diff options
| author | Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org> | 2024-01-02 03:33:39 +0530 |
|---|---|---|
| committer | Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org> | 2024-01-10 18:50:49 +0530 |
| commit | 146fbc595f5ae1f904abfee37af7731f29f05941 (patch) | |
| tree | 6d175e141b4b1df241528cfdfa15c3ac53ebf832 /gather-metrics.sh | |
| parent | 153e2cc3233c2d19cae2a379edbcdf7277715f7e (diff) | |
fixup-perf-csv.py: Merge status info in perf.csv and remove symbol entries for failed bmks.
The rationale for this patch is to avoid issues like LLVM-1070 in which metric
and comparison scripts still process symbols for failed benchmarks resulting in
spurious regressions. The patch simply removes any symbol entries for failed bmks,
since IMHO we should only care about bmk profile if it's run successfully.
Change-Id: I198dc15333d3bcd13a1c3e25e8d75009619978ed
Diffstat (limited to 'gather-metrics.sh')
| -rwxr-xr-x | gather-metrics.sh | 5 |
1 files changed, 2 insertions, 3 deletions
diff --git a/gather-metrics.sh b/gather-metrics.sh index 026470c..e15a2f3 100755 --- a/gather-metrics.sh +++ b/gather-metrics.sh @@ -52,7 +52,8 @@ function gather_perf_data () fi fi - $bmk_scripts/fixup-perf-csv.py "$csv_results_dir/perf-tmp.csv" > "$csv_results_dir/perf.csv" + $bmk_scripts/fixup-perf-csv.py "$csv_results_dir/perf-tmp.csv" "$csv_results_dir/status.csv" \ + > "$csv_results_dir/perf.csv" rm "$csv_results_dir/perf-tmp.csv" if [ "$(cat $csv_results_dir/perf.csv | wc -l)" == "1" ]; then @@ -137,8 +138,6 @@ for metric in "${metrics[@]}"; do merge_opts+=("$csv_results_dir/$metric.csv") done -merge_opts+=("$csv_results_dir/status.csv") - # Merge all metric csvs $bmk_scripts/merge-metric-csvs.py "${merge_opts[@]}" \ > "$csv_results_dir/results.csv" |