Re: [Lightning] JIT binary translation using lightning possible?

[Paulo César Pereira de Andrade]

2015-10-20 23:45 GMT-02:00 Kaz Nishimura <address@hidden>:

  Hi,

> Your guess is mostly correct.  I am trying to make an application-level
> emulator on a specific m68k-based machine.  The system services will be
> implemented in native code in C++.  I had wrote a direct interpreting
> emulator in the past but it was done when real CPU clocks were about 100
> MHz, so I am trying to rewrite it in a more object-oriented way.
>
> Since host CPUs have been more than 10 times faster now, even direct
> interpretation will be fast enough IMO.  So I think JIT translation as an
> additional performance boost and want to keep the binary translator as
> simple as possible because the largest performance bottleneck will be
> peripheral device emulation.  The key factors in the binary translator are
> simplicity in using the JIT code generator and portability on my plan.  That
> is why I waver between several JIT implementations.

  Ok :)

  Lightning interface is very simple, but my work on it, after becoming
maintainer, is targeting it to be a jit for method based, dynamically typed
languages. You can see an example, not very active lately, of a C/C++
dynamically typed language using lightning at https://github.com/pcpa/owl/
see the lib/oemit*c files for the actual jit translation.

  Lightning does not have some operations that would be desirable for
a binary translator, that is, 32 bit operations on a 64 bit build, or, 64
bit operations on a 32 bit build. So, it may be required to be "creative"
with some operations if generating jit to emulate a 32 bit system on a
64 bit computer.

  For binary translation, a tracing jit could be better. But then you would
need to likely have both, an interpreter and a jit generator.

  Depending on the size of the binaries, you may even be able to fully
translate it to jit. On modern cpus, I "guesstimate" one could generate
up to 1Gb native code in 1 sec, as long as it is a simple direct translation
and everything is already in memory.

Thanks,
Paulo