CLI Reference#

cutile-basic provides a command-line interface for compiling BASIC source files.

Usage#

python -m cutile_basic.cli [options] <input.bas>

Argument	Description
`input`	Input `.bas` file (required)

Option	Description
`-o, --output FILE`	Write the compiled `.cubin` to `FILE`. If omitted, the path to the `.cubin` is printed to stdout.
`--dump-tokens`	Dump the token stream and exit.
`--dump-ast`	Dump the parsed AST and exit.
`--dump-analyzed`	Dump the analyzed program (symbols, types, metadata) and exit.
`--gpu-arch ARCH`	GPU architecture for compilation (e.g. `sm_120`). Default: auto-detect.

Compile and print the path to the .cubin:

python -m cutile_basic.cli examples/vector_add.bas

Write the .cubin to a file:

python -m cutile_basic.cli examples/vector_add.bas -o vector_add.cubin

Inspect the token stream:

python -m cutile_basic.cli examples/hello.bas --dump-tokens

Inspect the AST:

python -m cutile_basic.cli examples/hello.bas --dump-ast

Compile to cubin with an explicit GPU architecture:

python -m cutile_basic.cli examples/vector_add.bas --gpu-arch sm_120 -o vector_add.cubin