[Simh] several SIMH instances

[Kevin Handy]

Michael Unger wrote:

>On 2006-12-04 23:04, "Kevin Handy" wrote:
>
>  
>
>>The current 'debug/test' version of simh has this OS idle loop code
>>(it looks for specific code for each OS supported), but it doesn't
>>    
>>
>   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>  
>
>>like Linux 2.6 due to the 250 HZ rate. Bypassing that test, it
>>creates a VMS emulation that idles on my 600Mhz machine at
>>< 1% of CPU usage when it's not doing anything. This is MUCH
>>    
>>
>  ^^^^^^^^^^^^^^^^^
>  
>
>>nicer than idling at 100%.
>>    
>>
>
>What about the situation when SIMH is started but an OS isn't booted
>yet? I.e., is this code depending on a "counterpart" at the OS side?
>
>  
>
Yes. If it doesn't see what it wants, it sucks CPU like
crazy (100%), just like before. At the chevron prompt
">>>", it uses 100% of the CPU time. Once VMS is
actually loaded, the idle code kicks in.

>>The idle-loop code can be turned on/off as desired.
>>    
>>
>
>Using 5...10% of the CPU resources would be acceptable as well (at least
>for me) but 1% is much better of course.
>  
>
When VMS is doing something, it uses whatever host
CPU resources it wants, but that is still usually less than
100%. (i.e. it idles while waiting for disks to respond.)

The code is being developed for the 3.7 version of
simh. After bypassing the sanity check that fails
on my Linux 2.6 kernels, it seems to run quite well.
(And yes, I mentioned my problems and what I believe
to be its cause on the development thread. Much
discussion ensued.)