ml_switcheroo.compiler.frontends.rdna.tokens

RDNA Tokenizer Definition.

Provides a Regex-based Lexer (RdnaLexer) that decomposes raw AMD RDNA assembly strings into a stream of typed Token objects. It handles specific assembly constructs like Registers (s0, v[0:3]), Immediates (0xFF), Directives (.text), and Modifiers (glc).

Classes

TokenType	Enumeration of valid RDNA token types.
Token	Represents a lexical unit.
RdnaLexer	Regex-based Lexer for AMD RDNA / GCN assembly.

Module Contents

class ml_switcheroo.compiler.frontends.rdna.tokens.TokenType(*args, **kwds)

Bases: enum.Enum

Enumeration of valid RDNA token types.

LABEL_DEF

DIRECTIVE

COMMENT

COMMA

RBRACKET

LBRACKET

COLON

SGPR

VGPR

SPECIAL_REG

IMMEDIATE

MODIFIER

IDENTIFIER

class ml_switcheroo.compiler.frontends.rdna.tokens.Token

Represents a lexical unit.

kind

The type of token.

Type:

TokenType

value

The raw string value.

Type:

str

line

Line number in source (1-based).

Type:

int

column

Column number in source (1-based).

Type:

int

kind: TokenType

value: str

line: int

column: int

class ml_switcheroo.compiler.frontends.rdna.tokens.RdnaLexer

Regex-based Lexer for AMD RDNA / GCN assembly.

PATTERNS: List[Tuple[TokenType, str]]

regex_pairs

tokenize(text: str) → Generator[Token, None, None]

Tokenizes the input string.

Parameters:

text (str) – Raw RDNA source code.

Yields:

Token – Token objects.

Raises:

ValueError – If an unrecognized character sequence is encountered.