Re: [gentoo-user] Emerge and CPU core usage

[Dale]

james wrote:
> On 12/9/19 1:31 AM, Dale wrote:
>> Howdy,
>>
>
>> I'm sure trying to get portage to do things in parallel would be a
>> programmers nightmare.� It may not even be doable given how the
>> tree is
>> done or that the complexity of calculating all the options is just to
>> much to run in parallel. 
>
> Hello Dale,
>
> Not sure this is what you are looking for, but it's pretty easy to set
> up.
>
>
> in my /etc/make.conf file I have this::
>
>
> MAKEOPTS="-j7 -l8"
>
>
> so it does what you are looking for on larger upgrades. Some files
> that do not compile properly, auto limit to one core. My understanding
> is there are a variety of mechanism to achieve this.
>
> The kernel version/setting surely has more options, but they are
> optimized according to the types of workloads and the scheduler you
> have selected. YOU or anyone can waist weeks and months going down
> that pathway.
>
> Then there are mechanisms to further refine how your system works.
>
>
> It's a wide open area so read up a bit and find your comfort level.
>
>
> hth,
> James
>
>


I have that type of setting already.  What I was talking about is when
you do a emerge -uaDN world and it is calculating what packages needs
updating.  When it is doing that, it only uses one core, thread I guess
for those who have threads, which means having a multi-core CPU doesn't
help speed things up. Basically, my question was about the emerge
command itself not when it is actually compiling packages. 

I read where it was discussed on -dev a couple years or so ago.  I'm not
a coder but from what I could understand, it sounded really complicated
to have the emerge command able to run calculations in parallel.  It
seems there is a couple things that just can not be done that way. 

Maybe one day.  Just maybe. 

Dale

:-)  :-)