def test_nvml_nvlink_error_counters(ngpus, handles, error_type):
for i in range(ngpus):
for j in range(pynvml.NVML_NVLINK_MAX_LINKS):
assert pynvml.nvmlDeviceResetNvLinkErrorCounters(
handles[i], j) == pynvml.NVML_SUCCESS
error_count = pynvml.nvmlDeviceGetNvLinkErrorCounter(
handles[i], j, error_type)
assert error_count >= 0
def test_nvml_nvlink_error_counters(ngpus, handles, error_type, driver):
if driver > 450.0:
pytest.xfail(XFAIL_LEGACY_NVLINK_MSG)
for i in range(ngpus):
for j in range(pynvml.NVML_NVLINK_MAX_LINKS):
assert (
pynvml.nvmlDeviceResetNvLinkErrorCounters(handles[i], j)
== pynvml.NVML_SUCCESS
)
error_count = pynvml.nvmlDeviceGetNvLinkErrorCounter(
handles[i], j, error_type
)
assert error_count >= 0