Merge branch 'main' into main-public

Author: ashahba

notebooks/performance/hf_text_classification_performance.ipynb

@@ -87,19 +87,22 @@
    "outputs": [],
    "source": [
     "# Get and display CPU/platform information\n",
-    "cpu_info = CPUInfo()\n",
-    "platform_util = PlatformUtil()\n",
-    "print(\"{0} CPU Information {0}\".format(\"=\" * 20))\n",
-    "print(\"CPU family:\", platform_util.cpu_family)\n",
-    "print(\"CPU model:\", platform_util.cpu_model)\n",
-    "print(\"CPU type:\", platform_util.cpu_type)\n",
-    "print(\"Physical cores per socket:\", cpu_info.cores_per_socket)\n",
-    "print(\"Total physical cores:\", cpu_info.cores)\n",
-    "cpufreq = psutil.cpu_freq()\n",
-    "print(\"Max Frequency:\", cpufreq.max)\n",
-    "print(\"Min Frequency:\", cpufreq.min)\n",
-    "cpu_socket_count = cpu_info.sockets\n",
-    "print(\"Socket Number:\", cpu_socket_count)\n",
+    "try:\n",
+    "    cpu_info = CPUInfo()\n",
+    "    platform_util = PlatformUtil()\n",
+    "    print(\"{0} CPU Information {0}\".format(\"=\" * 20))\n",
+    "    print(\"CPU family:\", platform_util.cpu_family)\n",
+    "    print(\"CPU model:\", platform_util.cpu_model)\n",
+    "    print(\"CPU type:\", platform_util.cpu_type)\n",
+    "    print(\"Physical cores per socket:\", cpu_info.cores_per_socket)\n",
+    "    print(\"Total physical cores:\", cpu_info.cores)\n",
+    "    cpufreq = psutil.cpu_freq()\n",
+    "    print(\"Max Frequency:\", cpufreq.max)\n",
+    "    print(\"Min Frequency:\", cpufreq.min)\n",
+    "    cpu_socket_count = cpu_info.sockets\n",
+    "    print(\"Socket Number:\", cpu_socket_count)\n",
+    "except Exception as e:\n",
+    "    print(\"Warning, Unable to parse CPU information. It is recommended to use a Xeon machine. Without a Xeon machine, behavior is unknown\")\n",
     "\n",
     "print(\"\\n{0} Memory Information {0}\".format(\"=\" * 20))\n",
     "svmem = psutil.virtual_memory()\n",
@@ -112,7 +115,7 @@
     "\n",
     "# Display PyTorch version information\n",
     "print(\"\\n{0} PyTorch Information {0}\".format(\"=\" * 20))\n",
-    "print(\"PyTorch version:\", torch.__version__)"
+    "print(\"PyTorch version:\", torch.__version__)\n"
    ]
   },
   {
@@ -810,7 +813,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.6"
+   "version": "3.9.17"
   }
  },
  "nbformat": 4,

notebooks/requirements.txt

@@ -31,5 +31,5 @@ tensorflow-hub==0.15.0
 torch==1.13.1
 torchvision==0.14.1
 transformers==4.34.0
-urllib3==2.0.6
+urllib3==2.0.7
 evaluate==0.4.0

tests/utils/files/sorted_membind_info_no_numa.json

@@ -1,22 +1,22 @@
 [
     [{
         "cpu_id": 0,
-        "node_id": 0,
-        "socket_id": -1
+        "node_id": -1,
+        "socket_id": 0
     }],
     [{
         "cpu_id": 1,
-        "node_id": 1,
-        "socket_id": -1
+        "node_id": -1,
+        "socket_id": 1
     }],
     [{
         "cpu_id": 2,
-        "node_id": 2,
-        "socket_id": -1
+        "node_id": -1,
+        "socket_id": 2
     }],
     [{
         "cpu_id": 3,
-        "node_id": 3,
-        "socket_id": -1
+        "node_id": -1,
+        "socket_id": 3
     }]
 ]

tlt/utils/platform_util.py

@@ -103,15 +103,15 @@ def _sort_membind_info(membind_bind_info):
             for entry in membind_bind_info:
                 cpu_id = int(entry[0])
                 core_id = int(entry[1])
-                node_id = int(entry[2])
+                socket_id = int(entry[2])
                 # On a machine where there is no NUMA nodes, entry[3] could be empty, so set socket_id = -1
                 if entry[3] != "":
-                    socket_id = int(entry[3])
+                    node_id = int(entry[3])
                 else:
-                    socket_id = -1
+                    node_id = -1
 
                 # Skip nodes other than current node number
-                if node_number != node_id:
+                if node_number != socket_id:
                     continue
 
                 # Add core info