
need to up that number by one to see if that keeps it more busy. Thing |
47 |
is, in the past, it didn't help any. |
48 |
|
49 |
Thanks for the info. |
50 |
|
51 |
Dale |
52 |
|
53 |
:-) :-) |