GCBASIC documentation
Reference Data

Introduction

This section explores the efficient implementation of lookup reference tables in embedded systems, specifically GCBASIC, focusing on the use of PROGMEM memory to store fixed data sets. It addresses common misconceptions about data storage and initialisation, compares different methods of data handling, and provides advanced techniques for optimising memory usage.

Lookup reference tables are essential in embedded systems for storing fixed data sets that can be accessed during runtime. This section aims to clarify the correct implementation of these tables, debunking common myths and providing practical solutions for efficient data management.

Conventional Misconceptions

A common misconception is that the data required by a fixed lookup table is defined by its content and declared in a Dim statement, with its data filled at runtime. This implies that an array (in RAM memory) is empty initially and populated during initialisation, leading to data duplication and wasted memory resources.

Correct Implementation

A fixed lookup table is a set of data (bytes, words, etc.) stored in PROGMEM memory. The correct implementation involves using the TABLE and READTABLE commands:

There is no DIM in the definition process, and the data is part of the hex file, not filled at runtime.

Memory Efficiency

Storing data in PROGMEM ensures that there is only one copy of the data, avoiding duplication. Copying data to an array is redundant, since reading the table can replace the array.

Practical Solutions

Using TABLE - END TABLE is the simplest way to handle data. For data sets smaller than the EEPROM in the chip, load the table directly to EEPROM and use READTABLE to read the data.

Advanced Techniques

Arrays in Embedded Systems

An array is a special type of variable that can store multiple values, addressed individually using an index. Arrays can be bytes, longs, integers, or words, and are held in RAM. Loading an array can be done element by element or all at once.

Comparison of Methods

Using arrays can be costly in terms of RAM and PROGMEM. The following examples illustrate the difference:

Using an Array

64 words PROGMEM / 13 bytes RAM

        #CHIP 18F2550
        #option Explicit

        Dim myResult, myIndex as Byte

        // Using an array 64 words Progmem / 13 bytes RAM
        Dim myArray(10)
            myArray = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
        For myIndex = 1 to 10
            myResult =  myArray(myIndex)          ' <<< reading a value out of RAM via the array index
        Next

Key line: myResult = myArray(myIndex) — reads element myIndex from myArray, which is stored in RAM; the whole array had to be populated in RAM first, which is why this approach costs 13 bytes of RAM against only 2 for the table-based equivalent below.

Using a Table

56 words PROGMEM / 2 bytes RAM

        #CHIP 18F2550
        #option Explicit

        Dim myResult, myIndex as Byte

        // Using a table 56 words Progmem / 2 bytes RAM
        For myIndex = 1 to 10
            ReadTable myTable, myIndex, myResult          ' <<< the ReadTable instruction
        Next

        Table myTable
            1, 2, 3, 4, 5, 6, 7, 8, 9, 10
        End Table

Key line: ReadTable myTable, myIndex, myResult — reads element myIndex directly out of the fixed myTable data stored in PROGMEM, with no RAM array needed to hold a copy of the data first; only the loop variables myIndex and myResult occupy RAM.

Usage

The ReadTable method provides data set capabilities to chips with limited RAM, using fewer resources and offering faster performance. Advanced techniques and a proper understanding of memory usage can significantly optimise embedded system performance.

Notes

By following these guidelines, developers can efficiently implement lookup reference tables in embedded systems, optimising memory usage and performance.

See Also: