...co.uk> wrote: m...@iae.nl (Marcel Hendrix) wrote: Jonah Thomas <jethom...@gmail.com> writes Re: Duff's device [Re: More fun with Miser's CASE] Yes, and he had to use tricks that aren't portable. ... things and so potentially a lot of nonstandard code could be a lot more portable than it first appears. If you come up with an elegant ...

...m...@iae.nl (Marcel Hendrix) wrote: Jonah Thomas <jethom...@gmail.com> writes Re: Duff's device [Re: More fun with Miser's CASE] Gerry <ge...@jackson9000.fsnet.co.uk> wrote: Jonah Thomas <jethom...@gmail.com> ...: I'm not sure I understand the point of Duff's device. [..] so that we could actually put the address of the code into a...

mhx@iae.nl (Marcel Hendrix) wrote: Jonah Thomas <jethomas5@gmail.com> writes Re: Duff's device [Re: More fun with Miser's CASE] Gerry <gerry@jackson9000.fsnet.co.uk> wrote: I haven't had a chance to look at Marcel'..., very, very much slower than memory operations. I predict that Duff's device has no real use (for normal I/O) at all....

...@gmail.com> writes Re: Duff's device [Re: More fun with Miser's CASE] Gerry <gerry@jackson9000.fsnet.co.uk> wrote: Jonah Thomas <jethom...@gmail.com> ...'m not sure I understand the point of Duff's device. [..] so that we could actually put the address of the code into a ...much slower than memory operations. I predict that Duff's device has no real use (for normal I/...

... writes Re: Duff's device [Re: More fun with Miser's CASE] ... part of your statement. -marcel -- ------------------------------------------- ANEW -duff 0 VALUE ctab : | ( n -- ) BL WORD ...9 CELLS ALLOT POSTPONE THEN POSTPONE duff POSTPONE EXIT ; IMMEDIATE \ -- test it --...... $0055AAB4 IDIS $0055AAB4 jmp duff+8 ( $0055A508 ) offset NEAR $0055AAB9 jmp...

... for smaller embedded targets where Duff's Device still makes sense. Well, ...targets are so excellent that Duff's Device doesn't help. Usually, ...And the key idea behind Duff's Device can be applied in ... clear to me how Duff's Device would help here, given ... What's common about Duff's Device and the coroutine hack ... shows that a switch's cases may arbitrarily occur within nested ...

... sure I understand the point of Duff's device. There isn't one any more. In the bad old days C compilers ... hand. These days the use of Duff's device tends to worsen code quality rather ...C compilers for smaller embedded targets where Duff's Device still makes sense. Well, maybe. ... embedded targets are so excellent that Duff's Device doesn't help. And the key ...

...not sure I understand the point of Duff's device. There isn't one any more. In the bad old days C compilers ...hand. These days the use of Duff's device tends to worsen code quality rather than improve it. Depends on... compilers for smaller embedded targets where Duff's Device still makes sense. And the key idea behind Duff's Device can be applied in other ways,...

...<i...@quirkster.com> wrote: I honestly would like to know how one could portably implement Duff's device efficiently in Forth (i.e. multiple entry points into the code of a while loop). ... to solve in practice. To answer the question of "why?", here is Tom Duff describing his motivation. (It solved an inner I/O loop bottleneck in the 1984 ...

...I understand the point of Duff's device. I'd never heard of... asked was how could Duff's device be done in standard ... code I presented does. Duff's device uses a C trick ... to use nested IFs or CASE they'll take on average 4... FOR NEXT likely to be more efficient than BEGIN WHILE REPEAT ... FOR NEXT for the mod case (or if you factor the ... you could re-create Duff's device in standard Forth. You ...