[Simh] Compiler differences

Bob Eager rde at tavi.co.uk
Fri May 27 17:13:20 EDT 2016


There's a VAX benchmark program. I ran it on my last machine and got
something in excess of an 8600!

I'd be surprised if it isn't around somewhere, but if it isn't then I
can upload it.

On Fri, 27 May 2016 17:03:18 -0400
Clem Cole <clemc at ccc.com> wrote:

> Is there a reasonable benchmark - moral equivalent of running the 741
> op-amp thru spice - Say a VAX under BSDx or VMS running some defined
> load?
> 
> 
> Also have you ever tried the DEC^h^h^hIntel C compiler on an x86 or
> INTEL*64 target (Linux, Winders, or OS X).
> 
> Clem
> 
> On Friday, May 27, 2016, Mark Pizzolato <Mark at infocomm.com> wrote:
> 
> > My observations are that clang does indeed compile faster than gcc,
> >
> > however that doesn’t say much about how fast the compiled results
> >
> > actually run.
> >
> >
> >
> > If you’re only compiling a single simulator (which most users are
> >
> > probably interested in) the compile time difference isn’t enough
> >
> > to worry about.
> >
> >
> >
> > When changes to simulators are made, compiling with both gcc
> >
> > and clang are used since these compilers tend to detect different
> >
> > errors.
> >
> >
> >
> > -        Mark
> >
> >
> >
> > *From:* Simh [mailto:simh-bounces at trailing-edge.com
> > <javascript:_e(%7B%7D,'cvml','simh-bounces at trailing-edge.com');>]
> > *On Behalf Of *Kevin Handy
> > *Sent:* Friday, May 27, 2016 12:04 PM
> > *To:* simh at trailing-edge.com
> > <javascript:_e(%7B%7D,'cvml','simh at trailing-edge.com');>
> > *Subject:* [Simh] Compiler differences
> >
> >
> >
> > Has anyone done a comparison (benchmarks) between simh emulators
> > compiled with g++ verses clang?
> >
> > clang seems to me to run quite a bit faster than gcc, but I haven't
> > run any actual comparisons between the two so it is completely
> > subjective.
> >
> > A full 'make' seems to build faster with clang than gcc.
> >
> > Just curious. I recently noticed bthe clang stuff in the makefile,
> > so I thought I'd try it out.
> >
> 
> 



More information about the Simh mailing list