[Simh] Throttle bug

[Villy Madsen]

Guess what Robert - it did it to me today.

You can run readthrottle to read the current state of the throttle.

In my case it was set to 1.

A quick fix is to

Stop throttle

run resetthrottle (out of whatever directory these programs are in)

submit sys$startup:throttle to resubmit the job

When I ran READTHROTTLE it showed a value of 1 - which means that a 1ms 
sleep is thrown in every 400 vax instructions - a more reasonable value for 
my system is a minimum of 4.


If you can send me a throttle log showing this problem - that would help me 
to debug the problem.

What I am looking for looks like this:
         11           176
         11           176
         11           170
          8           132
          8           139
          8           139
          6            96
          6            94
          6            94
          4            76
          4            58
MinVal =            4

The first number is the cpu sleep control value, and the 2nd number is 
expected performance in tenths of a VUP.  In other words  a control value 
of 4 - giving us a 1ms sleep every 1600 instructions equates to a .6 VUP 
vax running on my 1.5  ghz P3 running w/98SE.

Thanks for bringing this to my attention

Villy

On Wednesday, May 17, 2006 09:37, Robert Bustichi [SMTP:robert at smrn.com] 
wrote:
> On the latest version of throttle you posted on 5/9, there seems to be a
> bug.  Here is what happens.
>
> I boot up VMS on SIMH and Throttle gets submitted.
> After the two minutes, I see the host (Windows 2003) CPU usage drop like
> it should.  But, SIMH is very, very slow and remains slow.  The host CPU
> usage remains at 0%.  It does not appear that Throttle is working 
properly.
>
> If I re-execute RESETTHROTTLE, everything seems to work as your previous
> version did.
>
> As as test, I executed @THROTTLE interactivly so I can see what is going
> on.  Everything seems to be ok, as I see output and no errors.
>
> Any ideas?