Found your Mandelbrot program while browsing around Wikipedia - very impressive and obviously fast! I was quite pleased to see the Normalized Iteration algorithm added to the latest version. For a future versions, it would be awesome to be able to enable multi-sampling/anti-aliasing or even a record to .avi feature to make zoom-in/out videos.
Thanks for viewing and once again great job with QuickMan.
Taylor
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks for the compliments. The .avi zoom is definitely on my to-do list. I also want to implement saving of arbitrarily large images (for making huge printouts and such). Once that's done the AA/multisampling should follow naturally.
Now if only I could find some time to work on it...
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Great work! I hardly beleive that it runs so fast, really. And this is only one thing. The other is the user interface which makes it easy and convenient to surf the Mandelbrot landscape!
Now let me suggest an area to improve: precision.
The first step would be the 10bytes Extended arithmetic. Then -it's a big deal I know- further precision limits can be demolished by using more bits for representing the mantissa. It can be arbitrary but there are ways to insert some steps between extended and full arbitrary precision. (for example 2x64 bit, 4x64 bit etc (I don't know for sure if it's possible to do that with two or four double floats but it surely can be done with int64's by emulating floating point representation. The question is which is the faster or easier to implement?))
Ok, this is hard to make it work but I think it worth it. I u have no time to do that partial solutions are also acceptable. For examople the extended precision coz it's supported by hardware on x87. :P
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I've been meaning to add x87 80-bit precision mode, but I discovered that the Microsoft compiler I've been using doesn't support 80-bit floating point types, even in assembly language. It appears the only option with Microsoft is to __emit machine code directly into the source. :)
So this is a good reason to change compilers, maybe to MinGW. It seems like it would be a big job though, so I need to find some time to work on it (if anyone wants to take a shot at it, feel free...)
I believe there are arbitrary-precision libraries out there that could be added too. It should be pretty easy to add those, but of course the speed would drop drastically.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Found your Mandelbrot program while browsing around Wikipedia - very impressive and obviously fast! I was quite pleased to see the Normalized Iteration algorithm added to the latest version. For a future versions, it would be awesome to be able to enable multi-sampling/anti-aliasing or even a record to .avi feature to make zoom-in/out videos.
Thanks for viewing and once again great job with QuickMan.
Taylor
Thanks for the compliments. The .avi zoom is definitely on my to-do list. I also want to implement saving of arbitrarily large images (for making huge printouts and such). Once that's done the AA/multisampling should follow naturally.
Now if only I could find some time to work on it...
Great work! I hardly beleive that it runs so fast, really. And this is only one thing. The other is the user interface which makes it easy and convenient to surf the Mandelbrot landscape!
Now let me suggest an area to improve: precision.
The first step would be the 10bytes Extended arithmetic. Then -it's a big deal I know- further precision limits can be demolished by using more bits for representing the mantissa. It can be arbitrary but there are ways to insert some steps between extended and full arbitrary precision. (for example 2x64 bit, 4x64 bit etc (I don't know for sure if it's possible to do that with two or four double floats but it surely can be done with int64's by emulating floating point representation. The question is which is the faster or easier to implement?))
Ok, this is hard to make it work but I think it worth it. I u have no time to do that partial solutions are also acceptable. For examople the extended precision coz it's supported by hardware on x87. :P
I've been meaning to add x87 80-bit precision mode, but I discovered that the Microsoft compiler I've been using doesn't support 80-bit floating point types, even in assembly language. It appears the only option with Microsoft is to __emit machine code directly into the source. :)
So this is a good reason to change compilers, maybe to MinGW. It seems like it would be a big job though, so I need to find some time to work on it (if anyone wants to take a shot at it, feel free...)
I believe there are arbitrary-precision libraries out there that could be added too. It should be pretty easy to add those, but of course the speed would drop drastically.