MCPcopy Create free account
hub / github.com/apache/tvm / have_int8

Function have_int8

python/tvm/support/nvcc.py:1055–1067  ·  view source on GitHub ↗

Either int8 support is provided in the compute capability or not Parameters ---------- compute_version : str compute capability of a GPU (e.g. "6.1")

(compute_version)

Source from the content-addressed store, hash-verified

1053
1054
1055def have_int8(compute_version):
1056 """Either int8 support is provided in the compute capability or not
1057
1058 Parameters
1059 ----------
1060 compute_version : str
1061 compute capability of a GPU (e.g. "6.1")
1062 """
1063 major, _ = parse_compute_version(compute_version)
1064 if major >= 6:
1065 return True
1066
1067 return False
1068
1069
1070def have_tensorcore(compute_version=None, target=None):

Callers 1

check_cudaFunction · 0.90

Calls 1

parse_compute_versionFunction · 0.70

Tested by 1

check_cudaFunction · 0.72

Used in the wild real call sites across dependent graphs

searching dependent graphs…