tvm_tirx.testing.env.has_cudagraph

Contents

tvm_tirx.testing.env.has_cudagraph#

tvm_tirx.testing.env.has_cudagraph()#

True if a CUDA device is present and the toolkit supports CUDA Graphs.

Implies has_cuda(): nvcc.have_cudagraph() only checks the toolkit version, so the device guard must be explicit. Gate a test with @pytest.mark.skipif(not tvm_tirx.testing.env.has_cudagraph(), reason=...) (add @pytest.mark.gpu for CI selection).

Return type:

bool

Contents