[Simh] binary packages...

Brett Bump bbump at rsts.org
Thu Sep 6 18:20:04 EDT 2007


On Thu, 30 Aug 2007, Bob Supnik wrote:

> My goodness! So many suggestions, I hardly know how to deal with them all...
>
> 1. Source control
>
> There is already a SimH project on SourceForge.  It is not up to
> date.  There are several admins (other than me), feel free to bother
> them to bring it up to date.

Cool, I didn't realize this.  Might help the load on Tim's hardware if we
change links on our pages for SimH from trailing-edge to sourceforge.

On Wed, 5 Sep 2007, Jason Stevens wrote:

> Thanks to Tim Riker I've been making all kinds of updates to the sourceforge
> page for simh ( http://sourceforge.net/projects/simh/ ).
>
> I've also placed in the binaries & most of the system software from the main
> site into the sourceforge, giving the main site a bit of a break for
> bandwidth.  I plan on updating the CVS source to the current level once I
> feel comfortable with manipulating it via sourceforge.
>
> The big question is what platforms would you want binaries for?

All of them of course. ;-)  I would think most are running generic PCs or
an Apple equivalent.  I have some Alpha's running SimH, for no real reason
other than it's all DEC (without the weight, size and power to an H960).

> Also for win32 OS's does anyone have any good idea on how to benchmark the
> various compilers to see which is the fastest?  I was going to time building
> a kernel in single user mode on any *nix to see how fast that is (with idle
> on & off)..  Does that sound like a valid test?  I don't have any numbers,
> but the Watcom 11 compiler feels pretty snappy.  But it would be interesting
> to put up mingw/cygwin(GCC) VisualC++ & Watcom to see which is the fastest.

Depends on what your after.  Compile time isn't (normally) as important
as execution time.  For high-performance-computing, the people at NCSA
recommend using using the Intel C/C++ compiler for cluster computing on
the Grid (most likely, this is a better profiling compiler).  Outside
of the scientific world, programmers don't seem to be profiling code
anymore, so I don't know that it really matters.

> FWIW it seems there are some downloads from the site already.  I figure the
> exposure is a 'good thing'.
>
> Let me know if there is something that you feel should be on the site..
>
>
> Jason
>

Might not be a bad idea to also post a descriptive narrative on some
other high profile sites (that link to souceforge) such as osnews.com,
linux.com, freshmeat.org and slashdot.com (SimH latest release).
Just a thought.

Brett



More information about the Simh mailing list