On Tue, 9 Mar 2010 07:23:03 -0500 Peter Relson <[email protected]> wrote:
:>The question really cannot be answered as posed, because it depends on the :>contention rate. :>Looking only at path-length fastest to slowest: :>SETLOCK :>Latch :>ENQ :>But that is in the absence of contention. Once the requestor has to be :>suspended because of contention, you've lost. :>The Latch and ENQ approaches, being more granular, make it possible to :>have less contention as "no one else" can also be contending (unlike the :>local lock where any getmain or post or myriad other things being done in :>that space could also need the local lock). Latches have the benefit of :>not needing any other system serialization themselves to grant (again, in :>the absence of contention). :>I would be fairly certain that a complicated PLO operation (it takes a :>good deal of setup to issue even a simple PLO) with more-complex logic :>would lose out to SETLOCK and possibly even latch obtain when there is no :>contention. It will certainly make your code far harder to maintain (let :>alone to certify). PLO would likely be a winner if there is contention. Wouldn't ENQ implicitly do a SETLOCK as part of its process? -- Binyamin Dissen <[email protected]> http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
