[Oberon] Numerical CASE Statements in Project Oberon

[Jan de Kruyf]

Here is some news:

We get cleanly compiling code.

With this test module:
-------
MODULE TestIf;
VAR t : INTEGER;

BEGIN
   t := 10;
   IF ((t >= 9) & (t <= 11)) OR (t = 7) THEN
      t := 11;
   ELSIF (t = 13) OR ((t >= 14) & (t <= 16)) THEN
      t := 12
   END;

   CASE t OF
     9 .. 11, 7  : t := 11;
   |13, 14 .. 16 : t := 12
   END;

END TestIf.
-----------

When I only compile the IF structure I get 37 assembly instructions,
average seek for named values is 13.66 instruction times.
When I only compile the CASE structure I get 40 assembly instructions,
average seek for named values is 7.33 .
(this is from doing a hand count through the search structure)

A lot of waste in the IF is generated by reloading the 't' variable for
every compare, while I steal a register to keep 't' loaded.

j.


On Tue, Nov 24, 2015 at 8:38 PM, Jan de Kruyf <jan.de.kruyf at gmail.com>
wrote:

> John, thats a complex matter really.
>
> Anything we dont know at compiletime is left dangling for resolving
> 'later'.
> Part of that 'later' is resolved by the linker, in Linux or Windows. And
> still more is resolved when the code gets loaded.
>
> In Oberon, in general there is no linker, except when we want to build
> some boot-code.
>
> Everything not known at compile-time must then be fixed up when we load
> the module. And to complicate matters, some choices are made by the system
> designer about what we actually _want_ to know at compile time. It is
> basically a cost / benefit choice, sometimes it is easier to leave the
> complete solution until load time, although at least part of the knowledge
> was available at compile-time.
>
> In the case I queried I did not see a reference to the index register that
> LDR uses. That space as left at '0' to be patched at load-time (so it
> looked as if R0 was to be used for an index), although R12, the standard
> index register, could have been written there at compile-time for my
> specific code.
>
> Section 6.3 gives some small introduction to the art of loading. Near the
> end of the page is the line:
> "At the very end of the file three integers called fixorgP, fixorgD, and
> fixorgT are read."
> That is where he talks about the patching.
>
> From the code in Modules.Load and from the compiler description it follows
> that globals need to be treated specially, since they might come of any
> module. So although we might have a partial knowledge of the place of a
> 'local' global at compile time, it is obvious that we don't know at all
> about globals in other modules. So Wirth chose to fix them all in one
> place: the module loader.
>
> In any case: a good way to learn compilers is to target a small micro
> controller that you know very well, and adapt the Oberon 0 compiler. I have
> done it many years ago for an 8 bit AVR, and it has helped me ever since.
>
> Cheers,
>
> j.
>
>
> On Tue, Nov 24, 2015 at 1:22 PM, John Stout <JSS at kgv.ac.uk> wrote:
>
>> Jan
>>
>> Yes, that's the one although my copy is on faded line printer paper!
>>
>> I'm interested in your email about code generation: what does Module.Load
>> patch and to what?
>>
>> I've just bought an OberonStation and am getting to grips with it. I
>> think the first task is to get another nRF24L01+ system working so that I
>> can transfer stuff to and from it.
>>
>> John
>>
>>
>> ----------------------------------------------------------------------
>>
>> Message: 1
>> Date: Mon, 23 Nov 2015 12:09:45 +0200
>> From: Jan de Kruyf <jan.de.kruyf at gmail.com>
>> To: ETH Oberon and related systems <oberon at lists.inf.ethz.ch>
>> Subject: Re: [Oberon] Numerical CASE Statements in Project Oberon
>> Message-ID:
>>         <CAA85NCh6R8Xf-Tp53sFF0=
>> F_5rnH7WXoeTyeDQss-eDMsVMaLg at mail.gmail.com>
>> Content-Type: text/plain; charset="utf-8"
>>
>> Is it this one John?
>>
>> http://homepages.cwi.nl/~steven/pascal/
>>
>> There are links there to a 76 version from eth.
>>
>> j.
>>
>>
>> On Mon, Nov 23, 2015 at 9:47 AM, John Stout <JSS at kgv.ac.uk> wrote:
>>
>> > Well, I should have done a bit more research before making my comments,
>> > and/or done all the exercises in Compiler Construction, which would have
>> > lead me to Professor Wirth's preferred implementation for CASE.
>> >
>> > I have the source code for the Pascal P4 compiler from 1976 (courtesy of
>> > Lancaster University) and that uses this system, although it does place
>> a
>> > maximum of 1000 on the number of cases.
>> >
>> > John
>>
>> [King George V Logo]
>> John Stout
>> Management Information Systems Coordinator
>> Tel: 01704 530601
>> E-Mail: JSS at kgv.ac.uk
>> Web: www.kgv.ac.uk
>>
>>
>> Information contained in this e-mail is intended for the use of the
>> addressee only and is confidential.  Any dissemination, distribution,
>> copying or use of this communication without prior permission of the
>> addresseeis strictly prohibited.
>>
>> Every effort has been made to ensure that any attachment to this mail
>> does not contain virus. King George V College has taken every reasonable
>> precaution to minimise this risk, neither it nor the sender can accept
>> liability for any damage which you sustain as a result of software viruses.
>> You should carry outyour own virus checks before opening any attachments.
>>
>> [Think Before You Print]
>> --
>> Oberon at lists.inf.ethz.ch mailing list for ETH Oberon and related systems
>> https://lists.inf.ethz.ch/mailman/listinfo/oberon
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.inf.ethz.ch/pipermail/oberon/attachments/20151126/ea683917/attachment.html>