Echelon Neuron C User Manual
Page 187
Neuron C Programmer’s Guide
175
• Each address table entry requires 5 bytes. A maximum of 15 address
table entries are allowed. The minimum is 0. The default is 15 entries.
See
on page 188.
• Each network variable declared (input or output) uses 3 bytes for its
configuration information. In addition, it uses 3 bytes of read-only
memory for its fixed information. If you use the SNVT self-identification
(SI) feature, there is an additional 7-byte fixed overhead plus 2 additional
bytes per network variable (minimum).
• Each network variable alias table entry uses 4 bytes. There is no default
for the size of this table. See
on page 189.
• Variables declared as eeprom and config in your program use an amount
of EEPROM corresponding to its C data type. This includes network
variables of the config_prop (or cp) class and modifiable configuration
parameters declared using the cp_family keyword (the latter are stored
together in a writable value file). See
on page
• The when clause table is placed in the code memory area (ROM, if
available, or EEPROM). Each when clause uses a table entry from 3 to 6
bytes (most are 3 bytes). This code space is usually slightly smaller than
the equivalent code generated by an if statement. Additional code space
can result from when clauses containing user-defined events.
• The interrupt dispatcher and the interrupt_control( ) function are placed
in the code memory area (ROM, if available, or EEPROM or flash
memory). The interrupt_control( ) function takes a constant, small,
amount of code space (12 bytes at the time of release, but the exact
number is subject to change). The interrupt dispatcher is optimized for
the each application, and consumes the smallest amount of memory
possible. Its size varies with the number and type of interrupt tasks that
your application defines.
• The read-only value file is normally placed in the CODE memory area
(this can be ROM, if available, or EEPROM). The configuration value
files use only the number of data bytes required by the data types of the
configuration properties contained within. The #pragma codegen
put_read_only_cps_in_data_memory directive instructs the linker to
place the read-only value file in a modifiable memory area instead of the
CODE memory area. See the
Compiler Directives
chapter in the
Neuron
C Reference Guide
for more information about this directive. There is no
additional overhead. The configuration template file is also placed in the
CODE memory area. The template file uses a number of bytes to
describe each configuration property in the value files. This number
varies based on the type and characteristics of the configuration property,
but it is typically 12 or more bytes per configuration property instance.
See the
Compiler Directives
chapter in the
Neuron C Reference Guide
for
more information about the #pragma codegen
cp_family_space_optimization compiler directive. Use of this directive
can substantially reduce the size of the configuration property template
file.