On Sun, Jun 19, 2016 at 11:52 AM, Charles Mills <[email protected]> wrote:
> Oh sure. The basic "wait for one of many." My code waits for an any one of > an operator command, a timer expiration, or "real work." > > My question was does anyone issue a wait with a completion count of more > than one? "Wake me when two of these five ECBs have been posted"? > > Charles > > That is a good question. I can think of a reason, as you said, to wait for "1 of n". And I can think of a reason to wait for "n of n". I've even thought of a reason to wait for what I'll call "1+1 of n" (i.e. what until one specific ECB is posted, plus one of "n" other ECBs). The closest that I can envision for "2 of n" might be if the "n" are "n" worker threads which produce "some output" and, for some reason that I can't envision, the parent needs at least 2 worker products in order to do something else. But that idea is so vague that it is basically useless. Hum, maybe a case of "efficiency" in the aforementioned scenario. Suppose that you need "n" worker threads to produce "product". These thread each take a relatively "large" amount of time. Your main thread does something with this "product", but it does it quickly. It would be more efficient to accumulate "m" (<n) units of "product" to process in a loop for each redispatch of the main thread. Well, a little more reasonable, but still not too specific. -- "Pessimism is a admirable quality in an engineer. Pessimistic people check their work three times, because they're sure that something won't be right. Optimistic people check once, trust in Solis-de to keep the ship safe, then blow everyone up." "I think you're mistaking the word optimistic for inept." "They've got a similar ring to my ear." >From "Star Nomad" by Lindsay Buroker: Maranatha! <>< John McKown ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
