Improve baseline comparison (#817)

* Do not force unsupported metrics to be specified in older gpu
  architectures as None

* Remove metrics which are explicitly set to None

* Update CHANGELOG

* Fix analysis configuration to fix baseline comparisons across all gpu
  architectures
    * Add missing 1812 section for gfx908
    * Add missing 1812 section for gfx90a

* Baseline comparision will only show common metrics
   * First workload will be used to set Metric ID index column

[ROCm/rocprofiler-compute commit: d4c316a730]

projects/rocprofiler-compute/src/utils/file_io.py

@@ -75,6 +75,12 @@ def load_panel_configs(dir):
             if f.endswith(".yaml"):
                 with open(str(Path(root).joinpath(f))) as file:
                     config = yaml.safe_load(file)
+                    # metric key can be None due to some metric tables not having any metrics
+                    # metric key should be empty dict instead of None
+                    for data_source in config["Panel Config"]["data source"]:
+                        metric_table = data_source.get("metric_table")
+                        if metric_table and metric_table["metric"] is None:
+                            metric_table["metric"] = {}
                     d[config["Panel Config"]["id"]] = config["Panel Config"]
 
     # TODO: sort metrics as the header order in case they are not defined in the same order

projects/rocprofiler-compute/src/utils/tty.py

@@ -114,6 +114,34 @@ def show_all(args, runs, archConfigs, output, profiling_config, roof_plot=None):
                     show_roof_plot(roof_plot)
                     continue
 
+                # Metrics baseline comparison mode
+                # We cannot guarantee that all runs have the same metrics. Only show common metrics.
+                if (
+                    type == "metric_table"
+                    and "Metric" in table_config["header"].values()
+                    and len(runs) > 1
+                ):
+                    # Common metrics across all runs
+                    common_metrics = set()
+                    for _, data in runs.items():
+                        if not common_metrics:
+                            common_metrics = set(data.dfs[table_config["id"]]["Metric"])
+                        else:
+                            common_metrics &= set(data.dfs[table_config["id"]]["Metric"])
+                    # Apply common metrics across all runs
+                    # Reindex all runs based on first run
+                    initial_index = None
+                    for key in runs.keys():
+                        runs[key].dfs[table_config["id"]] = (
+                            runs[key]
+                            .dfs[table_config["id"]]
+                            .loc[lambda d: d["Metric"].isin(common_metrics)]
+                        )
+                        if initial_index is None:
+                            initial_index= runs[key].dfs[table_config["id"]].index
+                        else:
+                            runs[key].dfs[table_config["id"]].index = initial_index
+
                 # take the 1st run as baseline
                 base_run, base_data = next(iter(runs.items()))
                 base_df = base_data.dfs[table_config["id"]]