It may well be that the dependency graph forces theories to be built
sequentially, in which case –j2 won’t win you anything. And then, even when
concurrent execution is possible, we are still at the mercy of the OS’s
scheduler.
Michael
On 26/10/17, 09:22, "Mario Castelán Castro" <marioxcc...@yandex.com> wrote:
On 25/10/17 17:16, michael.norr...@data61.csiro.au wrote:
> The –j number dictates how many processes Holmake will fork/exec at once
to run theory scripts. On a 1 core machine (without hyperthreading), you will
only get time-sliced concurrency between the scripts. As scripts do minimal I/O
(reading the script and writing theory files at the end), you may not get much
advantage from scripts able to run while others are blocked.
>
> With –j1, Holmake doesn’t monitor the child process asynchronously, but
uses “system” to run scripts one after the other.
Sorry. I made a typing error in my previous message I meant to write:
I used “bin/build -j 2” and it still used only *1* core
most of the time. My CPU is *2* core.”.
So it mostly wastes the opportunity to use both cores.
--
Do not eat animals; respect them as you respect people.
https://duckduckgo.com/?q=how+to+(become+OR+eat)+vegan
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
hol-info mailing list
hol-info@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/hol-info
