Lee say: ""I think I can explain. I don't believe that there is such an arbitrary precision bug in later versions of Fractint. In your deepest zoom at a magnification of e^89, and using a maxiter of over 2 billion, the image does "freeze" once the showdot reaches the edge of the minibrot. But, unless you had a very powerful machine (much faster than your P3-733), it would be expected to take a very long time per pixel, and thus appear to freeze up."" No. That is the first thing I thought of. But if that was it ALL versions of fractint would show the same behavior. They do not. It does take longer to get through the lake area, but if you hit TAB a couple minutes apart, you will see different number for row/column etc showing it is still working, albiet slowly. When I say it freezes I mean it locks up on one pixel and doesn't procede any further - same row/column for hours. No problem with the E+55 minibrot (the second one) in the video, but on the final E+89 brot, shortly after the tiny black spot in the center begins to resolve itself, the freeze happens and I switch to Fractint95 and it works. I call it fractint95 because the date on the file is 1995, I think it is version 19.3 or 19.6. The freeze happens at the boundary of the minibrot, about half way through it. Exactly why or what is going on I do not know, I suspect an endless loop - it never bails for some reason. If I save the image at this point and reload it into fractint95 it continues normally. As to the depth at which it occurs, it is between E+55 and E+89 obviously, but I recall this happening many years ago also before I started making animations, just deep stills. I have all fractint versions since the "fractal creations" CD, which I recall is version 16. They do not all behave the same! Some work for some things and others don't. You may find this out if you do a lot of experimenting. I wish I could get the old versions just to check out - starting with version 1.0 in 1987! Anyone know where / if these are available? JoTz
JoTz, I have tested your e^89 par, and it does exactly as you say. It locks up on a single pixel for hours. I think Jonathan and Tim should look at this. I'll email them my current image so that they don't have to start it over. Lee ------------------------------- Original Message --------------------- Lee say: "I think I can explain. I don't believe that there is such an arbitrary precision bug in later versions of Fractint. In your deepest zoom at a magnification of e^89, and using a maxiter of over 2 billion, the image does "freeze" once the showdot reaches the edge of the minibrot. But, unless you had a very powerful machine (much faster than your P3-733), it would be expected to take a very long time per pixel, and thus appear to freeze up." No. That is the first thing I thought of. But if that was it ALL versions of fractint would show the same behavior. They do not. It does take longer to get through the lake area, but if you hit TAB a couple minutes apart, you will see different number for row/column etc. showing it is still working, albeit slowly. When I say it freezes I mean it locks up on one pixel and doesn't proceed any further - same row/column for hours. No problem with the E+55 minibrot (the second one) in the video, but on the final E+89 brot, shortly after the tiny black spot in the center begins to resolve itself, the freeze happens and I switch to Fractint95 and it works. I call it fractint95 because the date on the file is 1995, I think it is version 19.3 or 19.6. The freeze happens at the boundary of the minibrot, about half way through it. Exactly why or what is going on I do not know, I suspect an endless loop - it never bails for some reason. If I save the image at this point and reload it into fractint95 it continues normally. As to the depth at which it occurs, it is between E+55 and E+89 obviously, but I recall this happening many years ago also before I started making animations, just deep stills. I have all fractint versions since the "fractal creations" CD, which I recall is version 16. They do not all behave the same! Some work for some things and others don't. You may find this out if you do a lot of experimenting. I wish I could get the old versions just to check out - starting with version 1.0 in 1987! Anyone know where / if these are available? JoTz
jackoftradez wrote:
I wish I could get the old versions just to check out - starting with version 1.0 in 1987! Anyone know where / if these are available?
I have all these and have known for years that I needed to make them available and have promised it several times. I have a whole lot of things on my plate right now (as always - it's good news, I have a very full and fulfilling life :-), but I shall do my best to figure out a way to get my fractint collection out. I may have to write some batch files to unpack and recompile source. But if I just released a set of official versions, that would narrow it down to an official release whn it broke that would give us a start, and not be too time consuming for me. Is there anyone who could host a CD ISO image? We can't do that at fractint.org because of the download bandwidth. Having an ISO image online seems to attract many casual downloaders who have huge bandwidth available to them. I discovered this before when I posted an ISO of the Fractal Creations 2nd Edition CD. The nature of the deep zoom hang that's being discussed is that it will be very hard to diagnose without some kind of binary search through the various versions. But with those versions, since I have not just release versions but developer patch versions, we have a shot - if we can locate the exact version where the problem occurs, the number of lines of code to inspect to find the bug is relatively small. Tim
Tim,
I have all these and have known for years that I needed to make them available and have promised it several times. I have a whole lot of things on my plate right now (as always - it's good news, I have a very full and fulfilling life :-), but I shall do my best to figure out a way to get my fractint collection out. I may have to write some batch files to unpack and recompile source. But if I just released a set of official versions, that would narrow it down to an official release whn it broke that would give us a start, and not be too time consuming for me.
I have executables for 19.0 and 19.6. Version 19.0 works, and 19.6 doesn't. Looking at the dates on the diff files indicates that versions 19.0, 19.1 and 19.2 were released in 1995. Jonathan
Jonathan wrote:
I have executables for 19.0 and 19.6. Version 19.0 works, and 19.6 doesn't. Looking at the dates on the diff files indicates that versions 19.0, 19.1 and 19.2 were released in 1995.
See: http://www.fractint.org/ftp/archive/ I uploaded the official releases from 7.30 to 19.60. By patch-by patch collection starts from 19.6 to the present, or at least up to the start of CVS. I haven't uploaded the patches. Since you say the problem is between 19.0 and 19.6, the ones I have uploaded is the best I can do. Tim
Jonathan wrote:
The problem occurs between versions 19.2 and 19.3. My guess would > be version 19.21 patch 5. This patch rearranged the bignum routines.
Ah so. Upon further review, I do have all the diffs in that series. I just hadn't created source directories and compiled them. Do you have the diffs? While I do have other projects this long weekend, one of which you know about, I think I can come up with the exact changes. Tim
Tim,
Ah so. Upon further review, I do have all the diffs in that series. I just hadn't created source directories and compiled them. Do you have the diffs? While I do have other projects this long weekend, one of which you know about, I think I can come up with the exact changes.
Yes, I have the diffs. This particular patch is a collection of the changed bignum files plus a diff file. Jonathan
On Thu, 25 May 2006 at 22:17:42, Tim Wegner wrote:
See: http://www.fractint.org/ftp/archive/ I uploaded the official releases from 7.30 to 19.60.
On Fri, 26 May 2006 at 19:37:34, Tim Wegner replied:
The whole history of the official versions is now online.
For those wishing to acquire some of the missing versions, I will upload what I have available of the other ones to my site. You will be able to get v2.10, v4.00, v5.10, v7.00, etc..... On Fri, 26 May 2006 at 19:37:34, Tim Wegner replied:
On Thu, 25 May 2006 at 23:54:16, Paul N. wrote:
If you will take the time to burn the CD and mail it to me, then I will upload it to my website with the other items available for download.
Thanks for the offer. I may take you up on it. However I'm beginning to question the wisdom of a downloadable CD.
Yes, I know... it's like trying to get picture of Jim's fractal cats.
Last night it was effortless to put up all the official releases I have. Putting up many small files protects bandwidth better - people download what they need or are interested in.
The how about completing the set of downloads so that it has all possible versions that have ever been made publicly available?? Sincerely, P.N.L. ------------------------------------------------- http://home.att.net/~Paul.N.Lee/PNL_Fractals.html http://www.Nahee.com/Fractals/
For those wishing to acquire some of the missing versions, I will upload what I have available of the other ones to my site. You will be able to get v2.10, v4.00, v5.10, v7.00, etc.....
By all means, I would like a complete set, so I'm interested if you have some I don't. I may have some earlier ones also, I need to study my hard drive a bit more carefully. Between us maybe we can get close to a complete set.
The how about completing the set of downloads so that it has all possible versions that have ever been made publicly available??
I think it's close to that now except the very latest ones. Of course if you are referring to the last few years, there are many more. I have a very complete collection of those. There's actually a lot of value in having more than the "official releases" available. The development team has context diffs for each patch. The value of all these incremental versions is that using a binary search strategy one can narrow down the exact patch in which some future broke, as in the deep zoom bug issue. That requires building each source directory and making the executable. By making the whole set of executables available, the user community can help track this sort of thing down. Tim
Tim Wegner wrote:
Paul N. wrote:
For those wishing to acquire some of the missing versions, I will upload what I have available of the other ones to my site. You will be able to get v2.10, v4.00, v5.10, v7.00, etc.....
.....I'm interested if you have some I don't. I may have some earlier ones also, I need to study my hard drive a bit more carefully. Between us maybe we can get close to a complete set.
Besides those few mentioned above, I have a few more to fill in some blanks between v10.00 and v17.10, bringing the additional to a total of about ten. Hopefully you could find more in your own archives, other than what you have already uploaded, and can add those to the current FractInt.Org website archive.
Then how about completing the set of downloads so that it has all possible versions that have ever been made publicly available??
I think it's close to that now except the very latest ones. Of course if you are referring to the last few years, there are many more. I have a very complete collection of those.
I too have a fairly complete set from v19.00 upwards, including DIF files over the past seven years. But I was mainly referring to the really old versions. I know I am missing at least eleven of the older versions. Would love to get a copy of the first version. Looking forward to any others that you can find to complete the set. Sincerely, P.N.L. ------------------------------------------------- http://home.att.net/~Paul.N.Lee/PNL_Fractals.html http://www.Nahee.com/Fractals/
Tim Wegner wrote:
Is there anyone who could host a CD ISO image? We can't do that at fractint.org because of the download bandwidth.
If you will take the time to burn the CD and mail it to me, then I will upload it to my website with the other items available for download. Have been asking for this for a few years now, hopefully it will happen this time. :-) Sincerely, P.N.L. ------------------------------------------------- http://home.att.net/~Paul.N.Lee/PNL_Fractals.html http://www.Nahee.com/Fractals/
Paul N. wrote:
If you will take the time to burn the CD and mail it to me, then I > will upload it to my website with the other items available for download.
Thanks for the offer. I may take you up on it. However I'm beginning to question the wisdom of a downloadable CD. Last night it was effortless to put up all the official releases I have. Putting up many small files protects bandwidth better - people download what they need or are interested in. Putting up a full CD triggers downloads by people only casually interested who neverless download many megabytes.
Have been asking for this for a few years now, hopefully it will happen this time. :-)
The whole history of the official versions is now online. Tim
participants (5)
-
jackoftradez＠comcast.net -
Jonathan Osuch -
Lee H. Skinner -
Paul N. Lee -
Tim Wegner