Examples

This folder contains examples of code that can be compiled AOT and run via Ravi. Note that these examples can also be JIT compiled - for an example of how that works, please see 21_matrixmul.lua.

The examples assume that Ravi and Ravi Compiler have been built in parallel folders, and the executables are available in build folders.

You can run these examples as follows on a Linux host:

sh matrixmul_aot.sh
sh sieve_aot.sh

The following explains the steps, using the matrix multiplication example.

../build/trun --gen-C -main mymain -f matrixmul_lib.lua > matrixmul_lib.c

Above generates C code that represents the Ravi code in matrixmul_lib.lua. The C code will contain a function named mymain that will be the equivalent of the Lua chunk.

gcc -O2 -shared -fpic matrixmul_lib.c -o matrixmul_lib.so

Above creates a shared library.

../../ravi/build/ravi matrixmul_aot.lua

Above step loads the shared library and then executes it similar to the way a Lua chunk is executed when it is loaded.

local f = package.load_ravi_lib('matrixmul_lib.so', 'mymain')
assert(f and type(f) == 'function')
local matrix = f()

The package.load_ravi_lib() function is a special Ravi extension that loads the mymain function from the shared library. Note that this function behaves like a Lua load, rather than C functions that are normally loaded by Lua.

matrix will contain the table that was returned by the matrixmul_lib.lua script.