Gpc June 2010

gpc@gnu.de

8 participants
7 discussions

Inefficiency of `mod'
by Emil Jerabek 25 Oct '10

25 Oct '10

Hi, during some debugging I noticed that GPC generates very hairy code for the `mod' operator with signed arguments. A closer look revealed that the code contains a lot of branching on conditions which can be determined on compile time, and most surprisingly, the value of X mod Y is recomputed _twice_, if X is negative. (I've attached a commented assembler dump.) This may be partially a back-end problem (CSE?), but I feel there's something wrong in the relevant code in gpc-common.c (build_pascal_binary_op) as well. `exp1 mod exp2' gets translated into something like exp2 <= 0 ? error : exp1 >= 0 ? exp1 mod exp2 : ((-exp1) mod exp2) == 0 ? 0 : exp2 - (-exp1) mod exp2 This arranges that both arguments supplied to the "internal mod" are non-negative, but this information is not passed on -- the arguments are still declared as signed types, which apparently makes the back-end to produce the mess it does. I've done a little test: the following replacement for `mod', which differs from the built-in one essentially just by the type-casts to `Cardinal', inline operator xmod (X, Y: Integer) = Z: Integer; begin if Y <= 0 then RuntimeError (714); if X >= 0 then Z := Cardinal (X) mod Y else begin Z := Cardinal (-X) mod Y; if Z <> 0 then Z := Y - Z end end; indeed generates a better code, it measures about 15% faster for positive X and more than twice faster for negative X. YMMV. If this is possible in user code, it should work in the compiler internals as well :) The result type of the whole `mod' expression looks dubious too. IIUIC, `exp1 mod exp2' inherits the type of exp1. Now, the result of `mod' is guaranteed to be nonnegative and to fit in the type of exp2 (or more precisely, it is in 0 .. High (type of (exp2)) - 1), whereas it needn't have anything to do with the type of exp1: consider var X, Z: Integer; Y: LongInt; begin X := -8; Y := High (LongInt); Z := X mod Y { -> overflow } end Or am I missing something? Emil (BTW, where is fjf434c.pas?)

10 16

Re: close(file); always necessary??
by samiam＠moorecad.com 19 Jun '10

19 Jun '10

Hence the reference count stuffs. I did a paper analysis once, I decided that: 1. To do true automated storage recycling, you have to track each pointer leaving scope, which means that you have to know where every pointer is. 2. Be ready to reference count each pointer, that is, each copy operation increments the reference count, and each dynamic entry gets a reference counter that is initially one for the pointer that is returned. 3. Decrement the reference counter each time any pointer leaves scope. 4. produce an error if the programmer attempts to free a non-zero referred pointer. 5. The whole thing generates headaches. Seriously, it is quite doable with Pascal, but generates a lot of overhead. Each data structure needs a template for the contained pointers, and every data structure leaving scope or freed must be analysed for contained pointers, and those pointers decrementing reference counts. Wirth also described this scheme, although I don't think he ever tried it. I have thought about implementing it primarily to shut people up, i.e., lack of automated storage collection "is a problem with old languages like Pascal". It isn't, the price paid for reclaimation overhead is the same as Java or other languages, we are primarily burdened by users who expect true compiled performance code. Scott -------- Original Message -------- Subject: Re: close(file); always necessary?? From: Frank Heckenbach <ih8mj(a)fjf.gnu.de> Date: Fri, June 18, 2010 2:56 pm To: gpc(a)gnu.de samiam(a)moorecad.com wrote: > Then that would leave you pretty close to > automatic dynamic storage recycling, no ? Partly. When a pointer variable goes out of scope, we cannot automatically dispose it, since copies may exist. Fully automatic storage recycling still requries full-blown reference counting or garbage collection, AFAICS. Frank -- Frank Heckenbach, f.heckenbach(a)fh-soft.de, http://fjf.gnu.de/, 7977168E GPC To-Do list, latest features, fixed bugs: http://www.gnu-pascal.de/todo.html GPC download signing key: ACB3 79B2 7EB2 B7A7 EFDE D101 CD02 4C9D 0FE0 E5E8

2 1

Re: close(file); always necessary??
by samiam＠moorecad.com 18 Jun '10

18 Jun '10

2 1

Re: close(file); always necessary??
by samiam＠moorecad.com 15 Jun '10

15 Jun '10

2 1

close(file); always necessary??
by Rugxulo 15 Jun '10

15 Jun '10

Hey, In non-standard Pascal mode, does it matter if I explicitly close the file or not? E.g., "close(file);", is that done automatically or is it implementation defined? (I'm 99% sure DOS always closes files upon exit.) I'm just wondering if that's a fair assumption for Linux, etc. (though I kinda doubt it).

3 2

3 points
by spir 14 Jun '10

14 Jun '10

Hello, I'm am amateur programmer, new to gpc and rather new to Pascal as well. I wonder whether this list is a proper place to find help in learning the language, and in using it for a few projects of mine. If yes, then below is a first series of diverse questions. I will have some more as I discover the language (including some stupid ones, certainly ;-). [Note: I have a candidate answer to the questions, but ask anyway to be sure, and stop wondering.] -1- writing arrays & records What is the literal notation for arrays & records? Can it apply to (a) const definition (b) var initialisation (c) var assignment? Is there a way to write back arrays & records to output (tostring(x), according to standard notation)? If not, how to write such a tool func? (super useful for learning, prototyping, testing...) Is it at all possible to write the following 'show' tool func: show(x); // 'x : 1' --> stdout show(x+1); // 'x+1 : 2' --> stdout ? (my dream debugging tool) -2- traversing collections Is there a way to traverse arrays with elements directly? (--> foreach, or Eiffel's across) Is it possible to define traversal for custom collections? -3- TP-like object system Is there more information in fpc's documentation about its object system (than chapter 6.8 in manual)? Else, are there good places online to find information about TP's object system? Denis ________________________________ vit esse estrany ☣ spir.wikidot.com

4 5

mixed comments
by Rugxulo 06 Jun '10

06 Jun '10

With apologies if this has already been mentioned. (I don't feel like searching the mailing list archives right now, and I haven't checked the test suite.) I just thought it was funny / interesting. ;-) { using GPC 20070904 / GCC 3.4.4 via DJGPP (gpc.info) : "GPC's default behaviour is (like BP) not to allow mixed comments" ... except that it allows it by ignoring it, not reporting an error (which is what was implied, or so I thought!) } { $ extended-pascal} { says "Hello, mixed!" } { $ standard-pascal} { says "Hello, mixed!" } {GPC default} { says "Hello" } {FPC default} { says "Hello" } program comment(output); begin write('Hello'); { *) write(', mixed!'); (* } writeln; end.

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

Gpc June 2010