cuda.tile.printf

cuda.tile.printf(format, *args)

Print the values at runtime from the device

Parameters:

• format (str) – a c-printf style format string in the form of %[flags][width][.precision][length]specifier, where specifier is limited to integer and float for now, i.e. [diuoxXeEfFgGaA]

• *args (tuple[Tile, ...]) – Only tile input is supported.

Examples

>>> tile = ct.arange(4, dtype=ct.int32)
>>> ct.printf("one tile: %d", tile)
>>> ct.printf("two tiles: %d, %f", tile, tile * 2.0)

Notes

This operation has significant overhead, and should only be used for debugging purpose.

When printing from multiple tile blocks, outputs will be interleaved. One workaround is to set optimization level to 0:

@ct.kernel(opt_level=0)
def my_print_kernel():
    ct.printf("%d", 123)