ftp.xmission.com/2014.06.ftp.xmission.com.tar/ftp.xmission.com/pub/lists/fractint/archive/v01.n117

Internet Message Format | 1998-02-24 | 41KB

From: [e-mail] (fractint-digest)
To: [e-mail]
Subject: fractint-digest V1 #117
Reply-To: fractint-digest
Sender: [e-mail]
Errors-To: [e-mail]
Precedence: bulk


fractint-digest       Tuesday, February 24 1998       Volume 01 : Number 117




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

Date: Mon, 23 Feb 1998 22:24:23 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) Mac Port of FracInt?

DeRobertis wrote:

> I'm interesting in porting FracInt to the mac, but I have no clue about
> these silly .asm files!

If you start with Xfractint, you won't have any asm to worry about,
only C. You will only need to port the X user interface to the Mac.
This shouldn't be too hard.

If you are really serious about this let me know. We can help some,
even though we don't have macs. A lot of folks have asked about this.


Tim


- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Mon, 23 Feb 1998 22:24:23 -0600
From: "Tim Wegner" <[e-mail]>
Subject: Re: (fractint) A Win95 version of fractint

Damien said:

> There's nothing terribly special about the DOS environment that makes
> arbitrary precision possible.  I have no doubt it can be implemented in a
> Windows program.

Fractint's arbitrary precision is in both C and assembler. The C
should work immediately under Windows, with a speed penalty. It
should be possible to port the assembler to Windows, but it will take
special expertise.

Tim


- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Mon, 23 Feb 1998 22:29:53 -0800
From: Felix <[e-mail]>
Subject: Re: (fractint) ASCII and ye shall ...

Jay Hill wrote:

> Peter wrote:
> >Does anyone know of any really cool text-based fractal generating programs,
> preferably in the obfuscated C >language; I am only familiar with Jay Hill's
> signature program. There was a nice program posted some years ago to
> >sci.fractals.
>

 I have DLed  45k .exe  demo in 1995  (demoscene stuff ya know) called
"ascii.exe" that has several effects rendered completely in ascii, including a
spinning and zooming M-set.  It's pretty neat.
I still have it around and could post it if folks are interested.  I'm sure it
was done in C.

- --
| Andrew Duhan    |     Cereal is     |
| [e-mail]  |       g00d.       |
| http://chimera.acs.ttu.edu/~aduhan/ |



- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Mon, 23 Feb 1998 15:27:11 -0500
From: Sylvie Gallet <[e-mail]>
Subject: (fractint) Lee's "Fracta

Hi Lee,

>> I'm really surprised that you got the CD only 3 days after I mailed
>> it from New Mexico!

  Me too!  When the US mail works, it's really something<g>!

  Cheers,

        - Sylvie

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 00:27:46 -0500
From: "Brian E. Jones" <[e-mail]>
Subject: Re: (fractint) Lee's "Fractal Dimensions": a MUST!!!

Hi Lee,

I am really enjoying your Fractal Dimensions CD.

There are just too many (Great) images on this cd to try to view them all
from your web site.

For anyone interested in fractal art, I would HIGHLY recommend buying this
CD!!!  I guarantee that you'll be busy for a while.

Brian
- -
http://ourworld.compuserve.com/homepages/Brian_E_Jones
- -
[e-mail]
- -
[e-mail]




- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 00:53:56
From: "Lavondyss" <[e-mail]>
Subject: (fractint) map maker DLL

- ----- Forwarded Message Starts Here -----

From: Lavondyss
Date: Tue, 24 Feb 1998 00:50:54
To: [e-mail]
Subject: Re: (fractint) map maker

I also had that problem so I searched and WOALA!!..I found the site..I hope
it still works!!

http://babeard.simplenet.com/dll.html

Have fun!!

- -Lavondyss


- ----- Forwarded Message Ends Here -----




- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 00:50:54
From: "Lavondyss" <[e-mail]>
Subject: Re: (fractint) map maker

I also had that problem so I searched and WOALA!!..I found the site..I hope
it still works!!

http://babeard.simplenet.com/dll.html

Have fun!!

- -Lavondyss




- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Mon, 23 Feb 1998 09:54:12 -0600
From: "Justin A. Kolodziej" <[e-mail]>
Subject: Re: (fractint) Fractint problem w/ video card

[e-mail] wrote:
>
> In a message dated 98-02-22 19:38:34 EST, [e-mail] writes:
>
> << for the SciTech Display Doctor, a shareware utility that (among other
>  things) adds VESA support.  >>
>
>   The other things about it are that it makes Fractint all icky when you try
> to run it in 800x600x256 mode, and that it kills my monitor. :-P

Well, this is the best argument I've heard yet for producing a Win32
version of Fractint.  That way, it won't matter whether you're using a
Fire GL or an IBM 8514 (?), Fractint will work at your chosen
resolution.  Actually, you could pretty much make the resolution limited
by memory instead of the limits of your video card (i.e. make an
8000x6000 image and scroll around it to find the interesting parts).

Assuming, of course, that Windows likes your video card... ;-)

Justin K.
- --
Justin A. Kolodziej
I sense a great disturbance in the Source.
Justin Kolodziej is [e-mail]
Marquette University is www.mu.edu

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Mon, 23 Feb 1998 12:43:07 -0500
From: [e-mail]
Subject: (fractint) Arbitrary Precision pisses me off!

Sorry bout the subject, but it does. I was generating ƒberzoomed M-Sets at
320x200x256, c. 2,000,000,000 iterations, and suddenly it stopped and no matter
how long I waited it wouldn't draw another point. I tried again later, and at
17-decimal arbitrary precision mode, it stopped at the same place. I waited 24
hours, and it was still at the same point, with the time around 25:32:00. Is
there something wrong with my computer, (P133), with Windows 95. It does the
same thing in DOS mode too.
- --
Robert Hailman,
publisher of "Your Weekly Smile"

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Mon, 23 Feb 1998 12:43:07 -0500
From: [e-mail]
Subject: (fractint) Arbitrary Precision pisses me off!

Sorry bout the subject, but it does. I was generating ƒberzoomed M-Sets at
320x200x256, c. 2,000,000,000 iterations, and suddenly it stopped and no matter
how long I waited it wouldn't draw another point. I tried again later, and at
17-decimal arbitrary precision mode, it stopped at the same place. I waited 24
hours, and it was still at the same point, with the time around 25:32:00. Is
there something wrong with my computer, (P133), with Windows 95. It does the
same thing in DOS mode too.
- --
Robert Hailman,
publisher of "Your Weekly Smile"

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Mon, 23 Feb 1998 10:11:58 -0500
From: Gedeon Peteri <[e-mail]>
Subject: Re: (fractint) Two problems with Fractint

Morgan L Owens wrote:

> This is a problem I have as well; do you run Fractint through Windows (I
> don't mean Winfract)? Even though Fractint is running maximised, if
> something happens in a Windows window somewhere Fractint scruffs up.
> Attempting to save in these circumstances results in garbage, and sometimes
> Fractint chokes completely. I specift "textsafe=save", but this doesn't
> prevent the problem.

Exactly! I learned this the hard way when an image I have been generating for
about 3 hours was destroyed when System Agent kicked in to do a scandisk! Now I
close all windows and disable System Agent before I run Fractint. The strip of
junk across the top of the image I mentioned, however, is independent of Windows
interference. It occurs when I recall certain images and check the <z> screen.
Upon return to the image the strip of junk appears. Recalling the image now for
the second time always allows going back again to <z> and subsequently back to the
image without this time showing any junk. It does not matter whether I make any
changes on the <z> screen or not. The only difference is that if I do make some
changes, instead of starting to generate the new image on a clean screen, the
recalled image remains on the screen but all broken up into alternating strips of
the image and blank space, and the new image begins to generate atop this mess,
filling in the blank spaces but leaving the strips of the old image intact.
Moreover, I can usually tell whether this is going to happen as soon as I hit <z>
for the first time. Instead of the <z> screen popping up without hesitation, the
screen flickers for a fraction of a second, meanwhile showing a cursor at the top
left corner.  When I see this happening, I don't even bother to proceed until I
recalled the image once more; a procedure which ALWAYS clears up the problem. It
needs to be added that this problem does not occur with every image, only with
some, though I have not been able to notice any regularity allowing me to state
unequivocally which ones. I can only say that it appears to affect formula types
more than built in ones, but this is by no means an iron-clad rule.

I have had this problem for a long time and learned to live with it since the
remedy is simple, straightforward, and works every time. I mentioned it lately in
connection with another problem I have been having running Kerry Mitchell's
formulas, thinking that the two may be related with one another. This turned out
not to be the case, however. The latter problem disappeared when I followed
Sylvie's and Jay's suggestion about using the command periodicity=0.

Gedeon


- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 05:23:31 -0600
From: "Paul N. Lee" <[e-mail]>
Subject: Re: (fractint) a Mac version?

Robi Hamilton wrote:
>
> A friend with a Mac has asked if a program similar
> to Fractint exists for use with his computer system.
> Anyone have answers or suggestions?
>

There are many fractal programs for the MAC.  I can't personally say if
they are similar to FractInt, but you might try going to the following:

       http://spanky.triumf.ca/pub/fractals/programs/MAC/

I have a few links listed at my website that contain descriptions:

       http://home.att.net/~Paul.N.Lee/PNL_Fract_Sftw.html

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 10:42:25 -0500
From: Gedeon Peteri <[e-mail]>
Subject: (fractint) Formula in par file

There are evident advantages of the inclusion of formulas within par
files with the prefix frm: as outlined by Jay and others in recent
posts. But there appears to be a major disadvantage as well, which I
discovered only this morning. If one saves a partially generated image
and returns to it later  to finish it (after having in the meantime
exited Fractint), Fractint does not find the formula, it flashes the
message "Could not open file where formula located," or else finds
another formula with the same name elsewhere if such a duplicate formula
exists in the formula folder; a situation which the inclusion of frm: in
the par file was designed to avoid in the first place. Is there a remedy
for this problem?

Gedeon


- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 08:16:20 -0800
From: "Jay Hill" <[e-mail]>
Subject: Re: (fractint) Formula in par file

Hi Gedeon,

> There are evident advantages of the inclusion of formulas within par
> files with the prefix frm: as outlined by Jay and others in recent
> posts. But there appears to be a major disadvantage as well, which I
> discovered only this morning. If one saves a partially generated image
> and returns to it later  to finish it (after having in the meantime
> exited Fractint), Fractint does not find the formula, it flashes the
> message "Could not open file where formula located," or else finds
> another formula with the same name elsewhere if such a duplicate formula
> exists in the formula folder; a situation which the inclusion of frm: in
> the par file was designed to avoid in the first place. Is there a remedy
> for this problem?

Solution: Enter the par file in question, start the image anew. Then hit
r to read in the partly done image. It should continue, since it is now
pointed to the par file.

Jay


- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 11:14:58 -0500
From: "Peter Gavin" <[e-mail]>
Subject: Re: (fractint) Arbitrary Precision pisses me off!

2 BILLION Iterations!!??  That explains itself! :)
At 65536 iterations, using arbitrary precision, it sometimes takes one ho=
ur
per pixel, on my P60.  But 2 billion would take a CRAY! :)

- -----Original Message-----
From: [e-mail] <[e-mail]>
To: [e-mail] <[e-mail]>
Date: Tuesday, February 24, 1998 5:29 AM
Subject: (fractint) Arbitrary Precision pisses me off!


>Sorry bout the subject, but it does. I was generating =9Fberzoomed M-Set=
s at
>320x200x256, c. 2,000,000,000 iterations, and suddenly it stopped and no
matter
>how long I waited it wouldn't draw another point. I tried again later, a=
nd
at
>17-decimal arbitrary precision mode, it stopped at the same place. I wai=
ted
24
>hours, and it was still at the same point, with the time around 25:32:00.
Is
>there something wrong with my computer, (P133), with Windows 95. It does
the
>same thing in DOS mode too.
>--
>Robert Hailman,
>publisher of "Your Weekly Smile"
>
>-
>------------------------------------------------------------
>Thanks for using Fractint, The Fractals and Fractint Discussion List
>Post Message:   [e-mail]
>Get Commands:   [e-mail] "help"
>Administrator:  [e-mail]
>Unsubscribe:    [e-mail] "unsubscribe fractint"


- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 11:18:43 EST
From: [e-mail]
Subject: Re: (fractint) map maker

In a message dated 98-02-23 17:35:52 EST, you write:

<< I downloaded mapmaker but I don't seem to have the BWCC.dll file so I was
 wondering if somebody could send it to me or tell me where I could find it.

 thanks

 Also thanks for all the advice given to me it has really helped.

 lloyd >>

Woops!
I will include bwcc.dll with the zip file when I do web page maintenance
tonight.
Ron Barnett

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 11:40:51 EST
From: [e-mail]
Subject: Re: (fractint) Lee's "Fractal Dimensions": a MUST!!!

Hi Bob,

finally it had work to describe to the mailing list.
Welcome --Jo--

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 10:54:33 -0600
From: "Justin A. Kolodziej" <[e-mail]>
Subject: Re: (fractint) Mac Port of FracInt?

Peter Gavin wrote:
>
> Someone tell me if I'm wrong, but I believe anything written in asm is also
> in C... ?
> I dunno, I really just glanced through it for a coupla minutes...
Yes and no.  Assembler code can be put in a C program, but still is
assembler code.  And there's a problem.  There are going to be a LOT of
differences between Mac assembler and Intel assembler just because they
are different processors with different instruction sets.

A different route to take would be to take the XFractint code with no
assembler code, replace all the ncurses and X-Windows specific stuff
with things that work on a Mac, and compile that.  Don't know if it
would be less or more work, though ;-)

Justin K.
- --
Justin A. Kolodziej
I sense a great disturbance in the Source.
Justin Kolodziej is [e-mail]
Marquette University is www.mu.edu

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 12:30:13 EST
From: [e-mail]
Subject: Re: (fractint) map maker

In a message dated 98-02-23 19:32:53 EST, you write:

<<  I create my own palettes and when I re-use one of them, I often edit some
 of the colors.  The color map is the most important and the most difficult
 part of the work and it requires time and patience.
  >>

Amen to that Sylvie. To many really good images become just average due to
poor coloring.

- --Bob Carr--(Ocala, FL)

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 12:56:04 EST
From: [e-mail]
Subject: Re: (fractint) map maker

Ron,

Is there a way you could alter your map making program so that it could
generate map files without the user putting in colors? Also if that is
possible to put something in the program that a user could use for example:
black in the center of the color instead of at the end of the color. Also be
able to shade 2 or 3 or 4 etc. of the vertical color lines (that is, darker or
lighter)? Just wondering.

- --Bob Carr--

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 12:36:05 -0600
From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractint) fractal hardware engine

Paul,

 - Won't the serial port I/O speed become a serious bottleneck? It must be
 - orders of magnitude slower than a PC bus. I wouldn't think all the
 - megahertz in the world would make much difference if your FPU communicated
 - to the CPU/memory at only 57600 baud...

Depends on how much you have to communicate.  If you can simply give the
DSP a starting point, and have it generate a whole slew of points
unattended, your communication speed becomes less important.

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

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 12:45:07 -0600
From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractint) Mac Port of FracInt?

DeRobertis,

 - I'm interesting in porting FracInt to the mac, but I have no clue about
 - these silly .asm files!

Those "silly" .asm files are what makes FractInt fast on the PC. :)  If
you're going to port to the Mac, you need to either (a) use the Xfractint
code, which doesn't use any of the assembly, or (b) write PowerPC versions
of the various assembly routines.

The biggest problem you will have porting to the Mac is the switch to an
event-driven programming model.  In DOS, the program is doing things,
constantly checking to see if there's input.  For a windowed environment,
typically you must wait for the OS to tell you something has happened.
It's a different approach, and sometimes switching a program from one to
the other isn't easy!

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

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 12:40:02 -0600
From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractint) Fractint problem w/ video card

Justin,

 - Well, this is the best argument I've heard yet for producing a Win32
 - version of Fractint.

There's no argument--pretty much everyone agrees that at some point
FractInt will need a 32-bit (preferably Windows) version.  The biggest
problem right now is, I presume, the "32-bit" portion of that.  I'd imagine
a 32-bit DOS version (using DJGPP or something similar) would be the first
step, and then writing a Windows interface around that.  Which means it's
still a good ways off.

Until you start looking at the FractInt code, you don't really get a
feeling for what a *huge* program it is.

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

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 12:33:32 -0600
From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractint) fractal hardware engine

Rich,

Sorry it took me so long to get to this.  Trying to catch up on all my
forgotten e-mail messages today. :)

 - Having been looking into the most cost-effective way to get a "fractal
 - accelerator" for PCs, I think I've found something that might get you
 - some acceleration and also be affordable.
 -
 - The question is: how much faster is a TI DSP320Cxx at computing fractals
 - than a pentium or P-II?

Not faster at all.  Bear in mind this: a Pentium-200 can crank out about
fifteen million Mandelbrot iterations per second, peak; that means if
you're using a 50MHz TMS320C5x DSP, you have to be able to finish an
iteration in about three cycles just to keep up.  Also remember that the
Pentium uses 64-bit precision internally (not counting the 16 bits of
exponent, since it's floating-point) and that the 320C5x DSPs only use
16-bit precision.  Even in parallel, these are very inconvenient to use for
fractal work.

Using hardware to accelerate fractals will probably not be possible with
this particular family of DSPs.  An associate suggests you look into
Motorola DSPs, which can offer a lot more performance.  The best solution
might be to use a large FPGA, but putting a 64x64 multiply into it will
take a lot of gates.  (And if you don't, it's not worth the effort!)

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

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 12:47:55 -0600
From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractint) A Win95 version of fractint

Tim,

 - Fractint's arbitrary precision is in both C and assembler. The C
 - should work immediately under Windows, with a speed penalty. It
 - should be possible to port the assembler to Windows, but it will take
 - special expertise.

Special expertise in the form of knowing how to write 32-bit assembly code
in the Windows environment; I know of several programmers who can handle
that.  The biggest problem is the 32-bit part, not the Windows part, as far
as arbitrary precision is concerned.  (For the rest of FractInt, though,
the Windows part is DEFINITELY a big issue.)

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

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 13:20:52
From: "Nhan Nguyen" <[e-mail]>
Subject: Re: (fractint) map maker

Ron...thanks for the color mapmaker. I love it!





- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 13:22:42
From: "Nhan Nguyen" <[e-mail]>
Subject: Re: (fractint) map maker

On Tue, 24 Feb 1998 12:30:13 EST,
[e-mail] wrote...
>In a message dated 98-02-23 19:32:53 EST, you write:
>
><<  I create my own palettes and when I re-use one of them, I often edit
>some
> of the colors.  The color map is the most important and the most
difficult
> part of the work and it requires time and patience.
>  >>
>
>Amen to that Sylvie. To many really good images become just average due to
>poor coloring.
>
>--Bob Carr--(Ocala, FL)
>
>-
>------------------------------------------------------------
>Thanks for using Fractint, The Fractals and Fractint Discussion List
>Post Message:   [e-mail]
>Get Commands:   [e-mail] "help"
>Administrator:  [e-mail]
>Unsubscribe:    [e-mail] "unsubscribe fractint"
>
>
I would also have to agree. I've printed some images using my Epson and
they turned out very nice due to good color schemes. I love them on the
high gloss papers!..You've gotta try it!

- -Lav




- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 12:30:59 -0700
From: Rich Thomson <[e-mail]>
Subject: Re: (fractint) fractal hardware engine

In article <[e-mail]> ,
    "Damien M. Jones" <[e-mail]>  writes:
> Not faster at all.  Bear in mind this: a Pentium-200 can crank out about
> fifteen million Mandelbrot iterations per second, peak;

Damien, this is exactly the sort of information I was looking for.
This is for iterations that fit within the 80-bit floating point
format of the device, correct?  What is the approximate peak
performance when 1024-bit mantissas are concerned?  (i.e. arbitrary
precision where you're using ~32 words for the mantissa)  (Also, I
don't have a P200, I have a P90, which is about 1/2 as fast, but still
quite competitive to a 320Cxx.)

> Using hardware to accelerate fractals will probably not be possible with
> this particular family of DSPs.  An associate suggests you look into
> Motorola DSPs, which can offer a lot more performance.  The best solution
> might be to use a large FPGA, but putting a 64x64 multiply into it will
> take a lot of gates.  (And if you don't, it's not worth the effort!)

The main reason I considered the TMS320Cxx chips is because the DSK
kits are available for $100.  I assume by the Motorola devices, you're
referring to the 56K device?  Looking around Motorola's web site, they
babble on quite a bit about "low cost evaluation boards" but don't
bother to mention the price anywhere. :(  (Pet peeve: web sites that
tell you everything about a product but the single most important
thing: its price!)

Your point about buying beefy enough hardware to make it competitive
to a P200 is well taken.  I've mentioned this before as well.  Its not
enough just to "make some hardware"; you have to make hardware that
will outperform the general-purpose CPU solution otherwise you didn't
win anything.  The main reason I mention these DSK/evaluation boards
is that they generally have everything one would need for such an
application and are inexpensive.  However, there isn't any point to
using them unless they outperform the general-purpose CPU.  Unless
you just want to build hardware for the educational experience in
which case there also isn't much point to using one of these
DSKs/evaluation boards.
- --
                                               Rich Thomson
                                               [e-mail]

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 13:53:19 -0600
From: "Damien M. Jones" <[e-mail]>
Subject: Re: (fractint) TECH: fractal hardware engine

Rich,

Perhaps we should take this over to the FracProgrammer list?

 - > Not faster at all.  Bear in mind this: a Pentium-200 can crank out about
 - > fifteen million Mandelbrot iterations per second, peak;
 -
 - This is for iterations that fit within the 80-bit floating point
 - format of the device, correct?

Correct, this is all internal, using FPU registers.  And that's with the
insanely optimized code I posted to the FracProgrammer list a while back.

 - What is the approximate peak performance when 1024-bit mantissas are
 - concerned?  (i.e. arbitrary precision where you're using ~32 words for
 - the mantissa)

Obviously not as fast. :)  Arbitrary precision presents something of a
problem for the Pentium, where 32x32 integer multiplies (which typically
make up the core of arbitrary precision code) are nine cycles instead of
one for a 64x64 FPU multiply.  This has been discussed on sci.fractals
before.  It may be possible to reap some benefit from the Pentium FPU by
using it for 24x24 multiplies, and performing accumulations inside the FPU,
but I have not written code to test this.  Even so, the problem vanishes on
the Pentium Pro and Pentium II, where the 32x32 integer multiply is one
cycle (throughput; three-cycle latency).

What this means is, on a PPro or P-II, arbitrary precision gets a huge
speed boost over the Pentium.  Doing a 1024x1024 multiply (with a 2048-bit
result) involves 1024 multiplications, if your core multiply is 32x32.  On
the TI DSP, your core multiply is 16-bit, so it takes 4096 multiplies--four
times as many, and at a lower clock speed to boot.  The scheduling
capabilities of PPro and P-II processors make it capable of doing much in
parallel, similar to how DSPs do, so that advantage is mostly nullified.

The only way to beat a general-purpose CPU like the Pentium or PPro/P-II is
to use a bigger multiply, or do work in parallel.  If you have a 64x64
hardware multiply that produces a result in one cycle (AFAIK the Pentium
has a 64x32 engine that takes two cycles) then an M-set iteration will take
three cycles minimum if you can do the adds/subtracts/transfers in
parallel.  If you don't have to check for bailout, you can do it in two.
Applying similar loop-unrolling tricks as the Pentium code does, this would
let you do a block of eight iterations in just 17 cycles.  If you can run
such a chip at 50MHz (not a ridiculous number for an FPGA) then that's 50%
faster than a 200MHz Pentium, roughly.  And that leaves the CPU open for
doing other work; if it is generating points as well, this means adding a
board with such a device to your 200MHz Pentium PC would more than double
its performance on M-set fractals.  And if you can parallelize the board so
more chips can be added, it gets a lot faster still.

 - The main reason I considered the TMS320Cxx chips is because the DSK
 - kits are available for $100.  I assume by the Motorola devices, you're
 - referring to the 56K device?

Actually, no, I wasn't referring specifically to the 56K device.  I don't
have references for Motorola DSPs handy so I can't tell you exactly which
one--I was just passing on a side remark from a friend.  I do have
reference materials for the TMS320C5x DSP, though.

DSK kits being $100 is nice, but only if the DSP is fast enough to make a
difference. :)

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

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 16:03:42 -0500
From: Atkinson <[e-mail]>
Subject: Re: (fractint) map maker DLL

Hi, those who sent it thanks for the dll.  Those who didn't, don't send it.
Thank you.

lloyd


- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 19:18:14 -0800
From: Peter Jakubowicz <[e-mail]>
Subject: (fractint) TwinDragon

Can somebody pls tell me how I can generate the TwinDragon Julia set cited
in Knuth , vol. 2, using Fractint, or point me to where it is discussed in
the docs, if it is?
Thank you


- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

Date: Tue, 24 Feb 1998 19:20:10 -0500
From: Sylvie Gallet <[e-mail]>
Subject: Re: (fractint) map maker

>> I would also have to agree. I've printed some images using my Epson
>> and they turned out very nice due to good color schemes. I love them
>> on the high gloss papers!..You've gotta try it!

  I've got some stunning images with my Epson Stylus Color 600, at 720 dp=
i
on photo quality glossy film.

        - Sylvie

- -
- ------------------------------------------------------------
Thanks for using Fractint, The Fractals and Fractint Discussion List
Post Message:   [e-mail]
Get Commands:   [e-mail] "help"
Administrator:  [e-mail]
Unsubscribe:    [e-mail] "unsubscribe fractint"

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

End of fractint-digest V1 #117
******************************