[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Linrad] Re: Setting priorities in Windows

Hello Alberto,

> There is a profound difference between the now dead Windows 98 and Windows XP or later.
> If you try with Win98, you will see that setting your process as having the Real Time
> priority class nearly freezes everything else on that PC.
> In Windows XP and later, they took the needed measures to ensure that the kernel 
> of Windows wasn't starved of CPU cycles in such a circumstance, so that an user 
> program cannot freeze anymore Windows, no matter whet its priority class (for 
> the process) or priority level (for the threads) may be.

Now I am home so I can do real life tests under Windows 98.

I have set parameters for Linrad in such a way that the system monitor
shows 96% CPU load. Linrad runs fine at normal, high and real time priority.

When I start another program the difference is profound. At normal
or high priority Linrad crasches because it can not produce data in
time for the output during a too long time. In real time mode
however Linrad workes fine and the CPU load goes to 100% when
I start another program. That other program becomes very slow since
it can get only about 4% of the CPU time.

More obviously: When Linrad is the only program running (at 96% CPU)
an attempt to resize Linrad causes problems in normal priority as
well as in high priority. The resize operation works normally, but
the Linrad output becomes stuttered. It may even exit with error 
message 1058 since restariing after a sync error may fail.

With real time priority it is different. If the resize operation is fast,
meaning that the time interval while the mouse button is depressed,
the resize is sucessful without any errors in the Linrad output. The
completion of the resize operation takes about 10 seconds. (If the
Linrad window is made smaller, it takes about 10 seconds for the old 
big one to disappear while Linrad works normally.)

The Win98 computer used is a Pentium II-MMX at 350 MHz. Interestingly, 
after I had finished experimenting, the normal shut-down procedure of 
Windows did not work. I had to push the reset button. Presumably
that is a system bug that will not be present in modern Windows.

To me, a computer running Linrad is a radio receiver. It may do
other tasks simultaneously, but it should never allow other processes
to destroy receiver performance. It seems that real time priority
is the appropriate solution.

NOTE. It is a very good idea to run Linrad under normal priorty
while experimenting with parameter values. Only when setup is 
completed and you know for sure Linrad will not use too much CPU 
power (Less than 80% or so) it may be advisable to set realtime 


Leif / SM5BSZ

You received this message because you are subscribed to the Google Groups "Linrad" group.
To post to this group, send email to linrad@xxxxxxxxxxxxxxxx
To unsubscribe from this group, send email to linrad+unsubscribe@xxxxxxxxxxxxxxxx
For more options, visit this group at http://groups.google.com/group/linrad?hl=en