Is this something that needs to be preserved? It currently depends on a bunch of assembly code and in the XFRACT code an 'MP' is just a double. I'm not sure what this is buying us, if anything, in a world where every x86 machine has an FPU. -- "The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download <http://www.xmission.com/~legalize/book/download/index.html> Legalize Adulthood! <http://blogs.xmission.com/legalize/>
Rich asked (about MP_MATH)
Is this something that needs to be preserved?
No
I'm not sure what this is buying us, if anything, in a world where every x86 machine has an FPU.
Agreed. What MP_MATH bought us at one time was increased dynamic range that ran reasonably on FPU-less machines. Tim
In article <46814465.30576.887B7@twegner.swbell.net>, "Tim Wegner" <twegner@swbell.net> writes:
What MP_MATH bought us at one time was increased dynamic range that ran reasonably on FPU-less machines.
Thanks Tim, I'll kill it tonight. One less code path in all that math stuff would be great. -- "The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download <http://www.xmission.com/~legalize/book/download/index.html> Legalize Adulthood! <http://blogs.xmission.com/legalize/>
OK, that came out pretty cleanly, but we're left with the old _MP fractal types. Should these be converted to the integer or floating-point types when a GIF file is loaded with this type? -- "The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download <http://www.xmission.com/~legalize/book/download/index.html> Legalize Adulthood! <http://blogs.xmission.com/legalize/>
Rich asked:
OK, that came out pretty cleanly, but we're left with the old _MP fractal types. Should these be converted to the integer or floating-point types when a GIF file is loaded with this type?
My suggestion would be conversion right after reading the GIF to the corresponding floating point fractal type. I did this already in the float-only version (which Jonathan and I decided when we met in Iowa not to maintain), but it wouldn't be too hard to figure out an equivalent scheme. One can leave dummy elements in the fractalspecific array for deleted types or make a lookup table of indices that maps deleted types to an existing type. This is assuming that you haven't yet refactored the fractalspecific array. Needs to be done sometime, but probably later. Tim
participants (2)
-
Richard -
Tim Wegner