Henry, What's the upper bound on that pre-allocated list, that's a good technique actually. But what do I know? My main issue was more the fact I couldn't see another way to do it but both Raul and Clifford have shown me the way.
Seasons greeting to all, I'd express that in J but... On Thu, 24 Dec 2020 at 13:01, Henry Rich <[email protected]> wrote: > i. 100 returns a section of a pre-allocated list, so it doesn't even > have to generate anything. > > (? 100 $ y) is a smidgen slower than (100 ?@$ y) > > Henry Rich > > On 12/24/2020 7:37 AM, Raul Miller wrote: > > Generating that list of numbers is trivial, compared to typical > > language processing. > > > > Consider a classic "for" loop, which calls a function 100 times. Here, > > you are also generating 100 numbers. It's true that they are not > > stored in memory simultaneously, but you're also generating 100 stack > > frames, one after another -- and the cost of constructing those is > > significant. Writing to one memory location is less work than > > populating a typical stack frame. > > > > Also, in terms of actual memory consumed -- unless you're up in the > > megabytes, you aren't going to even notice it on modern machines. > > > > That said, if your concern is the *relevance* of the numbers, I'd go > > with something like: > > > > 1,.~? 100 $ ,: 2 1024 768 > > > > Which, looking at what Cliff Reiter suggested, is basically the same > concept. > > > > Good luck, > > > > > -- > This email has been checked for viruses by AVG. > https://www.avg.com > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
