|
Simon, Thanks for the detailed explanation. In my case I still use the sorted data queue, but it only contains an offset and length for a user space entry. I think the main improvement comes from a user space being able to store variable length entries more efficiently. I don't know where you find this stuff, so I can't go look myself, but if you know off hand, I would be interested in whether the the data queue sorted on every dequeue or only on a dequeue with when after an entry has arrived? Also, do you know whether the whole entry is sorted rather than just the key? Maybe that is why I could not recognize a difference between sorted data queues and a user index. David Morris >>> shc@flybynight.com.au 01/22/02 04:37PM >>> Hello David, You wrote: >I always believed the manuals that said data queues were the fastest >means of interprocess communication and were extremely lightweight. They are fast -- faster than the alternatives of message queue or file -- and user queues are faster still. The major problem with data queues occurs when they are keyed AND you fill them with data before starting the process that fetches the data (may also occur when entries arrive on the queue faster than the queue reader can process them). The reason is simple. On RISC AS/400 a keyed queue is not sorted when entries are added (as happened on IMPI) but rather sorted when a dequeue operation is performed. The sort is not very efficient (On**2 last time I looked) and I don't believe it has changed. So changing from keyed to sequential as you did would certainly make an improvement but I would expect a major benefit only if you were filling the queue before processing the entries. Regards, Simon Coulter. -------------------------------------------------------------------- FlyByNight Software AS/400 Technical Specialists http://www.flybynight.com.au/ Phone: +61 3 9419 0175 Mobile: +61 0411 091 400 /"\ Fax: +61 3 9419 0175 mailto: shc@flybynight.com.au \ / X ASCII Ribbon campaign against HTML E-Mail / \ --------------------------------------------------------------------
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.