ftp.xmission.com/2014.06.ftp.xmission.com.tar/ftp.xmission.com/pub/lists/fractdev/archive/fractdev.9806

Internet Message Format | 1998-06-20 | 57KB

From: "Tim Wegner" <[e-mail]>
Subject: (fractdev) Fractint porting
Date: 01 Jun 1998 22:28:25 -0600

I've started actively researching a new development environment for Fractint. I
bought Microsoft Visual C/C++ with that in mind.

My tentative conclusion after a survey this weekend is to port fractint to djgpp
with the allegro library. This is the extended DOS port of the GNU C compiler,
along with a game graphics library. Some of the pluses of this are:

1. free platform
2. excellent, fast compiler
3. ready porting to Win95 and Linux

The main downside is that porting the assembler would be tricky, so we'd take
a performance hook.

I'm particularly intrigued with the Win95/DirectX port of Allegro as a porting
possibility.

Comments?

Tim


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractdev) Fractint porting
Date: 01 Jun 1998 23:38:38 -0400

Tim,

 - I'm particularly intrigued with the Win95/DirectX port of Allegro as a
 - porting possibility.

DirectX would be overkill for what FractInt needs.  However, if it's part
of a Windows port of Allegro, then you're basically getting it for free.
Keep in mind, though, that some older systems don't have good DirectDraw
support, and can lock up or crash with despairing regularity with programs
that try to use DirectDraw.  Newer cards don't have this problem.

Damien M. Jones   \\
[e-mail]  \\  http://www.icd.com/tsd/ (temporary sanity designs)
                    \\  http://www.fractalus.com/ (fractals are my hobby)

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: <[e-mail]>
Subject: Re: (fractdev) Fractint porting
Date: 02 Jun 1998 09:59:47 EDT

In a message dated 98-06-01 23:29:02 EDT, you write:

<< I've started actively researching a new development environment for
Fractint. I
 bought Microsoft Visual C/C++ with that in mind.

 My tentative conclusion after a survey this weekend is to port fractint to
djgpp
 with the allegro library. This is the extended DOS port of the GNU C
compiler,
 along with a game graphics library. Some of the pluses of this are:

 1. free platform
 2. excellent, fast compiler
 3. ready porting to Win95 and Linux

 The main downside is that porting the assembler would be tricky, so we'd take
 a performance hook.

 I'm particularly intrigued with the Win95/DirectX port of Allegro as a
porting
 possibility.

 Comments? >>
Tim,
I am currently unemployed as a result of a consulting group that downsized my
company and eliminated a layer of management. I may finally have time to be a
real participant. I would at least like to contribute to the true color
algorithm side. I believe Terry is now using my algorithm as one of the
options with his program.
Ron Barnett

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: [e-mail]
Subject: Re: (fractdev) Fractint porting
Date: 03 Jun 1998 14:06:45 -0400 (EDT)

On Mon, 1 Jun 1998, Tim Wegner wrote:

> I've started actively researching a new development environment for Fractint. I
> bought Microsoft Visual C/C++ with that in mind.
>
> My tentative conclusion after a survey this weekend is to port fractint to djgpp
> with the allegro library. This is the extended DOS port of the GNU C compiler,
> along with a game graphics library.

I think this is a good idea.  I'm currently trying to compile fractint
with djgpp...if I ever get it working, I'll let y'all know how it goes...

Cliff Crawford                             |            Windows NT crashed.
http://www.people.cornell.edu/pages/cjc26/ | I am the Blue Screen of Death.
                                           |     No one hears your screams.

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Christian Strik" <[e-mail]>
Subject: Re: (fractdev) Fractint porting
Date: 03 Jun 1998 22:07:35 +0100

Cliff wrote:

<<snipped>>
>I'm currently trying to compile fractint with djgpp...if I ever get it
working

Is it so the problem you're having is an exception 0x0E (page fault) when
starting any djgpp compiler/debugger? If so, I'm having the same problem.
Does anybody know a solution? I'm currently using TurboC++ 3.0, but want to
switch to djgpp asap.

Christian




Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: [e-mail]
Subject: Re: (fractdev) Fractint porting
Date: 03 Jun 1998 16:27:31 -0400 (EDT)

On Wed, 3 Jun 1998, Christian Strik wrote:

> Is it so the problem you're having is an exception 0x0E (page fault) when
> starting any djgpp compiler/debugger? If so, I'm having the same problem.
> Does anybody know a solution? I'm currently using TurboC++ 3.0, but want to
> switch to djgpp asap.

Hmm, not exactly sure what's going on.  Did you install everything
correctly, including setting the path and the environment variable?
("set DJGPP=C:\DJGPP\DJGPP.ENV") I remember at first I forgot to do that
and all sorts of weird problems occured.

No, the problems I'm having are with the source code itself not being
compatible.  I think I'm gonna try compiling the xfractint source
instead (since djgpp is so unix-ish), I might have more luck with that.

Cliff Crawford                             |            Windows NT crashed.
http://www.people.cornell.edu/pages/cjc26/ | I am the Blue Screen of Death.
                                           |     No one hears your screams.

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Peter Gavin" <[e-mail]>
Subject: RE: (fractdev) Fractint porting
Date: 04 Jun 1998 00:06:16 -0400

The problems I'm having with compiling it using DJGPP are all the =
#define in 'port.h'.  It seem to me that the best way would be to make a =
section combining the XFRACT section and the MSDOS section.  Well, you =
may just be able to do it by copying the 'remove far' section to the =
MSDOS section.  I really haven't had the time to mess with it that much, =
unfortunately...  :(



Pete
<[e-mail]>

//  -----Original Message-----
//  From: [e-mail]
//  [mailto:[e-mail]]On Behalf Of =
[e-mail]
//  Sent: Wednesday, June 03, 1998 4:28 PM
//  To: [e-mail]
//  Subject: Re: (fractdev) Fractint porting
// =20
// =20
//  On Wed, 3 Jun 1998, Christian Strik wrote:
// =20
//  > Is it so the problem you're having is an exception 0x0E (page=20
//  fault) when
//  > starting any djgpp compiler/debugger? If so, I'm having the=20
//  same problem.
//  > Does anybody know a solution? I'm currently using TurboC++=20
//  3.0, but want to
//  > switch to djgpp asap.
// =20
//  Hmm, not exactly sure what's going on.  Did you install everything
//  correctly, including setting the path and the environment variable?
//  ("set DJGPP=3DC:\DJGPP\DJGPP.ENV") I remember at first I forgot=20
//  to do that=20
//  and all sorts of weird problems occured.=20
// =20
//  No, the problems I'm having are with the source code itself not =
being=20
//  compatible.  I think I'm gonna try compiling the xfractint source=20
//  instead (since djgpp is so unix-ish), I might have more luck with =
that.
// =20
//  -------------------------------------------+--------------------
//  -----------
//  Cliff Crawford                             |            Windows=20
//  NT crashed.
//  http://www.people.cornell.edu/pages/cjc26/ | I am the Blue=20
//  Screen of Death.
//                                             |     No one hears=20
//  your screams.
// =20
//  --------------------------------------------------------------
//  Thanks for using Fractdev, The Fractint Developer's Discussion List
//  Post Message:   [e-mail]
//  Get Commands:   [e-mail] "help"
//  Administrator:  [e-mail]
//  Unsubscribe:    [e-mail] "unsubscribe fractdev"
// =20


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Christian Strik" <[e-mail]>
Subject: Re: (fractdev) Fractint porting
Date: 04 Jun 1998 13:05:19 +0100

>Did you install everything
>correctly, including setting the path and the environment variable?
>("set DJGPP=C:\DJGPP\DJGPP.ENV")

Thanx. Indeed the problem was the djgpp env var. I wrote
    set DJGPP=F:\PRGRMMNG\DJGPP\DGJPP.ENV
instead of ...\DJGPP.ENV.

Christian





Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: [e-mail]
Subject: RE: (fractdev) Fractint porting
Date: 04 Jun 1998 16:35:01 -0400 (EDT)

On Thu, 4 Jun 1998, Peter Gavin wrote:

> The problems I'm having with compiling it using DJGPP are all the
#define in 'port.h'.  It seem to me that the best way would be to make a
section combining the XFRACT section and the MSDOS section.  Well, you may
just be able to do it by copying the 'remove far' section to the MSDOS
section.  I really haven't had the time to mess with it that much,
unfortunately...  :(

Yes, I took care of the near/far problem (by undefining 'far'), and I've
got everything compiled, I'm just having trouble with the linking stage.
Seems the pre-compiled object files (corresponding to the assembly files)
aren't compatible with djgpp's linker.  I really don't want to try and
assemble all those files with the djgpp assembler, because I suspect that
they were also written in the 16-bit memory model, so I would have to go
through and change it all to 32-bit, which I suspect will be a LOT harder
than changing the C source code.  Anyone know what else I can do?

Cliff Crawford                             |            Windows NT crashed.
http://www.people.cornell.edu/pages/cjc26/ | I am the Blue Screen of Death.
                                           |     No one hears your screams.

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Tim Wegner" <[e-mail]>
Subject: RE: (fractdev) Fractint porting
Date: 04 Jun 1998 17:46:40 -0600

Cliff wrote:

> Seems the pre-compiled object files (corresponding to the assembly files)
> aren't compatible with djgpp's linker.  I really don't want to try and
> assemble all those files with the djgpp assembler, because I suspect that
> they were also written in the 16-bit memory model, so I would have to go
> through and change it all to 32-bit, which I suspect will be a LOT harder
> than changing the C source code.  Anyone know what else I can do?

You definitely can't use the assembler files because they are compiled for the
medium memory model.

Your best bet is to start with Xfractint, but you'll have to deal with any Xwindows-
specific stuff.

One way or another you have to bite a bullet to do a port - it isn't just a
question of tweaking the compiler so the code works. But starting with Xfractint
eliminates all assembler problems.

Tim



Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: [e-mail] (Kragen)
Subject: RE: (fractdev) Fractint porting
Date: 05 Jun 1998 09:25:27 -0400 (EDT)

On Thu, 4 Jun 1998 [e-mail] wrote:
> Seems the pre-compiled object files (corresponding to the assembly files)
> aren't compatible with djgpp's linker.  I really don't want to try and
> assemble all those files with the djgpp assembler, because I suspect that
> they were also written in the 16-bit memory model, so I would have to go
> through and change it all to 32-bit, which I suspect will be a LOT harder
> than changing the C source code.

There's also the MASM syntax vs. gas syntax thing.  gas uses a
different assembly-language syntax (% on all the register names,
slightly different names for most opcodes, no BYTE PTR, different
operand order) which is fairly uniform across processor families.  So
it's likely to be even harder than you thought :)

> Anyone know what else I can do?

If you don't feel like translating 16-bit MASM assembly code to 32-bit
gas assembly -- a feat which doesn't sound like it should be
tremendously difficult, but is nevertheless much harder than
recompiling something with a different compiler -- you could use the C
versions of the assembly routines.  I don't know if they exist in the
main source code base, but they certainly exist in xfractint, because I
ran xfractint on an RS/6000 in 1995.

Kragen


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Tim Wegner" <[e-mail]>
Subject: (fractdev) Visual C/C++ 5
Date: 08 Jun 1998 18:45:25 -0600

I am sure lots of folks know this stuff, but remember, I'm a beginner <grin!>

I've been playing with Microsoft Visual C/C++ 5.0 and have noticed the following:

1. The compiler has a "console" mode which can be run from DOS or
Windows. The executable program runs as DOS program; under windows a
DOS box is created. In this mode the compiler works like the good ol' MSC 7
that I use for the DOS fractint. Two big differences: long filenames are
supported, and even when running from a DOS Box (inside Win95 of course), a
flat memory model is used, so large amounts of memory can be malloc'ed. Of
course a DOS Fractint compiled in this way wouldn't run under a true DOS. A
few diehard Fractint users like Lee SKinner don't run Windows at all <g!> He'd
have to install WIn95 to use such a program.

2. It isn't hard to link Visual C with assembler. I was able to link in some test
code compiled with MASM 6.1. However it does look like the documentation is
more oriented to inline assembler. For portability, we could keep the
"assembler" files separate, and they could be implemented as either pure
assembler or as one big inline file. All assembler, of course, must use the flat
memory model.

I'm not so sure it would be all that hard to port some of the more critical
assembler to Visual C/C++ 5. This compiler doesn't support Windows 3.1, but
I'm not worried about supporting 16 bit Windows.

Tim




Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: Lee Skinner <[e-mail]>
Subject: (fractdev) Visual C/C++
Date: 08 Jun 1998 21:06:02 -0400

>>A few diehard Fractint users like Lee SKinner don't run Windows at all
<g!> He'd have to install WIn95 to use such a program.<<

I heard that!<g>

Actually I do use Win95 occasionally when writing CDs or making thumbnail=
s
with CPIC, but still run Fractint under DOS 6.22

Lee

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: [e-mail] (F.J. Slijkerman)
Subject: Re: (fractdev) Visual C/C++ 5
Date: 09 Jun 1998 09:38:07 +0200 (MET DST)


Tim,

I think you need to take the same approach as POV-Ray.
Basically, create a 32-bit fractal engine completely written
in C. Then you can build a user interface on top of that,
and rewrite routines in assembler for each platform.
I would recommend to use inline assembly for these routines --
that's ok since you will only use them with VC++, not with
other platforms.

Regards,
Frederik.

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: [e-mail]
Subject: Re: (fractdev) Visual C/C++ 5
Date: 09 Jun 1998 11:34:05 +0000

> Basically, create a 32-bit fractal engine completely written
> in C. Then you can build a user interface on top of that,
>
    I like that.
---
  Linux et Cassoulet  ;===)  http://savage.iut-blagnac.fr/
  Rien a dire ...     ;===)  http://www.mygale.org/08/oulala/

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: Phil McRevis <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 09 Jun 1998 16:26:51 -0600


Hi, after a long hiatus, I am back (sort of) :).
I'm coming in late on this discussion, but I might be
able to lend some insights...please accept my apologies
in advance if I'm repeating what's already been said.

In article <[e-mail]>,
    [e-mail] (F.J. Slijkerman)  writes:
> Basically, create a 32-bit fractal engine completely written
> in C. Then you can build a user interface on top of that,

If you start with the xfractint sources, this will be much
easier than starting with the fractint sources.  The xfractint
code has all the assembly replaced with C routines which compile
on various flavors of unix.  Its a little slow (especially the
formula-based fractals) but it does indeed work.

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 09 Jun 1998 21:49:26 -0600

Phil wrote:

> If you start with the xfractint sources, this will be much
> easier than starting with the fractint sources.  The xfractint
> code has all the assembly replaced with C routines which compile
> on various flavors of unix.  Its a little slow (especially the
> formula-based fractals) but it does indeed work.

That's the obvious approach, but we won't be able to retire the DOS version
until there's an equally fast version. As a non-expert, I'm encouraged that flat
model assembler doesn't look so bad. We could probably port the assembler to
djgpp and Linux, also, but that would be more work.

My idea is to eliminate the integer math, which in one fell swoop eliminates a
lot of assembler. The mandel/julia assembler and the parser assembler are the
most important to keep.

Tim


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 09 Jun 1998 23:20:55 -0400

Tim,

 - As a non-expert, I'm encouraged that flat model assembler doesn't look
 - so bad.

It's actually enormously easier than 16-bit modes... you never have to
touch the segment registers, everything is in just one linear address space
(it isn't really, but that's how the OS makes it look to your app) and it's
just Way Cool.

Damien M. Jones   \\
[e-mail]  \\  http://www.icd.com/tsd/ (temporary sanity designs)
                    \\  http://www.fractalus.com/ (fractals are my hobby)

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: [e-mail] (Kragen)
Subject: Re: (fractdev) Visual C/C++ 5
Date: 10 Jun 1998 00:04:27 -0400 (EDT)

On Tue, 9 Jun 1998, Tim Wegner wrote:
> model assembler doesn't look so bad. We could probably port the assembler to
> djgpp and Linux, also, but that would be more work.

But productive work.

> My idea is to eliminate the integer math, which in one fell swoop eliminates a
> lot of assembler. The mandel/julia assembler and the parser assembler are the
> most important to keep.

I really like integer math a lot, especially if I'm going to run
xfractint on my 386-16.

Kragen


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: [e-mail] (Kragen)
Subject: Re: (fractdev) Visual C/C++ 5
Date: 10 Jun 1998 00:07:15 -0400 (EDT)

On Tue, 9 Jun 1998, Damien M. Jones wrote:
>  - As a non-expert, I'm encouraged that flat model assembler doesn't look
>  - so bad.
>
> It's actually enormously easier than 16-bit modes...

It's much easier to translate 16-bit assembler into 16-bit assembler
(using, for example, copy, cat, or xcopy) than to translate 16-bit
assembler into 32-bit assembler (which requires quite a bit of work).
Certain parts of fractint are currently written in 16-bit assembler.

Kragen


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 10 Jun 1998 00:45:48 -0400

Kragen,

 - It's much easier to translate 16-bit assembler into 16-bit assembler
 - (using, for example, copy, cat, or xcopy) than to translate 16-bit
 - assembler into 32-bit assembler (which requires quite a bit of work).

I didn't mean to imply otherwise.  On the whole, writing a routine in
32-bit assembly to begin with is a heckuva lot easier than writing the same
technique in 16-bit assembly.  Having done this, I can attest that this is
very much the case, and THAT is the point I was trying to make here.

Damien M. Jones   \\
[e-mail]  \\  http://www.icd.com/tsd/ (temporary sanity designs)
                    \\  http://www.fractalus.com/ (fractals are my hobby)

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 10 Jun 1998 22:56:37 -0600

kragen wrote:

> I really like integer math a lot, especially if I'm going to run
> xfractint on my 386-16.

yes, but with every passing day a 386 is an unusual platform, and the old DOS
fractint can always be used. My view is that these old platforms shouldn't block
Fractint's progress.

On the other hand, if we get help porting the assembler,  we could keep the
integer math.

Tim


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Peter Gavin" <[e-mail]>
Subject: RE: (fractdev) Visual C/C++ 5
Date: 11 Jun 1998 01:28:14 -0400

And besides, who in their right mind would even *try* to run Win95 on a =
386?  <g>   :)

Well, then again, DJGPP has a built in 32 bit extender, correct?

Pete
<[e-mail]>

//  -----Original Message-----
//  From: [e-mail]
//  [mailto:[e-mail]]On Behalf Of Tim Wegner
//  Sent: Thursday, June 11, 1998 12:57 AM
//  To: [e-mail]
//  Subject: Re: (fractdev) Visual C/C++ 5=20
// =20
// =20
//  kragen wrote:
// =20
//  > I really like integer math a lot, especially if I'm going to run
//  > xfractint on my 386-16.
// =20
//  yes, but with every passing day a 386 is an unusual platform,=20
//  and the old DOS=20
//  fractint can always be used. My view is that these old=20
//  platforms shouldn't block=20
//  Fractint's progress.
// =20
//  On the other hand, if we get help porting the assembler,  we=20
//  could keep the=20
//  integer math.
// =20
//  Tim
// =20
// =20
//  --------------------------------------------------------------
//  Thanks for using Fractdev, The Fractint Developer's Discussion List
//  Post Message:   [e-mail]
//  Get Commands:   [e-mail] "help"
//  Administrator:  [e-mail]
//  Unsubscribe:    [e-mail] "unsubscribe fractdev"
// =20


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Peter Gavin" <[e-mail]>
Subject: RE: (fractdev) Visual C/C++ 5
Date: 11 Jun 1998 01:41:27 -0400

Wait, can you explain this to me?  The console mode programs made by Vis =
C++ 5 are not DOS programs at all, as far as I know.  I don't think you =
can run a 32-bit console mode program under DOS, even with a 32-bit DOS =
extender.  First off, DOS'll spit out the typical 'This program must be =
run under Windows' BS.  (Try using the FTP console mode program that  =
comes w/ windows under DOS, with a DOS TCP/IP stack, like Wattcp.  It =
won't work.  And the 'This program...' BS is really spit out by the =
Windows program itself, you can hook in any code you want, that's just =
the default behavior.)  Second, I don't think console mode progs can =
access vid hardware the way fractint does, windows doesn't like that, =
'specially w/ protected mode etc.  Third, did you translate the assembly =
to 32-bit, or what? <g> :)

BTW, not to be an anal-retentive dork or anything <g>, but the Vis C++ =
IDE runs that same console  mode compiler you found, it just pipes the =
output through that little box in the bottom.  The wonders of =
Microsoft's programming team! ;)

I have Vis C 4, which is (nearly) identical to 5, other than the MFC =
version included (which is only about .01 version-points better) and =
cosmetic "enhancements".  (Again, MS's team at work! <g>)  I would =
really like to see what you did to get that working!  Could you email =
me?

Thanks!

Pete
<[e-mail]>

//  -----Original Message-----
//  From: [e-mail]
//  [mailto:[e-mail]]On Behalf Of Tim Wegner
//  Sent: Monday, June 08, 1998 8:45 PM
//  To: [e-mail]
//  Subject: (fractdev) Visual C/C++ 5
// =20
// =20
//  I am sure lots of folks know this stuff, but remember, I'm a=20
//  beginner <grin!>
// =20
//  I've been playing with Microsoft Visual C/C++ 5.0 and have=20
//  noticed the following:
// =20
//  1. The compiler has a "console" mode which can be run from DOS or=20
//  Windows. The executable program runs as DOS program; under windows a =

//  DOS box is created. In this mode the compiler works like the=20
//  good ol' MSC 7=20
//  that I use for the DOS fractint. Two big differences: long=20
//  filenames are=20
//  supported, and even when running from a DOS Box (inside Win95=20
//  of course), a=20
//  flat memory model is used, so large amounts of memory can be=20
//  malloc'ed. Of=20
//  course a DOS Fractint compiled in this way wouldn't run under a=20
//  true DOS. A=20
//  few diehard Fractint users like Lee SKinner don't run Windows=20
//  at all <g!> He'd=20
//  have to install WIn95 to use such a program.
// =20
//  2. It isn't hard to link Visual C with assembler. I was able to=20
//  link in some test=20
//  code compiled with MASM 6.1. However it does look like the=20
//  documentation is=20
//  more oriented to inline assembler. For portability, we could keep =
the=20
//  "assembler" files separate, and they could be implemented as=20
//  either pure=20
//  assembler or as one big inline file. All assembler, of course,=20
//  must use the flat=20
//  memory model.
// =20
//  I'm not so sure it would be all that hard to port some of the=20
//  more critical=20
//  assembler to Visual C/C++ 5. This compiler doesn't support=20
//  Windows 3.1, but=20
//  I'm not worried about supporting 16 bit Windows.
// =20
//  Tim
// =20
// =20
// =20
// =20
//  --------------------------------------------------------------
//  Thanks for using Fractdev, The Fractint Developer's Discussion List
//  Post Message:   [e-mail]
//  Get Commands:   [e-mail] "help"
//  Administrator:  [e-mail]
//  Unsubscribe:    [e-mail] "unsubscribe fractdev"
// =20


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: [e-mail] (F.J. Slijkerman)
Subject: Re: (fractdev) Visual C/C++ 5
Date: 11 Jun 1998 11:11:30 +0200 (MET DST)


> On the other hand, if we get help porting the assembler,  we could keep the
> integer math.

The assembler must not be "ported", it should be rewritten. The
difference between 16- and 32-bit assembler is just too much.

Regards,
Frederik.

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: <[e-mail]>
Subject: (fractdev) 3D Fractint
Date: 11 Jun 1998 06:23:43 EDT

Can anyone help me?

I am working on a physics project and have been using Fractint to make 2D
images for a proof.  What I need now is to be able to make a specific type of
3D image such that, on an x,y,z reference frame, a Mandelbrot set can be
generated on each of the three planes.
     y \   / z
   x ---X---
       /    \

Each plane x,yi; y,zi; and z,xi, would have a two dimensional Mandelbrot
fractal to make a three dimensional image.  The formulas that I use to
generate each MS contains a small variable so that each MS that in generated
is different from any other MS.  The reference frame is then rotated two
degrees and three new MS are generated to get a new image.  With multiple
images, a video can be made showing vibrent 3D rotating fractals.

Need to show only the 'front' of the image.  Where the MSs overlap, only the
portion of MSs that are in view can be seen.

Can anyone help?????


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: Michael Bell <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 11 Jun 1998 20:31:06 +0100

Tim Wegner wrote:
>
> kragen wrote:
>
> > I really like integer math a lot, especially if I'm going to run
> > xfractint on my 386-16.
>
> yes, but with every passing day a 386 is an unusual platform, and the old DOS
> fractint can always be used. My view is that these old platforms shouldn't block
> Fractint's progress.
>
> On the other hand, if we get help porting the assembler,  we could keep the
> integer math.
>
> Tim
>

There are other processor's for which integer math still has large
benefits (as far as I can tell from v19.6).  The K6 processor has a slow
FPU so integer math is better on this processor than the floating point
code (guess what processor I've got :-)).  This generally applies for
all non-Intel chips as far as I can tell, and there are a fair few of
them about.

Michael.

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 11 Jun 1998 15:43:17 -0400

Michael,

 - There are other processor's for which integer math still has large
 - benefits (as far as I can tell from v19.6).  The K6 processor has a slow
 - FPU so integer math is better on this processor than the floating point
 - code (guess what processor I've got :-)).  This generally applies for
 - all non-Intel chips as far as I can tell, and there are a fair few of
 - them about.

The AMD K6-2 has a fast FPU.  Apparently even AMD recognizes a weak spot
when they see one. :-)  Cyrix still hasn't figured it out yet, though,
more's the pity.

Damien M. Jones   \\
[e-mail]  \\  http://www.icd.com/tsd/ (temporary sanity designs)
                    \\  http://www.fractalus.com/ (fractals are my hobby)

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 11 Jun 1998 22:29:59 -0600

I wrote:

> > On the other hand, if we get help porting the assembler,  we could keep the
> > integer math.


And Frederik responded:

> The assembler must not be "ported", it should be rewritten. The
> difference between 16- and 32-bit assembler is just too much.

Ahh, the key word was "help", not "porting" <grin!>. If the "helper" wrote 100%
pure assembler from first principles, that's cool.

Seriously, talk of help is a bit premature, but I do hope at the appropriate time
in true Stone Soup spirit we get some talented volunteers.

Tim


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 11 Jun 1998 22:29:59 -0600

Michael wrote:

> There are other processor's for which integer math still has large
> benefits (as far as I can tell from v19.6).  The K6 processor has a slow
> FPU so integer math is better on this processor than the floating point
> code (guess what processor I've got :-)).  This generally applies for
> all non-Intel chips as far as I can tell, and there are a fair few of
> them about.

Ah so, but I have a float-only version I have already made. I know for certain
that I personally will not be writing any flat model  integer math code in
assembler. That doesn't mean it won't happen, it just means that it will depend
on someone who agrees it is important and who has the skills and time
steping forward.

Tim


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Tim Wegner" <[e-mail]>
Subject: RE: (fractdev) Visual C/C++ 5
Date: 11 Jun 1998 22:29:59 -0600

Peter asked:

> Wait, can you explain this to me?  The console mode programs made by Vis C++ 5
> are not DOS programs at all, as far as I know.  I don't think you can run a
> 32-bit console mode program under DOS, even with a 32-bit DOS extender.
>First off, DOS'll spit out the typical 'This program must be run under
> Windows' BS.

Right, but if you are running Win95 and create a DOS window (fullscreen or
not), a console application runs just like a DOS program, although internally it
accesses memory using Win95. Such programs would not run under "naked"
DOS. I run Fractint under Win95 anyway, as I expect many people do.

To run under true DOS, a DOS extender woul;d be needed such as that in
djgpp.

Tim




Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Tim Wegner" <[e-mail]>
Date: 11 Jun 1998 22:29:59 -0600

I'm taking off for a week, will be back June 23. This list doesn't have many
administration issues, just be aware that if anything crops up I won't be dealing
with it until I'm back.

It will help if list members unsubscribe themselves if their account is coming to
an end.

Carry on. I will read all the messages here when I get back (I can't promise the
same for the fractint list <g!>)

I am determined to pursue a 32 bit port of Fractint. We should also get the
current developer's fractint out the door as v. 20 soon after I get back.

Tim



Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Peter Gavin" <[e-mail]>
Subject: RE: (fractdev) Visual C/C++ 5
Date: 12 Jun 1998 04:56:46 -0400

Wow, I didn't realize console programs could do that!  But, umm, did you =
say you linked in the assembly?  It's 16bit, though, right?  So how'd =
that work?  (It'd be really cool if you would email me with your method =
of compiling it for win32, because that way I wouldnt have to reinvent =
the wheel, and I don't really have the time to figure that crud out.  My =
experience with assembly is very limited, unfortunately... <g> ) =20

Pete
<[e-mail]>

//  -----Original Message-----
//  From: [e-mail]
//  [mailto:[e-mail]]On Behalf Of Tim Wegner
//  Sent: Friday, June 12, 1998 12:30 AM
//  To: [e-mail]
//  Subject: RE: (fractdev) Visual C/C++ 5
// =20
// =20
//  Peter asked:
// =20
//  > Wait, can you explain this to me?  The console mode programs=20
//  made by Vis C++ 5
//  > are not DOS programs at all, as far as I know.  I don't think=20
//  you can run a=20
//  > 32-bit console mode program under DOS, even with a 32-bit DOS=20
//  extender. =20
//  >First off, DOS'll spit out the typical 'This program must be run =
under
//  > Windows' BS.
// =20
//  Right, but if you are running Win95 and create a DOS window=20
//  (fullscreen or=20
//  not), a console application runs just like a DOS program,=20
//  although internally it=20
//  accesses memory using Win95. Such programs would not run under =
"naked"=20
//  DOS. I run Fractint under Win95 anyway, as I expect many people do.
// =20
//  To run under true DOS, a DOS extender woul;d be needed such as that =
in=20
//  djgpp.
// =20
//  Tim
// =20
// =20
// =20
// =20
//  --------------------------------------------------------------
//  Thanks for using Fractdev, The Fractint Developer's Discussion List
//  Post Message:   [e-mail]
//  Get Commands:   [e-mail] "help"
//  Administrator:  [e-mail]
//  Unsubscribe:    [e-mail] "unsubscribe fractdev"
// =20


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Tim Wegner" <[e-mail]>
Subject: RE: (fractdev) Visual C/C++ 5
Date: 12 Jun 1998 08:07:19 -0600


> Wow, I didn't realize console programs could do that!  But, umm, did you say you linked in the assembly?
> It's 16bit, though, right?  So how'd that work?

No, the assembly MUST be 32 bit. The easiest way to see what the assemly
looks like is to create assembler listings of simple routines. This are
compilable in MASM.

> (It'd be really cool if you would email me with your method of compiling it for
>win32, because that way I wouldnt have to reinvent the wheel, and I don't
>really have the time to figure that crud out.  My experience with assembly is
>very limited, unfortunately... <g> )

I'd be happy to resume the conversation when I get back - out the door in a few
minutes!

Tim


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Paul N. Lee" <[e-mail]>
Subject: Re: (fractdev) Visual C/C++ 5
Date: 14 Jun 1998 11:48:20 -0500

Tim Wegner wrote:
>
> I am sure lots of folks know this stuff, but remember,
> I'm a beginner <grin!>
>
> I've been playing with Microsoft Visual C/C++ 5.0 ..........
>

NEW VISUAL C++ MAGAZINE TO LAUNCH IN JUNE

The Visual C++ Developer's Journal, a new magazine dedicated to the
needs of Visual C++ developers, will debut at Microsoft Tech Ed in
June.  Published 6 times per year, VCDJ will provide articles on Visual
C++, MFC, ATL, MTS, and COM as well as sample code, news, interviews,
and more.  VCDJ is free for qualified subscribers; an online
subscription request form is available at http://www.vcdj.com

P.N.L.
     Why do most folks hate cynics so much?
       Because we're almost always right.
http://www.fractalus.com/cgi-bin/theway?ring=fractals&id=43&go

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: Tim Gilman <[e-mail]>
Subject: (fractdev) performance diffs between X & DOS
Date: 17 Jun 1998 10:47:42 -0700

Can I get some performance marks from the XFract world?  I'm interested
in the basic 640x480x256 mandelbrot (default).  Using the XFract base on
my mac (PowerPC 604e @ 180MHz), and it takes a little over a minute to
generate this fractal.  What's the rest of the world look like?

-= Tim Gilman
[e-mail]
[e-mail]

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: Phil McRevis <[e-mail]>
Subject: Re: (fractdev) performance diffs between X & DOS
Date: 17 Jun 1998 14:58:09 -0600


In article <[e-mail]>,
    Tim Gilman <[e-mail]>  writes:
> Can I get some performance marks from the XFract world?  I'm interested
> in the basic 640x480x256 mandelbrot (default).  Using the XFract base on
> my mac (PowerPC 604e @ 180MHz), and it takes a little over a minute to
> generate this fractal.  What's the rest of the world look like?

The most significant performance differences I've seen between
xfractint and fractint are for formula type fractals.
--
http://www.xmission.com/~legalize/    Legalize Adulthood!
[e-mail]
``Ain't it funny that they all fire the pistol,     <URL: http://
  at the wrong end of the race?''--PDBT     www.eden.com/~thewho>

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Frederik Slijkerman" <[e-mail]>
Subject: Re: (fractdev) performance diffs between X & DOS
Date: 18 Jun 1998 11:03:56 +0200


> Can I get some performance marks from the XFract world?  I'm interested
> in the basic 640x480x256 mandelbrot (default).  Using the XFract base on
> my mac (PowerPC 604e @ 180MHz), and it takes a little over a minute to
> generate this fractal.  What's the rest of the world look like?

The DOS world looks much better. :)

With the following parameters: type=mandel periodicity=0 symmetry=none
maxiter=1000 passes=1 float=yes  I was able to generate a 640x480 image
on my Pentium 166MMX in 14.94 seconds. With the following formula

Mandelbrot{
  z = 0:
  z = sqr(z) + pixel
  |z| < 4
}

the same image took 1 minute 19.42 seconds. It seems to me a PowerPC could
do much better with a little bit of assembler...

Regards,
Frederik.


Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Christian Strik" <[e-mail]>
Subject: Re: (fractdev) performance diffs between X & DOS
Date: 18 Jun 1998 14:09:50 +0100

My compatriot Frederik wrote:

>
>With the following parameters: type=mandel periodicity=0 symmetry=none
>maxiter=1000 passes=1 float=yes  I was able to generate a 640x480 image
>on my Pentium 166MMX in 14.94 seconds. With the following formula
>
>Mandelbrot{
>  z = 0:
>  z = sqr(z) + pixel
>  |z| < 4
>}
>

Maxiter=1000? Isn't the default setting 150?

Christian




Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: DeRobertis <[e-mail]>
Subject: Re: (fractdev) performance diffs between X & DOS
Date: 18 Jun 1998 09:29:38 -0500

>Can I get some performance marks from the XFract world?  I'm interested
>in the basic 640x480x256 mandelbrot (default).  Using the XFract base on
>my mac (PowerPC 604e @ 180MHz), and it takes a little over a minute to
>generate this fractal.  What's the rest of the world look like?

Hmmm...MacFractInt is up and running? Can I see the sources?

I might be able to help with some assembly code, or get it to run under
MPW's MrC (which optimizes better than the CodeWarrior compiler)


--
VirtualLawyers say: There is a significant chance that this message may
contain alliteration, facetious remarks, spelling slips and gramatical
goofs. Should that be the case, all were intended or were computer errors.



Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: Tim Gilman <[e-mail]>
Subject: Re: (fractdev) performance diffs between X & DOS
Date: 18 Jun 1998 10:27:45 -0700

>Hmmm...MacFractInt is up and running? Can I see the sources?
>
>I might be able to help with some assembly code, or get it to run under
>MPW's MrC (which optimizes better than the CodeWarrior compiler)

Yeah, but its not quite ready for public consumption, or even developer
consumption.  Let me clean it up a lot and then I'll spread it around.

I've done some testing, and 40 of those "little over a minute" seconds
comes from drawing a pixel-at-a-time to a window.  XFract already has an
'update after each scan-line' strategy to get around this;  anyone have
other strategies that might shave off them seconds?

-= Tim Gilman
[e-mail]
[e-mail]

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractdev) performance diffs between X & DOS
Date: 18 Jun 1998 14:14:55 -0400

Tim Gilman,

 - I've done some testing, and 40 of those "little over a minute" seconds
 - comes from drawing a pixel-at-a-time to a window.  XFract already has an
 - 'update after each scan-line' strategy to get around this;  anyone have
 - other strategies that might shave off them seconds?

Here's a technique I use in JuliaSaver.  I keep track of the current
plotting position and the last drawn position.  Five times per second (this
is run on a Windows timer) I check to see how much has been drawn, redraw
only the portion that has changed, and update the "last-drawn" position.

By putting the screen update on a regular timer, the program will seem more
responsive than if it's line-by-line, since the amount of time a single
line takes varies quite a bit.  When plotting is slow, you will see
incremental plotting, perhaps even point-by-point for really slow fractals.
 But when plotting is fast, you don't waste time at every pixel by plotting
the point.  The five-redraws-per-second value was arbitrary, but it seems
to work out pretty well.  Then again, JuliaSaver isn't meant to be terribly
interactive.

To implement this in Windows I used multi-threading, putting the generator
on a background, non-GUI thread, and leaving the GUI thread free to respond
to messages (like timer and redraw messages).  I don't know how you'd go
about this under MacOS.

Damien M. Jones   \\
[e-mail]  \\  http://www.icd.com/tsd/ (temporary sanity designs)
                    \\  http://www.fractalus.com/ (fractals are my hobby)

Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"


-------------------------------------------------------------------------------

From: DeRobertis <[e-mail]>
Subject: Re: (fractdev) performance diffs between X & DOS
Date: 21 Jun 1998 01:45:00 -0500

>>Hmmm...MacFractInt is up and running? Can I see the sources?
>>
>>I might be able to help with some assembly code, or get it to run under
>>MPW's MrC (which optimizes better than the CodeWarrior compiler)
>
>Yeah, but its not quite ready for public consumption, or even developer
>consumption.  Let me clean it up a lot and then I'll spread it around.
>
>I've done some testing, and 40 of those "little over a minute" seconds
>comes from drawing a pixel-at-a-time to a window.  XFract already has an
>'update after each scan-line' strategy to get around this;  anyone have
>other strategies that might shave off them seconds?


Draw it pixel-by-pixel offscreen, in a 8-byte aligned buffer. Then perform
one move using a FP register to bring it onscreen.

Basicly, this results in one bus cycle per 8 bytes of pixel data (if you
can manage to keep that 8 bytes in the on-chip data cache)

You could also use a larger size (perhaps 32 bytes), but by using a one bus
cycle one, you might be able to avoid some stalls (that is the write to RAM
occurs during pixel calculation).

But whatever you do, don't even get near "SetCPixel."


--
VirtualLawyers say: There is a significant chance that this message may
contain alliteration, facetious remarks, spelling slips and gramatical
goofs. Should that be the case, all were intended or were computer errors.



Thanks for using Fractdev, The Fractint Developer's Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractdev"