Re: [fpc-pascal] Freepascal Floating Point Issue

[Paul Nance]

Turbo Pascal also had a BCD unit.

On Aug 22, 2017 6:56 PM, "Ralf Quint"  wrote:

> On 8/22/2017 1:39 PM, Mark Morgan Lloyd wrote:
> > On 21/08/17 22:15, Ralf Quint wrote:
> >> On 8/21/2017 3:02 PM, James Richters wrote:> I am having an issue
> >> with a simple floating point application.  I am setting a variable to
> >> a specific value and immediately after I set it,  it is not exactly
> >> what I set it to.  Here's an example>>Draw_GX_Min:=999.999;>
> >> Writeln(Draw_GX_Min:3:30);>> The writeln results in
> >> 999.999002  >> Why is it not
> >> 999.999000  where did 0.02 come
> >> from?>Out of thin air... Well, kind of. Double floating point means
> >> 16 digitsof precision, so when you force a 30 digit precision output,
> >> anythingpast 16 digits is random garbage, at best...
> >
> > And in any event, that's probably much more precision than the GPU is
> > using to generate the final image :-)
> >
> Well, older GPUs (at least NVidia, pretty sure similar restrictions
> apply to AMD) use(d) only 32bit "single" floats, giving a 7 digit
> precision, though newer ones can also handle 64bit doubles. But there
> are quite a few differences in how certain FP operations are handled on
> those GPUs, which result in even doubles only having 14 (instead of 16)
> digits of precision at best. So while NVidia keeps mentioning IEEE754,
> their GPUs are in practice not 100% compliant.
>
> As I mentioned before, if someone needs to work a lot with floating
> point arithmetic, it really helps to get yourself acquainted to the way
> those works and all the possible pitfalls.
>
> Btw, anyone know about a BCD math implementation for Free Pascal, like
> it used to be implemented in DR CBASIC? (those were the days... ;-) )
>
> Ralf
>
>
>
> ---
> This email has been checked for viruses by Avast antivirus software.
> https://www.avast.com/antivirus
>
> ___
> fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
> http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal
___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

[fpc-pascal] Check In Interface Type Helpers

[Anthony Walter]

I just wanted to point out that revision 37023 Sven added type helper
support for interfaces.

Good job and thank you Sven!
___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Re: [fpc-pascal] Freepascal Floating Point Issue

[Ralf Quint]

On 8/22/2017 1:39 PM, Mark Morgan Lloyd wrote:
> On 21/08/17 22:15, Ralf Quint wrote:
>> On 8/21/2017 3:02 PM, James Richters wrote:> I am having an issue
>> with a simple floating point application.  I am setting a variable to
>> a specific value and immediately after I set it,  it is not exactly
>> what I set it to.  Here's an example>>    Draw_GX_Min:=999.999;>   
>> Writeln(Draw_GX_Min:3:30);>> The writeln results in
>> 999.999002  >> Why is it not 
>> 999.999000  where did 0.02 come
>> from?>Out of thin air... Well, kind of. Double floating point means
>> 16 digitsof precision, so when you force a 30 digit precision output,
>> anythingpast 16 digits is random garbage, at best...
>
> And in any event, that's probably much more precision than the GPU is
> using to generate the final image :-)
>
Well, older GPUs (at least NVidia, pretty sure similar restrictions
apply to AMD) use(d) only 32bit "single" floats, giving a 7 digit
precision, though newer ones can also handle 64bit doubles. But there
are quite a few differences in how certain FP operations are handled on
those GPUs, which result in even doubles only having 14 (instead of 16)
digits of precision at best. So while NVidia keeps mentioning IEEE754,
their GPUs are in practice not 100% compliant.

As I mentioned before, if someone needs to work a lot with floating
point arithmetic, it really helps to get yourself acquainted to the way
those works and all the possible pitfalls.

Btw, anyone know about a BCD math implementation for Free Pascal, like
it used to be implemented in DR CBASIC? (those were the days... ;-) )

Ralf



---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Re: [fpc-pascal] Freepascal Floating Point Issue

[Mark Morgan Lloyd]

On 21/08/17 22:15, Ralf Quint wrote:

On 8/21/2017 3:02 PM, James Richters wrote:> I am having an issue with a simple floating point 
application.  I am setting a variable to a specific value and immediately after I set it,  it is not 
exactly what I set it to.  Here's an example>>Draw_GX_Min:=999.999;>
Writeln(Draw_GX_Min:3:30);>> The writeln results in 999.999002  >> Why 
is it not  999.999000  where did 0.02 come from?>Out of thin air... 
Well, kind of. Double floating point means 16 digitsof precision, so when you force a 30 digit precision 
output, anythingpast 16 digits is random garbage, at best...


And in any event, that's probably much more precision than the GPU is 
using to generate the final image :-)


--
Mark Morgan Lloyd
markMLl .AT. telemetry.co .DOT. uk

[Opinions above are the author's, not those of his employers or colleagues]
___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Re: [fpc-pascal] Freepascal Floating Point Issue

[Sven Barth via fpc-pascal]

Am 22.08.2017 18:11 schrieb "Peter" :
>
> Hi James,
>
> Its unlikely that 999.999 has an EXACT representation in floating point.
>
> 999.999002  maybe the closest value at your chosen precision.
>
> Extended type has more precision, but still probably won't be exact.

Additionally Extended is not cross platform.

Regards,
Sven
___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Re: [fpc-pascal] Freepascal Floating Point Issue

[Peter]

Hi James,

Its unlikely that 999.999 has an EXACT representation in floating point.

999.999002  maybe the closest value at your chosen precision.

Extended type has more precision, but still probably won't be exact.


Regards,
Peter B

___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Re: [fpc-pascal] H2pas tool vs others, C header translation

[noreply]

On 2017-08-22 04:04, mar...@stack.nl wrote:

In our previous episode, Dmitry Boyarintsev said:


I'm using Chelper.  (http://wiki.freepascal.org/Chelper)
It's my choice, because it preserves comments (which is usually an
important part of open-source libs)
plus, can be configured to handle library header-specific defines.
Naturally, the process is semi-automatic, but get be done quite 
quickly.


Note that to convert C (as in source code, not headers), I did use 
other

tools in the past. Can't remember which though.



I did mean headers, not C source... But C source is also interesting to 
convert of course
I wonder if the Quake to Delphi and Doom To Delphi folks wrote tools for 
conversion, or used existing ones to do the bulk of the work.  If 
superman was involved, they likely did it by hand 100 percent the hard 
way, but if they are A.I. then they have no hands, likely.

___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Re: [fpc-pascal] H2pas tool vs others, C header translation

[noreply]

On 2017-08-21 13:08, mar...@stack.nl wrote:

In our previous episode, Michael Van Canneyt said:

> I am looking to convert the GNU MPFR library header files, or hire
> someone on a bounty to do it if header translation is not so easy with
> h2pas and requires lots of hand work, and close checking.

I do all header translations with h2pas. Most of the time, I run the 
code
first through gcc -E, which will take care of all macro magic. Usually 
h2pas

has little trouble with the result.


Before everything I do a couple of substs with an editor,



What do you mean by substs..

Search/replace, (substitutions? regex?)


gcc -E I use small pascal filterprograms to sort the symbols based on
heuristics.  Usually "clean" defines, structs, function calls and 
"rest".


A custom filter program for each case by case basis, or one major filter 
program you wrote for all header files?


Thanks for the tips.
___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Re: [fpc-pascal] H2pas tool vs others, C header translation

[Marco van de Voort]

In our previous episode, Dmitry Boyarintsev said:
> 
> I'm using Chelper.  (http://wiki.freepascal.org/Chelper)
> It's my choice, because it preserves comments (which is usually an
> important part of open-source libs)
> plus, can be configured to handle library header-specific defines.
> Naturally, the process is semi-automatic, but get be done quite quickly.

Note that to convert C (as in source code, not headers), I did use other
tools in the past. Can't remember which though.
___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Re: [fpc-pascal] Freepascal Floating Point Issue

[bernd.oppol...@t-online.de]

https://www.facebook.com/notes/new-stanford-pascal-compiler/floating-point-output-differences-between-platforms/365845753799072

 
/

other Pascal Implementations have to provide solutions to this issue, too

Kind regards

Bernd




Gesendet mit der Telekom Mail App



--- Original-Nachricht ---
Von: Sven Barth via fpc-pascal
Betreff: Re: [fpc-pascal] Freepascal Floating Point Issue
Datum: 22.08.2017, 7:54 Uhr
An: FPC-Pascal users discussions
Cc: Sven Barth



Am 22.08.2017 00:02 schrieb "James Richters" <
ja...@productionautomation.net  >:

The others already wrote about floating point precision, so I won't repeat 
that.

> The reason I noticed this is because I have some conditional statements 
like
> If Draw_GX_Min<>99.999 then
> Something

*DON'T* compare floating point numbers like that. Instead use SameValue() 
from the Math unit which has an additional epsilon value (with a default 
value) that defines a lower and upper bound from your desired value to 
compare to.

Regards,
Sven___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal

Re: [fpc-pascal] H2pas tool vs others, C header translation

[noreply]

On 2017-08-21 20:34, Dmitry Boyarintsev wrote:

On Sun, Aug 20, 2017 at 5:37 PM,  wrote:


Is there a competing tool that does more, or is updated more?


How about this list?
http://wiki.freepascal.org/C_to_Pascal


That's a good one, I missed that.. could not find it previously when 
searching google


Did you write CHelper because H2pas was not to your satisfaction?
___
fpc-pascal maillist  -  fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal