Board index » off-topic » Re: fpc beats gcc? delphi?

Re: fpc beats gcc? delphi?


2007-05-15 03:51:08 PM
off-topic3
yannis wrote:
Quote
I suggest to both of you to go to Non technical group and read up the
thread with the same title. It is something that have been asked
there and it is not for every body to use or trast for that matter.
if you do have something usefull to add then do it on that thread.
I wont comment any more, but I wont venture into non-tech it just
waffles on and too much bickering
--
Liz the Brit
Delphi things I have released: www.xcalibur.co.uk/DelphiThings
 
 

Re:Re: fpc beats gcc? delphi?

Liz used his keyboard to write :
Quote
yannis wrote:


>I suggest to both of you to go to Non technical group and read up the
>thread with the same title. It is something that have been asked
>there and it is not for every body to use or trast for that matter.
>if you do have something usefull to add then do it on that thread.

I wont comment any more, but I wont venture into non-tech it just
waffles on and too much bickering
aggred for most of the threads in there. This one seems a little bit
different though. But in your potision i wouldn't venture in that group
either. I am {*word*60}ed to non-tech (what a waste of time).
Regards
Yannis.
 

Re:Re: fpc beats gcc? delphi?

On 2007-05-15, Michael Anonymous < XXXX@XXXXX.COM >wrote:
Quote
Marco van de Voort wrote:
<snip>

>>I turned on "Create smartlinkable units" and "Link to smart libraries" in the FPC IDE.
>>I couldn't find stripping.
>
>compiler options ->linking ->debugging groupbox last option.

In the Free Pascal IDE ... hmm.
There's no compiler options menu ... there's an "Options" menu.
Then options->de{*word*81}, first item.
Quote
>Do you use a snapshot or a release? (is the last digit of your FPC release
>odd or even? Even is release)

For Lazarus I used 0.9.23 beta.
The reason for the question was the speedup the internal linker provides. This to get an idea
what's behind your ocmplaint. (e.g. a IDE with a version with internal
linker enabled, and lazarus not, because they originate from different
installer packages)
And for that I need the FPC version.
Quote
>It's not a product, and nobody is an user. It is a community project
>remember ? :-)

One of the definitions of a "product" is something that is procedure.
Lazarus was something that was produced.
Therefore, it's a product.
However strictly a produced good is not the strict context you used
"product" in, so is irrelevant to this discussion.
 

{smallsort}

Re:Re: fpc beats gcc? delphi?

On 2007-05-13, Alan Garny < XXXX@XXXXX.COM >wrote:
Quote
Daniël Mantione has given some of the reasons as to why C++ Builder/Delphi
are not efficient compilers.
IMHO "efficient compiler" is way a too general term for this. I'd notice the
smallest ansistring performance increase or loop optimization in my apps,
but anything floating point would have to be emulated to noticably slow
down. (or maybe even not even that)
Efficient is not a scalar, but a bunch of vectors.
 

Re:Re: fpc beats gcc? delphi?

Lars G formulated the question :
Quote
"Maël Hörz" < XXXX@XXXXX.COM >skrev i en meddelelse
news: XXXX@XXXXX.COM ...
>>You cannot take any of these tests seriously. It all depends on what you
>>are interested in. I know that for what I do (scientific computation),
>>Delphi and C++ Builder are really bad compared to other commercial C++
>>compilers...
>Yeah, what a pity. Since Delphi 3 I hoped they would change that. Maybe for
>the x64 compiler...


Please vote :-)

FPU optimizations 1 to 7 :-)
qc.codegear.com/wc/qcmain.aspx
qc.codegear.com/wc/qcmain.aspx
qc.codegear.com/wc/qcmain.aspx
qc.codegear.com/wc/qcmain.aspx
qc.codegear.com/wc/qcmain.aspx
qc.codegear.com/wc/qcmain.aspx
qc.codegear.com/wc/qcmain.aspx

Function/procedure Entry Point Alignment
qc.codegear.com/wc/qcmain.aspx
Well, I think that sums it up.
27 total votes for 7 reports, that are
over 3 years old. That's an average of
just over 1 vote per issue per year.
Why should CG bother.
Backing up a bit, there was some discussion
about adding optimizations to the compiler.
I think that is a great idea.
Some thought that was too hard and would be
too much work, too time consuming when there
were other things that needed to be done.
From my perspective, I completely trust the
FastCode guys to vote on and suggest the "low
hanging fruit" ideas that would make the most
difference and be the easiest to implement.
After those are done, then we can start
working on some harder issues. There
are some optimizations that can't be done
in a FastCode routine. They really need
to be done by the compiler. Stack frames
comes to mind. Loop unrolling.
But if the issues posted so far get hardly any
votes, I don't think that we can complain
about CG not putting enough time into such
an "important" issue. And I don't blame Dennis
or other FastCoders for not entering any
more reports as long as these easy ones are
lingering.
Excuse me, I have some voting to do.
Brad.
 

Re:Re: fpc beats gcc? delphi?

Brad White wrote :
Quote
Lars G formulated the question :
>"Maël Hörz" < XXXX@XXXXX.COM >skrev i en meddelelse
>news: XXXX@XXXXX.COM ...
>>>You cannot take any of these tests seriously. It all depends on what you
>>>are interested in. I know that for what I do (scientific computation),
>>>Delphi and C++ Builder are really bad compared to other commercial C++
>>>compilers...
>>Yeah, what a pity. Since Delphi 3 I hoped they would change that. Maybe
>>for the x64 compiler...
>
>
>Please vote :-)
>
>FPU optimizations 1 to 7 :-)
>qc.codegear.com/wc/qcmain.aspx
>qc.codegear.com/wc/qcmain.aspx
>qc.codegear.com/wc/qcmain.aspx
>qc.codegear.com/wc/qcmain.aspx
>qc.codegear.com/wc/qcmain.aspx
>qc.codegear.com/wc/qcmain.aspx
>qc.codegear.com/wc/qcmain.aspx
>
Report No: 1142 (RAID: 241739) Status: Open
Retention of statement results on FP stack (floating point
optimization)
The forced save/load on each iteration can easily make this routine
3 times slower than it need be.
qc.CodeGear.com/wc/qcmain.aspx
is another one. 5 years old with a paltry
3 votes. One of which, I'm glad to see, was mine.
Brad.
 

Re:Re: fpc beats gcc? delphi?

John Herbster wrote:
Quote
"Bruce McGee" < XXXX@XXXXX.COM >wrote in message
news: XXXX@XXXXX.COM ...
>Free Pascal benchmark application built with Delphi 2007.

Who is going run an EXE from an untrusted source?
You don't trust Bruce?
--
TJSDialog - TaskDialog for other operating systems:
www.jed-software.com/jsd.htm
Visual Forms IDE Add In: www.jed-software.com/vf.htm
Blog: jedqc.blogspot.com
 

Re:Re: fpc beats gcc? delphi?

Quote
>Who is going run an EXE from an untrusted source?
"JED" < XXXX@XXXXX.COM >wrote
You don't trust Bruce?
Yes, I do not trust Bruce! I do not know Bruce.
Further, I do not trust the distribution mechanism.
--JohnH