OSNews

> Yes one emulator used in the AmigaOS XL package uses
> special JIT emulation techniques (still estimated to be
> only around 40-60% of the performance compared to a
> native AmigaOS port, would it exist),

Mike, I know what I'm talking about, and I also know what you are talking about, unfortunately you don't seem to know what you yourself are talking about.

I know that a JIT compiler is used and I know what JIT means and how it works, I also know that by the sole fact that emulation is involved and that the memory is kept in big endian order, despite the complete disliking of that format by the x86 family of processors, the emulated AmigaOS environment can and will never be faster than a native implementation.

As a matter of facts, I did some tests and published them on ANN, in which an app, compiled with the amithlon port of gcc, which directly emits native x86 code, albeit accessing the memory in big endian mode, is shown to be 15% slower than the same app running on linux. If a "native" Amithlon app is slower than a native linux one, can you tell me how can an emulated app be faster than a native one?

> this combined with the efficiency of the *original* 68K
> AmigaOS and the fact that emulator uses native x86
> drivers and can use native x86 Amiga applications, it
> truly *is* alot faster and a much more feature complete
> solution than AROS currently.

More feature complete, without any doubts, faster? LOL

> Note that this emulator sits on top of a small optimised
> x86 kernel and is not running on top of a bloated OS like
> Windows.

The "small optimized x86 kernel" is just a plain linux kernel, patched a little to suit the amithlon needs, surely not optimized at all. See the benchs mentioned above for further information.

I'd suggest you to make correct information, next time.