VXi VT1422A 사용자 설명서
198 Creating and Running Algorithms
How Does it Work?
The example algorithm definition above will be used for this discussion.
When a value for <swap_size> is specified at algorithm definition, the
VT1422A allocates two identical algorithm spaces for ALG3, each the size
specified by <swap_size> (in this example 1000 words). This is called a
"double buffer," arbitrarily called space A and space B. The algorithm is
loaded into ALG3's space A at first definition. Later, while algorithms are
running, "replace" ALG3 by again executing:
When a value for <swap_size> is specified at algorithm definition, the
VT1422A allocates two identical algorithm spaces for ALG3, each the size
specified by <swap_size> (in this example 1000 words). This is called a
"double buffer," arbitrarily called space A and space B. The algorithm is
loaded into ALG3's space A at first definition. Later, while algorithms are
running, "replace" ALG3 by again executing:
ALG:DEF ALG3,#42435<2435char_alg_source>
Notice that <swap_size> is not (must not be) included this time.
The ALG:DEF command works like an Update Request. The VT1422A
translates and downloads the new algorithm into ALG3's space B while the
old ALG3 is still running from space A. When the new algorithm has been
completely loaded into space B and an ALG:UPDATE command has been
sent, the VT1422A simply switches to executing ALG3's new algorithm
from space B at the next Update Phase (see Figure 6-2). If yet another ALG3
were sent, it would be loaded and executed from ALG3's space A.
The ALG:DEF command works like an Update Request. The VT1422A
translates and downloads the new algorithm into ALG3's space B while the
old ALG3 is still running from space A. When the new algorithm has been
completely loaded into space B and an ALG:UPDATE command has been
sent, the VT1422A simply switches to executing ALG3's new algorithm
from space B at the next Update Phase (see Figure 6-2). If yet another ALG3
were sent, it would be loaded and executed from ALG3's space A.
Determining an
Algorithm's Size
In order to define an algorithm for swapping, it is necessary to know how
much algorithm memory to allocate for it or any of its replacements. It can
be queried from the VT1422A using the following sequence:
much algorithm memory to allocate for it or any of its replacements. It can
be queried from the VT1422A using the following sequence:
1. Define the algorithm without swapping enabled. This will cause the
VT1422A to allocate only the memory actually required by the
algorithm.
algorithm.
2. Execute the ALG:SIZE? <alg_name> command to query the amount
of memory allocated. The minimum amount of memory required for
the algorithm is now known.
the algorithm is now known.
3. Repeat 1 and 2 for each of the algorithms that are to be swapped with
the original. From this, the minimum amount of memory required for
the largest is known.
the largest is known.
4. Execute *RST to erase all algorithms.
5. Re-define one of the algorithms with swapping enabled and specify
<swap_size> at least as large as the value from step 3 above (and
probably somewhat larger because as alternate algorithms declare
different variables, space is allocated for total of all variables
declared).
probably somewhat larger because as alternate algorithms declare
different variables, space is allocated for total of all variables
declared).
6. Swap each of the alternate algorithms for the one defined in step 5,
ending with the one that is to be run now. Remember, not to send the
<swap_size> parameter with these. If an "Algorithm too big" error is
not received, then the value for <swap_size> in step 5 was large enough.
<swap_size> parameter with these. If an "Algorithm too big" error is
not received, then the value for <swap_size> in step 5 was large enough.
7. Define any other algorithms in the normal manner.