Re: NetBSD/vax compiler bounty?

[Dave McGuire <mcguire%neurotica.com@localhost>]

On 4/16/20 2:23 PM, Alexander Schreiber wrote:
> On Thu, Apr 16, 2020 at 07:12:07PM +0200, Martin Husemann wrote:
>> On Thu, Apr 16, 2020 at 12:18:46PM -0400, emanuel stiebler wrote:
>>> What kind of performance would one get of the simh/va/netbsd on a decent
>>> PC this days?
>>
>> I would expect it to outperform any real hardware :-)
> 
> It does, by a wide margin. I'm running a simulated MicroVAX 3100 in
> SIMH on a i5-3570K. I also found the source for the Dryhstone benchmark.
> According to it, my virtual VAX clocks in at 34.4 VUP (the real metal
> apparently is supposed to be at 3.8 VUP). According to the same
> benchmark, the host machine has 21858 VUP. While I advise taking these
> numbers with industrial amounts of salt, SIMH VAX on current x86 has
> been known to outperform the original metal for a while.

  Ah, you're comparing it against a very old, relatively slow low-end
VAX.  Even the very common 4000-60 is more than three times the speed of
that 3100.  I'm not sure this is a fair comparison.  There's plenty of
real VAX iron out there that can hold its own against an emulator on
modern hardware.

             -Dave

-- 
Dave McGuire, AK4HZ
New Kensington, PA