Merge pull request #97 from ORNL/amd

Amd

README.md

@@ -95,6 +95,14 @@ R. Souza, T. Skluzacek, S. Wilkinson, M. Ziatdinov, and R. da Silva
 
 ```
 
+## Disclaimer & Get in Touch
+
+Please note that this a research software. We encourage you to give it a try and use it with your own stack. We
+are continuously working on improving documentation and adding more examples and notebooks, but we are still far from
+a good documentation covering the whole system. If you are interested in working with FlowCept in your own scientific
+project, we can give you a jump start if you reach out to us. Feel free to [create an issue](https://github.com/ORNL/flowcept/issues/new), 
+[create a new discussion thread](https://github.com/ORNL/flowcept/discussions/new/choose) or drop us an email (we trust you'll find a way to reach out to us :wink: ).
+
 ## Acknowledgement
 
 This research uses resources of the Oak Ridge Leadership Computing Facility 

extra_requirements/amd-requirements.txt

@@ -0,0 +1 @@
+pyamdgpuinfo==2.1.6

extra_requirements/analytics-requirements.txt

@@ -1,4 +1,3 @@
 seaborn==0.13.2
-h2o==3.44.0.3
 plotly==5.18.0
 scipy==1.10.1

extra_requirements/data_augmentation-requirements.txt

@@ -0,0 +1 @@
+h2o==3.44.0.3

flowcept/__init__.py

@@ -14,9 +14,8 @@
         model_explainer,
         model_profiler,
     )
-except Exception as _exp:
-    flowcept.commons.logger.exception(_exp)
-
+except:
+    pass
 
 if Vocabulary.Settings.ZAMBEZE_KIND in flowcept.configs.ADAPTERS:
     try:
@@ -50,7 +49,7 @@
             MLFlowInterceptor,
         )
     except Exception as _exp:
-        flowcept.commons.loggerr.error(
+        flowcept.commons.logger.error(
             flowcept.commons._get_adapter_exception_msg(
                 Vocabulary.Settings.MLFLOW_KIND
             )

flowcept/commons/flowcept_dataclasses/telemetry.py

@@ -53,26 +53,26 @@ class Process:
         executable: str
         cmd_line: List[str]
 
-    @dataclass(init=False)
-    class GPU:
-        @dataclass
-        class GPUMetrics:
-            total: int
-            free: int
-            used: int
-            usage_percent: float
-            temperature: float
-            power_usage: float
-
-        gpu_sums: GPUMetrics
-        per_gpu: Dict[int, GPUMetrics] = None
+    # @dataclass(init=False)
+    # class GPU:
+    #     @dataclass
+    #     class GPUMetrics:
+    #         total: int
+    #         free: int
+    #         used: int
+    #         usage_percent: float
+    #         temperature: float
+    #         power_usage: float
+    #
+    #     gpu_sums: GPUMetrics
+    #     per_gpu: Dict[int, GPUMetrics] = None
 
     cpu: CPU = None
     process: Process = None
     memory: Memory = None
     disk: Disk = None
     network: Network = None
-    gpu: GPU = None
+    gpu: Dict = None  # TODO: use dataclasses
 
     def to_dict(self):
         ret = {}
@@ -87,6 +87,7 @@ def to_dict(self):
         if self.network is not None:
             ret["network"] = self.network.__dict__
         if self.gpu is not None:
-            ret["gpu"] = asdict(self.gpu, dict_factory=remove_none_values)
+            # ret["gpu"] = asdict(self.gpu, dict_factory=remove_none_values)
+            ret["gpu"] = self.gpu
 
         return ret

flowcept/commons/utils.py

@@ -83,8 +83,9 @@ def default(self, obj):
 def _get_adapter_exception_msg(adapter_kind):
     return (
         f"You have an adapter for {adapter_kind} in"
-        f" {SETTINGS_PATH} but we couldn't import its interceptor. "
-        f" Consider fixing the following exception or remove that adapter "
+        f" {SETTINGS_PATH} but we couldn't import its interceptor."
+        f" Consider fixing the following exception (e.g., try installing the"
+        f" adapter requirements -- see the README file remove that adapter"
         f" from the settings."
         f" Exception:"
     )

flowcept/configs.py

@@ -107,6 +107,26 @@
 
 TELEMETRY_CAPTURE = settings["project"].get("telemetry_capture", None)
 
+
+##################################
+# GPU TELEMETRY CAPTURE SETTINGS #
+#################################
+
+N_GPUS = dict()
+if TELEMETRY_CAPTURE.get("gpu", False):
+    try:
+        from pynvml import nvmlDeviceGetCount
+
+        N_GPUS["nvidia"] = nvmlDeviceGetCount()
+    except:
+        pass
+    try:
+        import pyamdgpuinfo
+
+        N_GPUS["amd"] = pyamdgpuinfo.detect_gpus()
+    except:
+        pass
+
 ######################
 # SYS METADATA #
 ######################