On Jul 14, 2010, at 4:02 PM, Gary Weinhold wrote:

> Paul Raulerson wrote:
>> I pretty much agree, except there is not any "extreme" overhead in doing so. 
>> And there is not any issue with catching instructions
>> executed out of the data section via an EX instruction.
>>
> I agree that this can be easily trapped if you're trapping every
> instruction, all EX instructions, or all instruction executed from
> within (or outside of) a specified virtual storage range).
>>
>> As a counter example, our production systems run with embedded debugger 
>> code, which does *not* affect execution speed at all.
>>
> I understand that you can run with embedded debugger code without
> affecting execution speed (on MVS ESTAEX can do that, a z/XDC trap can
> also do that).  It would be able to produce diagnostic information if an
> error was detected.  But are you saying (within the context of this
> discussion) that this embedded debugger code could be set up to trap,
> for example, every MVC instruction or every EPSW instuction with no loss
> of execution speed for all other instructions?
>
No, I think you will always take a measure of performance hit doing those kinds 
of traps.
Just that the magnitude of the hit is not really all that bad. I suppose it 
depends upon the
processor though.


> Gary Weinhold
> Data Kinetics, Ltd.
>

Reply via email to