From brewer@fcrao1.astro.umass.edu Thu Jul 27 14:53:46 2000
Date: Thu, 27 Jul 2000 14:53:29 -0400
From: "Michael K. Brewer" <brewer@fcrao1.astro.umass.edu>
Subject: Re: Test Results
To: rmillan@cfa.harvard.edu
Cc: smorel@cfa.harvard.edu, wtraub@cfa.harvard.edu, schloerb@astro.umass.edu,
        jmonnier@cfa.harvard.edu
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
Content-transfer-encoding: 7BIT
X-Priority: 3
X-MSMail-priority: Normal

Rafael,

   I measured the DIO output times in 2 ways.  First I output a square
wave on one bit and hooked it up to a frequency counter.  Second I
measured the time to output 10000 bytes with the AuxClk.  Both
measurements were consistent.  For the input, I just measured the time
to input 10000 bytes with the AuxClk.

Sebastien's code ran slower because to do an output he first did an
input to read the data from the latches before modifying the bit he
wanted to send out.  I am holding the data in shared memory, so I can
just modify some bits and write them out to the latches.

The tracking tasks are the celestial sphere update, the telescope
update, the delayline update which run once per second and the axis
update (Axis_Proc) which runs 4 times per second.  The tasks that run
once per second do a lot of calculations and therefore use significant
CPU time.  The axis task does very few calculations and so executes
quite quickly.

Synchronization with the tracking programs would yield the same result
as the tests without the tracking programs running so long as the data
acquisition/fringe tracking task ran at a higher priority than
anything else that was executing concurrently.

I think this DIO card should be adequate.


                   Mike