On Mon, Feb 24, 2020 at 10:44 PM David Kastrup <d...@gnu.org> wrote: > > The result is 25 minutes of purely CPU bound grinding (this is with > > Guile 2.2). Then building the remaining docs takes about 15 minutes. > > In this last phase, there is some inefficiency: we process the > > documents per language directory, but for each directory there is a > > bunch of small files, and a humongous notation manual. We can't move > > on to the next directory until the notation manual PDF file finishes. > > Where would the point be in moving on to the next directory when > CPU_COUNT processors are already working on the notation manual?
I didn't get the problem across clearly. Here is the 'de' directory: $ ls -rsS1 out-www/*.pdf 388 out-www/extending.pdf 544 out-www/usage.pdf 1320 out-www/essay.pdf 1832 out-www/learning.pdf 6584 out-www/notation.pdf The notation manual is so big that processing it takes much longer than the other files. This means that while processing the directory, there are times where make is running just the texi => pdf step for the notation manual (1 CPU), and nothing else, because there is nothing left to do in the directory. > > In order to fix this, we would have to reorganize the build system so > > it builds everything out of one directory. If we do all language directories in parallel, there would be other things to do while finalizing de/notation.tely -- Han-Wen Nienhuys - hanw...@gmail.com - http://www.xs4all.nl/~hanwen
