FW: [Simh] SimH vax.exe cpu usage on Windows.

Wed May 19 04:32:50 EDT 2004

Mark Pizzolato wrote:
> Hi Lewis,
>
> I believe that you are worried about the 100% thing when you really
> don't have to.
>
> When I first touched simh, one of the first changes I submitted to
> Bob was a lowering of the CPU priority on both Windows and Unix
> systems when instructions are executing.  This has been in the code
> base for almost 2 years now.  This is done internal to simh, so any
> external efforts to lower priority are merely making it lower than it
> already made itself.
>
> Since the simh process is running at a lower priority, it will
> simulate instructions only when the hosting system (Windows in your
> case) is otherwise idle.  Usually the only process that "suffers"
> from simh running at 100% is the system idle process, which exists
> merely to soak up CPU cycles that are otherwise unused.
>
> So, in the end, you don't really need separate system to run your VAX
> continuously.
>
> Your comments about lowering the priority having no apparent gain in
> performance is true, since it is already low, you could observe the
> same "no net gain" if you stopped the simh process.  Hence it really
> doesn't affect anything.
>
> Since your system is NOT a Notebook (which could benefit from
> dynamically turning down clock speed when the system is idle to
> reduce battery power consumption), running the CPU at 100% won't hurt
> anything.  It won't create more heat, or make your disks work harder,
> or anything.
>

Is this correct ?  My understanding is that the CPU heat is directly related
to how much work the CPU is actually doing.  I quick monitoring of my own
CPU temperature seems to reveal an increased temperature when the CPU is
blocked at 100% busy.

The overclocker guys are obsessive about this fact.  I have also seen
suggested that CPU life expectancy is directly related to thelong term
thermal load under which it is placed - cooler being better.

I may be wrong on all counts, but I think Bishops concern is that he may
decrease the life expectancy of his CPU by always having SIMH running, due
to greater than normal thermal loads.

Unless it is overclocked (and probably not even then) I do not think this
should be a problem.

> Andreas suggested that if you could figure out when the simulated
> system is "idle", you could get the simulator to idle itself as well.
>  As I've already said this is likely not worth the effort since there
> really is no net gain.
>
> There is, however, one case where implementing such an "idle"ing
> concept would benefit things.  This would come up if you somehow
> wanted or needed to run more than one simulator on the same host
> system simultaneously.
>
> All of that said, I've actually have a version of simh VAX which does
> support idling (or yielding) the CPU when the simulated system is
> idling.  The changes to support this affect several chinks of simh
> code.  I submitted it to Bob about a year and a half ago, and he
> didn't see the need.
>

Personally though, I would have included these changes, unless they have
some massive downside.

philip lewis

> - Mark Pizzolato
>
> --- "Bishop, Lewis" <lewis.bishop at eds.com> wrote:
>> Thanks for the responses so far...
>>
>> I have a dilemma whereby I cannot really justify a separate system
>> as this
>> is a home setup and whilst it's nice to have up and running it
>> won't
>> honestly get much use. The performance of the system itself is
>> actually
>> okay, however, at times with the load on the system it does affect
>> it quite
>> a bit - the system is also functioning as a PDC and exchange server
>> and runs
>> virtual SuSE Linux and Solaris systems with Oracle installed in
>> each - so it
>> does a fair bit! ;-) Memory is not a problem as it has 1gb -
>> normally
>> leaving 256mb (25%) free after everything else.
>>
>> I'm keen to reduce the CPU as I don't like to run the CPU at 100%
>> for long.
>> I'm not convinced it's good for it, especially as it's not a server
>> grade
>> system (i.e. it's a beefed up home PC or it was when I bought it 4
>> years
>> ago!).  There are a number of discussions around discussing this
>> type of
>> usage and a lot of concerns are raised not so much over the high
>> CPU usage
>> but over the other system components that have to run - i.e.
>> cooling fans,
>> power supply etc... all of these components, having a limited
>> lifetime are
>> going to fail at some stage - running a fan 24/7 is obviously going
>> to
>> stress it more than a few hours running per day.
>>
>> I have set the priority down to low and to be honest it doesn't
>> seem to make
>> any difference to the overall system performance. I've tried some
>> of the PC
>> slow utilities but these have not reduced the CPU usage (didn't
>> think they
>> would). If it cannot be done within SimH then maybe some form of
>> resource
>> manager would  be good?
>>
>> Anyone have any further ideas?
>>
>> Lewis Bishop
>>
>>
>> -----Original Message-----
>> From: Parker, David [mailto:David.Parker at stjohn.org]
>> Sent: Tuesday, 18 May 2004 4:29 p.m.
>> To: Bishop, Lewis
>> Subject: RE: [Simh] SimH vax.exe cpu usage on Windows.
>>
>>
>> You could try setting the priority lower on the application
>> process.
>> Or, if performance is just that important, run it on another
>> machine.
>> I'm running mine (pretty much the same setup) on an old 850Mhz CPU
>> and it's
>> decent.
>>
>> $0.02
>>
>> -----Original Message-----
>> From: Bishop, Lewis [mailto:lewis.bishop at eds.com]
>> Sent: Mon 5/17/2004 6:20 PM
>> To: 'simh at trailing-edge.com'
>> Cc:
>> Subject: [Simh] SimH vax.exe cpu usage on Windows.
>>
>>
>>
>> Hi all,
>>
>> I have OpemVMS 7.3 running under SimH v2.10-3 under Windows 2000
>> with the
>> networking option using WinPCap and a 2nd NIC. I'm really happy
>> with
>> the
>> fact that i can get VMS running on a windows machine (after not
>> using it for
>> 6 years or so) but i'm a bit concerned to see the cpu usage by
>> vax.exe
>> sitting at max (as much as it can get) and taking complete system
>> cpu to
>> 100%.
>>
>> Although system performance is not 'hammered' by this (it releases
>> resource
>> when requested) the system performance has suffered slightly. I
>> have
>> tested
>> the latest distribution (v3.2) but it is the same - does anyone
>> know
>> if
>> there is a way to limit the CPU usage for vax.exe?
>>
>> Thanks,
>>
>> Lewis Bishop.
>>
>> _______________________________________________
>> Simh mailing list
>> Simh at trailing-edge.com
>> http://mailman.trailing-edge.com/mailman/listinfo/simh
>>
>>
>>
>>
>> CONFIDENTIALITY NOTICE:  This email message and any accompanying
>> data are
>> confidential, and intended only for the named recipient(s).  If you
>> are not
>> the intended recipient(s), you are hereby notified that the
>> dissemination,
>> distribution, and or copying of this message is strictly
>> prohibited.  If you
>> receive this message in error, or are not the named recipient(s),
>> please
>> notify the sender at the email address above, delete this email
>> from your
>> computer, and destroy any copies in any form immediately.
>> _______________________________________________
>> Simh mailing list
>> Simh at trailing-edge.com
>> http://mailman.trailing-edge.com/mailman/listinfo/simh
>
> _______________________________________________
> Simh mailing list
> Simh at trailing-edge.com
> http://mailman.trailing-edge.com/mailman/listinfo/simh