Groups | Search | Server Info | Login | Register
Groups > comp.lang.misc > #11279
| From | Andy Walker <anw@cuboid.co.uk> |
|---|---|
| Newsgroups | comp.lang.misc |
| Subject | Re: Algol 68 / Genie - new release |
| Date | 2025-09-30 23:23 +0100 |
| Organization | Not very much |
| Message-ID | <10bhl97$3hoao$2@dont-email.me> (permalink) |
| References | (18 earlier) <10b78bq$1cen9$1@dont-email.me> <10b8pc1$1jb4p$1@dont-email.me> <10bcj33$2mnrh$1@dont-email.me> <10be779$2scen$1@dont-email.me> <10behvb$365mk$1@dont-email.me> |
On 29/09/2025 19:08, bart wrote:
>> Loop/recursion unrolling can do that sort of thing. Space for an
>> extra 225 m/c instructions is neither here nor there if it saves a
>> few jumps or frames, esp with caches measured in megabytes.
> You're right, if the program was just this 7-line function plus one
> more line to call it.
> But why go to those lengths for this one tiny function out of all
> the thousands across an application?
If you have "thousands" of functions in one application, then
you're talking about something more substantial than anything I've
ever written. I suspect that such monsters are essentially not
maintainable and not understandable by humans, which is why new bugs
are always being discovered, usually only when something catastrophic
happens. I'm all for keeping things simple! There's a lot to be
said for the original "software tools" approach of Unix; pity it
doesn't entirely scale to major projects. [OTOH, it's surprising
what you can do with shell scripts, edit scripts and small programs
linked by pipes.]
> For a C-style HLL, then on x64 you can reckon on about 10 bytes of
> machine per line of source. As it happens, the 7-line fib() in my
> language generates 72 bytes of machine code.
> But in optimised C, the same function (where it is 6 lines),
> generates 950 bytes of machine code. That's about 150 bytes per line
> of source code, which is unacceptable and will result in bloated
> executables that are more likely to slow things down.
I two-thirds agree. But the way storage and cache are going,
the slow-down soon won't be noticeable [if it is even now].
[Unoptimised A68G vs "-O3":]> [...] I think I got an 8.5x speed-up (15s vs. 1.75s
> measuring elapsed time for 100 loops).
Perhaps worth noting that the Whetstone test in A68G is in
three parts; using standard reals, using long reals [for the same
code otherwise] and using long long reals [ditto]. These three parts
optimise differently -- on my machine there is a 9x speed-up for
reals, 15% for long reals and 2% for long long reals [as measured
by the "MWhets" figures]. If you have an old copy of the benchmark
you may have only the single-length real results, and of course if
you have different word lengths the results are incomparable anyway.
[If you have recent source, then the Whetstone source is in
".../src/test-set/34-whetstones.a68" where "..." is where you put
the installation, and "34" may vary with version.]
> However the behaviour with -O3 was odd. It seems -O3 causes it to
> spend 3 seconds optimising first before starting the program proper.
The optimiser is "gcc"'s, so nothing directly to do with A68G.[But at least it makes sense to optimise /before/ starting the program
rather than after! :-)]
> So for some short runtimes, it could well make it slower!
If you're running the program once, that may well be the case.
But of course the intention is that you should optimise once and run
the executable lots of times [if the run-time matters to you].
--
Andy Walker, Nottingham.
Andy's music pages: www.cuboid.me.uk/andy/Music
Composer of the day: www.cuboid.me.uk/andy/Music/Composers/Kirnberger
Back to comp.lang.misc | Previous | Next — Previous in thread | Next in thread | Find similar
Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-23 06:49 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-08-23 15:22 +0100
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-23 17:49 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-08-24 00:58 +0100
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-25 01:49 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-08-26 23:53 +0100
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-23 19:28 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-08-24 00:27 +0100
Re: Algol 68 - formatting with fixed point format Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-08-24 02:49 +0000
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-25 01:57 +0200
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-25 02:14 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-08-27 00:58 +0100
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-27 08:23 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-08-27 23:23 +0100
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-28 07:35 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-08-29 00:43 +0100
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-30 05:40 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-08-30 20:38 +0100
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-31 10:10 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-09-01 17:20 +0100
Algol 68 / Genie - new release (was Re: Algol 68 - formatting with fixed point format) Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-02 00:56 +0200
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-02 01:15 +0200
Re: Algol 68 / Genie - new release (was Re: Algol 68 - formatting with fixed point format) Andy Walker <anw@cuboid.co.uk> - 2025-09-02 17:02 +0100
Re: Algol 68 / Genie - new release (was Re: Algol 68 - formatting with fixed point format) Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-02 22:17 +0200
Re: Algol 68 / Genie - new release (was Re: Algol 68 - formatting with fixed point format) Andy Walker <anw@cuboid.co.uk> - 2025-09-04 17:01 +0100
Re: Algol 68 / Genie - new release (was Re: Algol 68 - formatting with fixed point format) Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-05 04:32 +0200
Re: Algol 68 / Genie - new release (was Re: Algol 68 - formatting with fixed point format) Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-09-02 23:04 +0000
Re: Algol 68 / Genie - new release (was Re: Algol 68 - formatting with fixed point format) Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-03 01:46 +0200
Re: Algol 68 / Genie - new release (was Re: Algol 68 - formatting with fixed point format) Andy Walker <anw@cuboid.co.uk> - 2025-09-04 00:30 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-10 17:15 +0200
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-09-10 23:57 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-11 07:27 +0200
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-11 09:42 +0200
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-09-11 17:16 +0100
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-09-27 00:41 +0100
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-09-27 14:38 +0100
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-09-29 01:15 +0100
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-09-29 16:05 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-29 18:57 +0200
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-09-30 15:58 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-09-30 17:48 +0200
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-09-29 19:08 +0100
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-09-30 23:23 +0100
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-01 23:00 +0100
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-10-02 16:55 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-02 21:51 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-04 12:38 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-04 14:04 +0200
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-10-04 13:47 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-04 15:03 +0200
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-04 15:07 +0100
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-10-04 18:40 +0100
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-04 19:46 +0100
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-10-07 16:37 +0100
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-07 17:16 +0100
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-10-07 23:58 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-08 01:20 +0200
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-10-08 16:10 +0100
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-08 17:41 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-09 00:53 +0200
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-09 01:02 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-08 01:09 +0200
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-07 22:27 +0100
Re: Algol 68 / Genie - new release richard@cogsci.ed.ac.uk (Richard Tobin) - 2025-10-08 00:20 +0000
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-04 22:00 +0000
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-05 00:31 +0200
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-05 01:52 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-05 03:15 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-05 11:05 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-05 20:20 +0000
Re: Algol 68 / Genie - new release David Brown <david.brown@hesbynett.no> - 2025-10-06 09:44 +0200
Higher order functions (was Re: Algol 68 / Genie - new release) Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-06 10:01 +0200
Re: Higher order functions (was Re: Algol 68 / Genie - new release) David Brown <david.brown@hesbynett.no> - 2025-10-06 10:56 +0200
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-06 11:03 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-06 22:01 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-06 23:46 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-06 23:05 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-07 10:32 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-08 04:57 +0000
Re: Algol 68 / Genie - new release David Brown <david.brown@hesbynett.no> - 2025-10-05 13:24 +0200
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-08 20:21 +0000
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-08 21:30 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-08 22:40 +0100
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-10 04:18 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-10 16:21 +0100
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-11 00:05 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-11 11:25 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-11 13:08 +0200
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-11 15:25 +0100
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-11 12:39 +0000
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-09 00:34 +0200
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-09 03:15 +0000
Re: Algol 68 / Genie - new release ram@zedat.fu-berlin.de (Stefan Ram) - 2025-10-09 08:59 +0000
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-09 14:31 +0000
Re: Algol 68 / Genie - new release ram@zedat.fu-berlin.de (Stefan Ram) - 2025-10-09 14:53 +0000
Re: Algol 68 / Genie - new release ram@zedat.fu-berlin.de (Stefan Ram) - 2025-10-09 10:35 +0000
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-09 21:37 +0000
Re: Algol 68 / Genie - new release David Brown <david.brown@hesbynett.no> - 2025-10-10 17:40 +0200
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-10 20:04 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-10 21:45 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-11 00:29 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-11 00:16 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-11 01:59 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-11 01:42 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-11 10:42 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-31 06:51 +0000
Re: Algol 68 / Genie - new release David Brown <david.brown@hesbynett.no> - 2025-10-13 20:58 +0200
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-13 23:13 +0100
Re: Algol 68 / Genie - new release David Brown <david.brown@hesbynett.no> - 2025-10-14 00:39 +0200
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-12 03:56 +0000
Re: Algol 68 / Genie - new release David Brown <david.brown@hesbynett.no> - 2025-10-13 22:38 +0200
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-24 04:33 +0000
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-03 22:24 +0000
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-04 03:39 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-04 12:19 +0100
Re: Algol 68 / Genie - new release Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-10-04 14:16 +0200
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-04 21:05 +0000
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-05 01:40 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-05 03:13 +0000
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-05 01:29 +0000
Re: Algol 68 / Genie - new release Andy Walker <anw@cuboid.co.uk> - 2025-10-07 15:25 +0100
Re: Algol 68 / Genie - new release bart <bc@freeuk.com> - 2025-10-07 17:03 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-07 22:09 +0000
Re: Algol 68 / Genie - new release Adam Sampson <ats@offog.org> - 2025-10-08 18:22 +0100
Re: Algol 68 / Genie - new release Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-10-08 21:48 +0000
Re: Algol 68 / Genie - new release antispam@fricas.org (Waldek Hebisch) - 2025-10-08 23:45 +0000
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-30 06:54 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-08-31 00:44 +0100
Re: Algol 68 - formatting with fixed point format Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-08-31 08:12 +0000
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-31 10:42 +0200
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-09-01 00:10 +0100
Re: Algol 68 - formatting with fixed point format Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-09-01 07:30 +0000
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-09-01 16:13 +0100
Re: Algol 68 - formatting with fixed point format Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-09-01 22:41 +0000
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-09-02 14:23 +0100
Re: Algol 68 - formatting with fixed point format Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-09-02 23:08 +0000
Re: Algol 68 - formatting with fixed point format Andy Walker <anw@cuboid.co.uk> - 2025-09-04 00:54 +0100
Re: Algol 68 - formatting with fixed point format Janis Papanagnou <janis_papanagnou+ng@hotmail.com> - 2025-08-31 10:29 +0200
Re: Algol 68 - formatting with fixed point format Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-08-24 02:47 +0000
csiph-web