IP Core Database


PoC internal IP core database uses INI files and advanced interpolation rules provided by ExtendedConfigParser. The database consists of 5 *.ini files:

  • pyconfig.boards.ini
    This files contains all known FPGA boards and FPGA devices.
  • pyconfig.defaults.ini
    This files contains all default options and values for all supported node types.
  • pyconfig.entity.ini
    This file contains all IP cores (entities) and theirs corresponding testbench or netlist settings.
  • pyconfig.private.ini
    This files is created by .\poc.ps1 configure and contains settings for the local PoC installation. This files must not be shared with other PoC instances. See Configuring PoC’s Infrastructure on how to configure PoC on a local system.
  • pyconfig.structure.ini
    Nodes in this file describe PoC’s namespace tree and which IP cores are assigned to which namespace.

Additionally, the database refers to *.files and *.rules files. The first file type describes, in an imperative language, which files are needed to compile a simulation or to run a synthesis. The latter file type contains patch instructions per IP core. See Files Formats for more details.

Database Structure

The database is stored in multiple INI files, which are merged in memory to a single configuration database. Each INI file defines an associative array of sections and option lines. The content itself is an associative array of options and values. Section names are inclosed in square brackets [...] and allow simple case-sensitive strings as names. A section name is followed by its section content, which consists of option lines.

One option is stored per option line and consists of an option name and a value separated by an equal sign =. The option name is also a case-sensitive simple string. The value is string, starts after the first non-whitespace character and end before the newline character at the end of the line. The content can be of any string, except for the newline characters. Support for escape sequences depends on the option usage.

Values containing ${...} and %{...} are raw values, which need to be interpolated by the ExtendedConfigParser. See Value Interpolation and Node Interpolation for more details.

Sections can have a default section called DEFAULT. Options not found in a normal section are looked up in the default section. If found, the value of the matching option name is the lookup result.


option1 = value1
opt2 =    val ue $2

option1 = ${section1:option1}
opt2 =    ${option1}

Option lines can be of three kinds: An option, a reference, or a user defined variable. While the syntax is always the same, the meaning is infered from the context.

Option Line Kind Distinguishing Characteristic
Reference The option name is called a (node) reference, if the valueof an option is a predefined keyword for the current nodeclass. Because the option’s value is a keyword, it can notbe an interpolated value.
Option The option uses a defined option name valid for the currentnode class. The value can be a fixed or interpolated string.
User Defined Variable Otherwise an option line is a user defined variable. It canhave fixed or interpolated string values.
Name =
Prefix =
arith =         Namespace
bus =           Namespace

addw =          Entity
prng =          Entity

stream =        Namespace
wb =            Namespace
Arbiter =       Entity

Buffer =        Entity
DeMux =         Entity
Mirror =        Entity
Mux =           Entity

fifo_adapter =  Entity
ocram_adapter = Entity
uart_wrapper =  Entity


The database is build of nested associative arrays and generated in-memory from 5 *.ini files. This implies that all section names are required to be unique. (Section merging is not allowed.) A fully qualified section name has a prefix and a section name delimited by a dot character. The section name itself can consist of parts also delimited by dot characters. All nodes with the same prefix shape a node class.

The following table lists all used prefixes:

Prefix Description
INSTALL A installed tool (chain) or program.
SOLUTION Registered external solutions / projects.
CONFIG Configurable PoC settings.
BOARD A node to describe a known board.
CONST A node to describe constraint file set for a known board.
PoC Nodes to describe PoC’s namespace structure.
IP A node describing an IP core.
TB A node describing testbenches.
COCOTB A node describing Cocotb testbenches.
CG A node storing Core Generator settings.
LSE A node storing settings for LSE based netlist generation.
QMAP A node storing settings for Quartus based netlist generation.
XST A node storing settings for XST based netlist generation.
VIVADO A node storing settings for Vivado based netlist generation.
XCI A node storing settings for IP Catalog based netlist generation.

The database has 3 special sections without prefixes:

Section Name Description
PoC Root node for PoC’s namespace hierarchy.
BOARDS Lists all known boards.
SPECIAL Section with dummy values. This is needed by synthesis and overwritten at runtime.

Example section names


The fully qualified section name PoC.bus.stream has the prefix PoC and the section name bus.stream. The section name has two parts: bus and stream. The dot delimited section name can be considered a path in a hierarchical database. The parent node is PoC.bus and its grandparent is PoC. (Note this is a special section. See the special sections table from above.)


Whatever:this is handy to create new field

Supported Options


See py\config.defaults.ini for predefined default values (options) and predefined variables, which can be used as a shortcut.