From: Robert M. <mr...@gm...> - 2012-06-16 20:09:19
|
Also, Basilisk II's 680x0 emulation is waaaaaay faster than SheepShaver running Apple's 68K emulation. Sometimes it's even faster than a PowerPC version of the same program running on SheepShaver. Here are some benchmarks for my Mandelbrot program, showing floating-point performance of IEEE 32-bit and 64-bit floating-point add/subtract/multiply. All are on the same 2.2 GHz Nehalem Core i7: BasiliskII's JIT compiler running a 68K native version of the program: Single precision: 227 MFLOPs/sec Double precision: 55.2 MFLOPs/sec SheepShaver running a PowerPC native version of the app: Single precision: 148 MFLOPs/sec Double precision: 164 MFLOPs/sec SheepShaver's JIT compiler running Apple's 68K emulator, running the 68K version of the program and using SANE library for floating point: Single precision: 0.74 MFLOPs/sec Double precision: 0.69 MFLOPs/sec >> I disagree with this one. Basilisk II is quite useful for running late >> 68k Mac stuff, and it would be a shame for it not to pick up any improvements >> that get made to the shared code. >> >> And Basilisk II seems to be a bit more flexible about the memory model . -- Robert Munafo -- mrob.com Follow me at: gplus.to/mrob - fb.com/mrob27 - twitter.com/mrob_27 - mrob27.wordpress.com - youtube.com/user/mrob143 - rilybot.blogspot.com |