ROCm · vedithal-amd · Jan 20, 2026 · Jan 2, 2026 · Jan 5, 2026 · Jan 5, 2026
@@ -49,6 +49,8 @@ Full documentation for ROCm Compute Profiler is available at [https://rocm.docs.
 
 * Fix issue where counter collection data was empty when profiling workload which spawn multiple child processes
 
+* Fix issue where dispatch filtering in a range (e.g. >2) was not working
+
 * Fix redundant warnings for compute/memory partition not found for < MI 300 series GPUs by skipping partition checks
 
 ### Removed

@@ -346,6 +346,7 @@ Show System Speed-of-Light and CS_Busy blocks only
    this case, ``1`` is the ID for System Speed-of-Light and ``5.1.0`` the ID for
    GPU Busy Cycles metric.
 
+
 Filter kernels
   First, list the top kernels in your application using `--list-stats`.
 
@@ -562,12 +563,6 @@ WARNING Failed to evaluate expression for 3.1.39 - Value: to_round((to_avg(
    (pmc_df.get("pmc_perf_ACCUM") / pmc_df.get("SQC_ICACHE_REQ")).where((pmc_df.get("SQC_ICACHE_REQ") != 0), None)) * 100), 0) - unsupported operand type(s) for /: 'NoneType' and 'float'
 WARNING Failed to evaluate expression for 3.1.39 - Value: to_round((to_avg(
    (pmc_df.get("pmc_perf_ACCUM") / pmc_df.get("SQC_ICACHE_REQ")).where((pmc_df.get("SQC_ICACHE_REQ") != 0), None)) * 100), 0) - unsupported operand type(s) for /: 'NoneType' and 'float'
-  DEBUG Calculating expressions for kernel: _ZN12rocrand_impl6system6detail17trampoline_kernelINS_4host33static_block_size_config_providerILj256EEEjLb0ELNS3_11target_archE942EZNS3_25xorwow_generator_templateINS0_13device_systemENS3_23default_config_providerIL16rocrand_rng_type401EEEE4initEvEUlT_DpT0_E_JPN14rocrand_device13xorwow_engineEjjyyEEEvT3_DpT4_
-  DEBUG Calculating expressions for kernel: _ZN12rocrand_impl6system6detail17trampoline_kernelINS_4host23default_config_providerIL16rocrand_rng_type401EEEfLb0ELNS3_11target_archE942EZZNS3_25xorwow_generator_templateINS0_13device_systemES6_E8generateIfNS3_20uniform_distributionIfjEEEE14rocrand_statusPT_mT0_ENKUlSF_E_clISt17integral_constantIbLb0EEEEDaSF_EUlSF_DpT0_E_JPN14rocrand_device13xorwow_engineEjPfmSD_EEEvT3_DpT4_
-  DEBUG Calculating expressions for kernel: void bodyForce_block<256>(HIP_vector_type<float, 4u> const*, HIP_vector_type<float, 4u>*, float, int)
-  DEBUG Calculating expressions for kernel: _ZN12rocrand_impl6system6detail17trampoline_kernelINS_4host33static_block_size_config_providerILj256EEEjLb0ELNS3_11target_archE942EZNS3_25xorwow_generator_templateINS0_13device_systemENS3_23default_config_providerIL16rocrand_rng_type401EEEE4initEvEUlT_DpT0_E_JPN14rocrand_device13xorwow_engineEjjyyEEEvT3_DpT4_
-  DEBUG Calculating expressions for kernel: _ZN12rocrand_impl6system6detail17trampoline_kernelINS_4host23default_config_providerIL16rocrand_rng_type401EEEfLb0ELNS3_11target_archE942EZZNS3_25xorwow_generator_templateINS0_13device_systemES6_E8generateIfNS3_20uniform_distributionIfjEEEE14rocrand_statusPT_mT0_ENKUlSF_E_clISt17integral_constantIbLb0EEEEDaSF_EUlSF_DpT0_E_JPN14rocrand_device13xorwow_engineEjPfmSD_EEEvT3_DpT4_
-  DEBUG Calculating expressions for kernel: void bodyForce_block<256>(HIP_vector_type<float, 4u> const*, HIP_vector_type<float, 4u>*, float, int)
   DEBUG Calculated metric values
   DEBUG Calculated roofline data points
   DEBUG [analysis] generating analysis

@@ -28,7 +28,9 @@ Launch the standalone GUI analyzer
 ----------------------------------
 
 To launch the ROCm Compute Profiler GUI analyzer, include the ``--gui`` flag with your
-desired analysis command. For example:
+desired analysis command.
+
+For example:
 
 .. code-block:: shell-session
 

@@ -386,26 +386,24 @@ def sanitize(self) -> None:
             sys.exit(0)
 
         # Ensure analysis output does not overwrite existing files
-        if not args.output_name:
-            return
-
-        if not re.match(r"^[A-Za-z0-9_-]+$", args.output_name):
-            console_error(
-                "analysis",
-                "Analysis output file/folder name must "
-                "contain only alphanumeric characters "
-                "or underscores (_), hyphens (-).",
-            )
+        if args.output_name:
+            if not re.match(r"^[A-Za-z0-9_-]+$", args.output_name):
+                console_error(
+                    "analysis",
+                    "Analysis output file/folder name must "
+                    "contain only alphanumeric characters "
+                    "or underscores (_), hyphens (-).",
+                )
 
-        path_to_check = args.output_name
-        if args.output_format in ("txt", "db"):
-            path_to_check += f".{args.output_format}"
+            path_to_check = args.output_name
+            if args.output_format in ("txt", "db"):
+                path_to_check += f".{args.output_format}"
 
-        if Path(path_to_check).exists():
-            console_error(
-                f"Analysis output file/folder {path_to_check} already exists. "
-                "Please choose a different name."
-            )
+            if Path(path_to_check).exists():
+                console_error(
+                    f"Analysis output file/folder {path_to_check} already exists. "
+                    "Please choose a different name."
+                )
 
         # Check if any kernel's counters are missing due to iteration multiplexing
         if (

@@ -51,7 +51,6 @@ def __init__(
         self.app = dash.Dash(
             __name__, title=PROJECT_NAME, external_stylesheets=[dbc.themes.CYBORG]
         )
-        self.dest_dir = str(Path(args.path[0][0]).absolute().resolve())
         self.arch: Optional[str] = None
 
         self.__hidden_sections = ["Memory Chart"]
@@ -90,6 +89,7 @@ def build_layout(
         kernel_top_df = base_data.dfs[1]
         for kernel_id in base_data.filter_kernel_ids:
             filt_kernel_names.append(str(kernel_top_df.loc[kernel_id, "Kernel_Name"]))
+        input_filters["kernel"] = filt_kernel_names
 
         # setup app layout
         from utils.gui_components.header import get_header
@@ -338,6 +338,7 @@ def pre_processing(self) -> None:
             )
 
         args = self.get_args()
+        self.dest_dir = str(Path(args.path[0][0]).absolute().resolve())
 
         # create 'mega dataframe'
         self._runs[self.dest_dir].raw_pmc = file_io.create_df_pmc(

@@ -61,7 +61,7 @@ def multi_bar_chart(
 
 def create_instruction_mix_bar_chart(display_df: pd.DataFrame, df_unit: str) -> px.bar:
     display_df = display_df.copy()
-    display_df["Avg"] = display_df["Avg"].apply(lambda x: int(x) if x != "" else 0)
+    display_df["Avg"] = display_df["Avg"].apply(lambda x: int(x) if x != "N/A" else 0)
 
     return px.bar(
         display_df,
@@ -78,7 +78,7 @@ def create_multi_bar_charts(
     display_df: pd.DataFrame, table_id: int, df_unit: str
 ) -> list[px.bar]:
     display_df = display_df.copy()
-    display_df["Avg"] = display_df["Avg"].apply(lambda x: int(x) if x != "" else 0)
+    display_df["Avg"] = display_df["Avg"].apply(lambda x: int(x) if x != "N/A" else 0)
 
     nested_bar = multi_bar_chart(table_id, display_df)
     charts = []
@@ -103,7 +103,9 @@ def create_multi_bar_charts(
 
 def create_sol_charts(display_df: pd.DataFrame, table_id: int) -> list[px.bar]:
     display_df = display_df.copy()
-    display_df["Avg"] = display_df["Avg"].apply(lambda x: float(x) if x != "" else 0.0)
+    display_df["Avg"] = display_df["Avg"].apply(
+        lambda x: float(x) if x != "N/A" else 0.0
+    )
 
     charts = []
 
@@ -144,7 +146,7 @@ def create_sol_charts(display_df: pd.DataFrame, table_id: int) -> list[px.bar]:
     elif table_id == 1101:
         # Special formatting reference 'Pct of Peak' value
         display_df["Pct of Peak"] = display_df["Pct of Peak"].apply(
-            lambda x: float(x) if x != "" else 0.0
+            lambda x: float(x) if x != "N/A" else 0.0
         )
         charts.append(
             px.bar(

@@ -1290,14 +1290,16 @@ def apply_dispatch_filter(df: pd.DataFrame, workload: schema.Workload) -> pd.Dat
     # NB: support ignoring the 1st n dispatched execution by '> n'
     #     The better way may be parsing python slice string
     for dispatch_id in workload.filter_dispatch_ids:
+        if isinstance(dispatch_id, str) and ">" in dispatch_id:
+            dispatch_id = re.match(r"\>\s*(\d+)", dispatch_id).group(1)
         if int(dispatch_id) >= len(df):  # subtract 2 bc of the two header rows
             console_error("analysis", f"{dispatch_id} is an invalid dispatch id.")
 
     if (
         isinstance(workload.filter_dispatch_ids[0], str)
         and ">" in workload.filter_dispatch_ids[0]
     ):
-        dispatch_match = re.match(r"\> (\d+)", workload.filter_dispatch_ids[0])
+        dispatch_match = re.match(r"\>\s*(\d+)", workload.filter_dispatch_ids[0])
         df = df[
             df[schema.PMC_PERF_FILE_PREFIX]["Dispatch_ID"]
             > int(dispatch_match.group(1))