Re: [Simulavr-devel] Re: [avr-chat] Tenfold acceleration of Similar is possible.

[Paul Schlie]

> From: "E. Weddington" <address@hidden>
>> Joerg Wunsch wrote:
>>> "Dmitry K." <address@hidden> wrote:
>>>
>>> In result speed of simulation increase:
>>> 
>>>   Computer #1:   7.54 --> 3.01 sec
>>>   Computer #2:  18.28 --> 1.51 sec
>> 
>> Impressing -- but simulavr is a dead end.  Only simulavrxx is
>> maintained anymore.
> 
> It's not completely a dead end on Windows. SimulAVR builds for Windows;
> it's unknown if simulavrxx builds for Windows.
> 
> Also, according to Björn (CC'd), it's the only known simulator that will
> run the GCC test suite. So if there's a patch that will speed up
> execution of the latest release of simulavr, I'm open to it. Even if it
> is unmaintained in the long term.

- I'm sure a patch which speeds up simulavr by this much would be greatly
  appreciated by all those who do rely on it (for whatever reason); and
  although the tool is technically without an active maintainer, I suspect
  that someone with cvs access to the sources might even consider patching
  them as a favor after being able confirm that the patch doesn't introduce
  any undesirable functional regressions. (Possibly enabled by first posting
  the patch to allow folks to experiment with it, and then posting their
  corresponding findings?)