mirror of
https://github.com/zebrajr/pytorch.git
synced 2025-12-06 12:20:52 +01:00
9e71de398b
Adding NUMA awareness through numa_node_id in DeviceOption. Blobs of operators with numa_node_id are allocated on corr. memory banks, using CPU pools with NUMA affinity set to run operators.
39 lines
1.3 KiB
Python
39 lines
1.3 KiB
Python
from __future__ import absolute_import
|
|
from __future__ import division
|
|
from __future__ import print_function
|
|
|
|
from caffe2.python import core, workspace
|
|
from caffe2.proto import caffe2_pb2
|
|
from caffe2.python.test_util import TestCase
|
|
import unittest
|
|
|
|
core.GlobalInit(["caffe2", "--caffe2_cpu_numa_enabled=1"])
|
|
|
|
|
|
@unittest.skipIf(not workspace.IsNUMAEnabled(), "NUMA is not enabled")
|
|
@unittest.skipIf(workspace.GetNumNUMANodes() < 2, "Not enough NUMA nodes")
|
|
class NUMATest(TestCase):
|
|
def test_numa(self):
|
|
net = core.Net("test_numa")
|
|
net.Proto().type = "async_scheduling"
|
|
|
|
numa_device_option = caffe2_pb2.DeviceOption()
|
|
numa_device_option.device_type = caffe2_pb2.CPU
|
|
numa_device_option.numa_node_id = 0
|
|
|
|
net.ConstantFill([], "output_blob_0", shape=[1], value=3.14,
|
|
device_option=numa_device_option)
|
|
|
|
numa_device_option.numa_node_id = 1
|
|
net.ConstantFill([], "output_blob_1", shape=[1], value=3.14,
|
|
device_option=numa_device_option)
|
|
|
|
workspace.RunNetOnce(net)
|
|
|
|
self.assertEqual(workspace.GetBlobNUMANode("output_blob_0"), 0)
|
|
self.assertEqual(workspace.GetBlobNUMANode("output_blob_1"), 1)
|
|
|
|
|
|
if __name__ == '__main__':
|
|
unittest.main()
|