I agree that startup on a warm cache is mostly a kernel / filesystem ram cache benchmark. The actual drive hardware is irrelevant, but background processes hitting the disk cache at the same time do matter. On a warm cache and without background IO the sage -startuptime (i.e. importing sage.all) is pretty consistently about 800ms for me. Which I don't think can be much improved without changing the way how Python imports stuff; Surely we can still shave off a few tens of ms but thats it.
For server use you should be forking processes from a zygote process, this is how chrome works. The only way to fix the import speed on a cold cache with high-latency drive hardware is to load a single big file (aka database) instead of thousands of small files. And cache the results of stat calls at build time, and then use those via a custom importer. But thats a lot of work for a pretty narrow use case. Its also fragile as the database can easily go out of sync with the actual filesystem, e.g. when you pip install stuff. > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
