| 1 |
On 12/13/2014 10:10 AM, Diamond wrote: |
| 2 |
> On Sat, 13 Dec 2014 09:44:21 +0100 |
| 3 |
> Michał Górny <mgorny@g.o> wrote: |
| 4 |
> |
| 5 |
>> Dnia 2014-12-11, o godz. 11:36:29 |
| 6 |
>> Michał Górny <mgorny@g.o> napisał(a): |
| 7 |
>> |
| 8 |
>>> Hello, everyone. |
| 9 |
>>> |
| 10 |
>>> Following a similar change in distutils-r1, I would like to remove |
| 11 |
>>> the parallel run support in multilib eclasses, and effectively from |
| 12 |
>>> multibuild completely. |
| 13 |
>>> |
| 14 |
> |
| 15 |
> What is parallel run support? |
| 16 |
|
| 17 |
It uses multiprocessing.eclass, which parses the make -j setting from |
| 18 |
the MAKEOPTS variable, and uses that setting to internally parallelize |
| 19 |
the build (as make would if we were using make for that build). |
| 20 |
|
| 21 |
> Will it affect (the speed of) parallel emerging of packages? |
| 22 |
|
| 23 |
The effect will be to eliminate the internal parallelization of some |
| 24 |
builds that were internally parallelized via multiprocessing.eclass |
| 25 |
usage. This will reduce the load incurred by those builds, which reduces |
| 26 |
the global load average, allowing emerge to spawn more concurrent builds |
| 27 |
when using emerge --jobs with the --load-average option. |
| 28 |
-- |
| 29 |
Thanks, |
| 30 |
Zac |
Archives