Atmel CAVR-4 Manual De Usuario

CAVR-4

Part 1. Using the compiler

By selecting a memory model, you have selected a default memory type that your 
application will use. However, it is possible to specify—for individual variables or 
pointers—different memory types. This makes it possible to create an application that 
can contain a large amount of data, and at the same time make sure that variables that 
are used often are placed in memory that can be efficiently accessed.

The AVR IAR C/C++ Compiler provides a set of extended keywords, which can be used 
as data memory attributes. These keywords let you override the default memory type for 
individual data objects, which means that you can place data objects in other memory 
areas than the default memory. This also means that you can fine-tune the access method 
for each individual data object, which results in smaller code size.

The following table summarizes the available memory attributes:

_ _tiny

1 byte

Data

0-0xFF

128 bytes

_ _near

2 bytes

Data

0-0xFFFF

32 Kbytes

_ _far

3 bytes

Data

0-0xFFFFFF

 (16-bit pointer 

arithmetics)

32 Kbytes

_ _huge

3 bytes

Data

0-0xFFFFFF

8 Mbytes

_ _tinyflash

1 byte

Code

0-0xFF

128 bytes

_ _flash

2 bytes

Code

0-0xFFFF

32 Kbytes

_ _farflash

3 bytes

Code

0-0xFFFFFF

 (16-bit pointer 

arithmetics)

32 Kbytes

_ _hugeflash

3 bytes

Data

0-0xFFFFFF

8 Mbytes

_ _eeprom

1 bytes

EEPROM

0-0xFF

128 bytes

_ _eeprom

2 bytes

EEPROM

0-0xFFFFFF

32 Kbytes

_ _io

N/A

I/O space

0–0x3F

4 bytes

_ _io

N/A

Data

0x60–0xFF

4 bytes

_ _ext_io

N/A

Data

0x100–0xFFFF

4 bytes