Basic log parsing tool

Add numpy to setup

geosongpu_ci/pipeline/held_suarez.py

@@ -8,7 +8,8 @@
 from geosongpu_ci.pipeline.geos import set_python_environment
 from geosongpu_ci.pipeline.gtfv3_config import GTFV3Config
 from geosongpu_ci.utils.progress import Progress
-from geosongpu_ci.tools.benchmark import parse_geos_log, report
+from geosongpu_ci.tools.benchmark.geos_log_parser import parse_geos_log
+from geosongpu_ci.tools.benchmark.report import report
 from typing import Dict, Any, Optional
 import shutil
 import os
@@ -361,10 +362,13 @@ def check(
                 bench_raw_data = []
                 for log in logs:
                     shutil.copy(log, benchmark_artifact)
-                    if ".0.out" in log:
+                    # Grab all rank 0 that are not caching runs
+                    if ".0.out" in log and "cache" not in log:
                         bench_raw_data.append(parse_geos_log(log))
                 benchmark_report = report(bench_raw_data)
                 print(benchmark_report)
+                with open(f"{benchmark_artifact}/report_benchmark.out", "w") as f:
+                    f.write(str(benchmark_report))
 
         return True
 

geosongpu_ci/tools/benchmark/__init__.py

@@ -1,2 +0,0 @@
-from .geos_log_parser import geos_log_parser
-from .report import report, BenchmarkReport

geosongpu_ci/tools/benchmark/geos_log_parser.py

@@ -1,4 +1,4 @@
-from geosongpu_ci.tools.benchmark import BenchmarkRawData
+from geosongpu_ci.tools.benchmark.raw_data import BenchmarkRawData
 from typing import Iterable, Optional
 import re
 
@@ -50,11 +50,12 @@ def parse_geos_log(filename: str) -> BenchmarkRawData:
         benchmark.backend = "fortran"
     else:
         backend_pattern = "backend: "
-        benchmark.backend = (
-            _grep(filename, backend_pattern, exclude_pattern=True)[0]
-            .strip()
-            .replace("\n", "")
-        )
+        grepped = _grep(filename, backend_pattern, exclude_pattern=True)
+        if grepped == []:
+            benchmark.backend = "gtfv3 (details failed to parse)"
+        else:
+            backend = grepped[0].strip().replace("\n", "")
+            benchmark.backend = f"gtfv3: ({backend})"
 
     # Get timings of FV
     if is_gtfv3:
@@ -118,9 +119,10 @@ def parse_geos_log(filename: str) -> BenchmarkRawData:
                 end_pattern=superdyn_profiler_exit,
             )
         )
-        benchmark.fv_gridcomp_detailed_profiling.append(
-            (shortname, measures[4], parent)
-        )
+        if measures != []:
+            benchmark.fv_gridcomp_detailed_profiling.append(
+                (shortname, measures[4], parent)
+            )
 
     # Model throughput
     gloabl_profiler_entry = "profiler: Model Throughput"

geosongpu_ci/tools/benchmark/raw_data.py

@@ -0,0 +1,15 @@
+from dataclasses import dataclass, field
+from typing import Tuple, Iterable, List
+
+
+@dataclass
+class BenchmarkRawData:
+    backend: str = ""
+    grid_resolution: Tuple[int, int, int] = (0, 0, 0)  # nx / ny / nz
+    node_setup: Tuple[int, int, int] = (0, 0, 0)  # NX / NY / Total ranks used
+    global_run_time: float = 0  # seconds fort the global RUN
+    fv_dyncore_timings: Iterable[float] = field(default_factory=list)  # seconds
+    inner_dycore_timings: Iterable[float] = field(default_factory=list)  # seconds
+    fv_gridcomp_detailed_profiling: List[Tuple[str, float]] = field(
+        default_factory=list
+    )

geosongpu_ci/tools/benchmark/report.py

@@ -1,20 +1,8 @@
 import itertools
-from typing import Any, Iterable, Tuple, List
+from typing import Any, List
 from dataclasses import dataclass, field
 import numpy as np
-
-
-@dataclass
-class BenchmarkRawData:
-    backend: str = ""
-    grid_resolution: Tuple[int, int, int] = (0, 0, 0)  # nx / ny / nz
-    node_setup: Tuple[int, int, int] = (0, 0, 0)  # NX / NY / Total ranks used
-    global_run_time: float = 0  # seconds fort the global RUN
-    fv_dyncore_timings: Iterable[float] = field(default_factory=list)  # seconds
-    inner_dycore_timings: Iterable[float] = field(default_factory=list)  # seconds
-    fv_gridcomp_detailed_profiling: List[Tuple[str, float]] = field(
-        default_factory=list
-    )
+from geosongpu_ci.tools.benchmark.raw_data import BenchmarkRawData
 
 
 @dataclass

setup.py

@@ -19,6 +19,7 @@
     install_requires=[
         "pyyaml",
         "click",
+        "numpy",
     ],
     data_files=[
         ("./geosongpu/experiments", ["./experiments/experiments.yaml"]),