Question for Matthew: the docs for raco setup [1] mention something interesting:
-jobs ‹n›, --workers ‹n›, or -j ‹n› — > use up to ‹n› parallel processes. By default, raco setup uses ( > processor-count > <https://docs.racket-lang.org/reference/futures.html#%28def._%28%28lib._racket%2Ffuture..rkt%29._processor-count%29%29> > ) jobs, which typically uses all of the machine’s processing cores. > Indeed when raco setup is rebuilding all the scribble docs, it maxes out my CPU. By contrast, when rendering a large batch of Pollen documents, my CPU util% is generally between 10–20%. Would it be feasible to add some optional parallelism to Pollen, either as an argument to raco pollen render or as an alternate form of render-batch? Authors would have to take care not to shoot themselves in the feet with it, of course, but the same is true of programming generally. I can see where there may be something in Pollen's design that prevents any Pollen project from being amenable to parallel processing, but I don't know enough about the subject to figure out what that would be. If you think it could work but you'd rather not add it or have to support it right now (please don’t do anything that would noticeably interrupt your work on Quad!), that would be useful information as well; it would mean it might be worth my while to try rolling my own experimental render-batch-parallel. [1]: https://docs.racket-lang.org/raco/running.html -- You received this message because you are subscribed to the Google Groups "Pollen" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
