nd@doc.ic.ac.uk (N Dulay) (10/04/90)
I would interested to learn whether anyone has used ACI's new 4th Dimension Compiler on their existing 4th Dimension databases. And if so, how much improvement they have noticed? N Dulay Dept of Computing Imperial College 180 Queen's Gate London SW7 2BZ Tel: +44 71 589 5111 ext 5091
paulr@syma.sussex.ac.uk (Paul T Russell) (10/05/90)
From article <2299@gould.doc.ic.ac.uk>, by nd@doc.ic.ac.uk (N Dulay): > I would interested to learn whether anyone has used ACI's new 4th Dimension > Compiler on their existing 4th Dimension databases. And if so, how much > improvement they have noticed? The compiler only really speeds up scripts and procedures. If your database is mostly automatic (ie. using 4D's default behaviour) then you won't see much improvement. If you spend a lot of time interpreting long and/or complex procedures then you will get a considerable speed increase (an order of magnitude or thereabouts). To squeeze the maximum performance out of the compiled version you need to declare all your string and integer variables (otherwise they will default to the (slower) text and real types). This applies to procedure parameters and function results too. It took me about a day to get one of my systems 'compiler-ready' and then a little more time to debug a few minor run-time problems. The best speed improvement I have seen is x 50 on a particularly long-winded procedure... //Paul -- Paul Russell, Department of Experimental Psychology University of Sussex, Falmer, Brighton BN1 9QG, England Janet: paulr@uk.ac.sussex.syma Nsfnet: paulr@syma.sussex.ac.uk Bitnet: paulr%sussex.syma@ukacrl.bitnet Usenet: ...ukc!syma!paulr
bmug@garnet.berkeley.edu (BMUG) (10/05/90)
In article <2299@gould.doc.ic.ac.uk> nd@doc.ic.ac.uk (N Dulay) writes: >I would interested to learn whether anyone has used ACI's new 4th Dimension >Compiler on their existing 4th Dimension databases. And if so, how much >improvement they have noticed? > I've spent some time with the compiler, and in general it's a good (though expensive) complement to 4th Dimension. Depending on what your particular database application does, it can make a minor or major difference in speed. For apps which are almost all disk I/O intensive, little or no difference. For those which use a lot of calculations, iterative structures, looping, or other interpreter- heavy operations, the increase in speed can be amazing. If you have access to MacWEEK, check out the October 1st issue, which has a review of the compiler and fairly typical, "real world" tests. A less obvious use for the compiler is its debugger, which will catch things the source-level debugger in 4D's interpreter doesn't. The interpreter can live with a certain amount of programming ambiguity; the compiler (because it's reducing the interpreted code to machine code) can't. Using the compiler can train you to be a more careful programmer by catching sloppy techniques. John Heckendorn /\ BMUG ARPA: bmug@garnet.berkeley.EDU A__A 1442A Walnut St., #62 BITNET: bmug@ucbgarne |()| Berkeley, CA 94709 Phone: (415) 549-2684 | |
francis@giza.cis.ohio-state.edu (RD Francis) (10/05/90)
In article <1990Oct5.150511.14473@agate.berkeley.edu> bmug@garnet.berkeley.edu (BMUG) writes: >In article <2299@gould.doc.ic.ac.uk> nd@doc.ic.ac.uk (N Dulay) writes: >>I would interested to learn whether anyone has used ACI's new 4th Dimension >>Compiler on their existing 4th Dimension databases. And if so, how much >>improvement they have noticed? As long as we're talking about the compiler, have they ever gotten the version 2.0 xref utility done? As I recall, it was farmed out to After Hours, but I know that a number of people, some of whom work at Acius, expressed a concern that without a cross-referencing tool (like what *came* *with* version 1.0), writing a program for compilation would become much more difficult, and it was likely that the programs compiled without reference to the cross reference dump would be much less efficient. In particular, large programs rewritten from verison 1.0 might be likely to have numerous "temporary" global variables (I used to make all variables fair game for being cleared (i.e., removed from memory) unless they began G_; in particular, v* variables were wiped out at the end of every menu procedure). -- R David Francis francis@cis.ohio-state.edu
rad@genco.uucp (Bob Daniel) (10/06/90)
In article <2299@gould.doc.ic.ac.uk> nd@doc.ic.ac.uk (N Dulay) writes: >I would interested to learn whether anyone has used ACI's new 4th Dimension >Compiler on their existing 4th Dimension databases. And if so, how much >improvement they have noticed? The speed improvment is simply astronomical! Even on a Plus, 4D is actually tolerable. A loop of 65000 took over 10 minutes to run noncompiled and took less than 3 seconds compiled! Get it! It is worth $700 (US developers cost)
bmug@garnet.berkeley.edu (BMUG) (10/06/90)
In article <84448@tut.cis.ohio-state.edu> francis@giza.cis.ohio-state.edu (RD Francis) writes: > >As long as we're talking about the compiler, have they ever gotten the >version 2.0 xref utility done? The xref utility is still being worked on by After Hours Software (who produced 4D Mover); I suspect we'll see it around the same time as the 4D Environment modules (4DWrite, etc.) appear. I'm betting by November 15th... BTW, word is that Xref 2.1 will be several orders of magnitude more powerful than the wimpy one that was shipped with 4D 1.0 (and will cost more, as well, but you get what you pay for). John Heckendorn /\ BMUG ARPA: bmug@garnet.berkeley.EDU A__A 1442A Walnut St., #62 BITNET: bmug@ucbgarne |()| Berkeley, CA 94709 Phone: (415) 549-2684 | |