Wire breakdown.csv into feedsim_autoscale_dlrm + add preprocessing/postprocessing logs (#736)

Summary:
Pull Request resolved: #736

Three related changes so perfpub can identify the actual benchmarking phase and report metrics filtered to that window (instead of averaging across warmup + drain + idle periods).

1. jobs.yml: add `benchmarks/feedsim/breakdown.csv` to the copymove `after:` list for feedsim_autoscale_dlrm. The other feedsim jobs (search-mode, mini, etc.) already had it; only the DLRM autoscale variant was missing. With this, perfpub's automatic breakdown-based metric filtering kicks in for `./perfpub --no-xdb --no-manifold --dir benchmark_metrics_<id>` (see perfpub/README.md "Automatic Metric Filtering with breakdown.csv").

2. run.sh: add preprocessing/postprocessing log entries around the existing search_qps.sh main_benchmark window. search_qps.sh already logs main_benchmark start/end around `sleep $experiment_time` (the actual measurement window); this commit adds the surrounding phase info so the full timeline is visible in breakdown.csv:
   - preprocessing  = run.sh start → driver launch (server bring-up, graph/model load, warmup)
   - main_benchmark = experiment_time window (unchanged)
   - postprocessing = driver exit → run.sh exit (queue drain, leaf shutdown)
   perfpub uses only main_benchmark for default filtering; pre/post are informational for humans inspecting the CSV.

3. runtime_breakdown_utils.sh: make `create_breakdown_csv` idempotent (skip if file exists). Without this, multi-instance feedsim runs (run-feedsim-multi.sh launching N instances of run.sh) race on truncating the CSV — the second instance's create_breakdown_csv call would silently drop the first instance's already-logged entries. Verified t20 q=100/q=120 runs now record both inst1+inst2 preprocessing/main_benchmark/postprocessing entries cleanly. benchpress's copymove hook (is_move: true) moves the file out between iterations, so stale data across runs is not a concern.

Verified on rtptest3440 BGM (qps=100, 120 each at 180s/300s):
- breakdown.csv lands in benchmark_metrics_<id>/ after each run
- `./perfpub --no-xdb --no-manifold --dir benchmark_metrics_<id>` succeeds and emits overall-metrics.csv with the breakdown-filtered window
- All 4 expected operation entries present per instance: preprocessing start/end, main_benchmark start/end, postprocessing start/end (no race-induced drops)

Reviewed By: YifanYuan3

Differential Revision: D105756766

fbshipit-source-id: 7ce888f9d8e962a0c255c16992ead4ece23879a5

Author: excelle08

benchpress/config/jobs.yml

@@ -781,6 +781,7 @@
           - 'benchmarks/feedsim/feedsim_results*.txt'
           - 'benchmarks/feedsim/feedsim-multi-inst-*.log'
           - 'benchmarks/feedsim/src/perf.data'
+          - 'benchmarks/feedsim/breakdown.csv'
           - '/tmp/feedsim_log.txt'
 
 - name: feedsim_autoscale_dlrm_mini

packages/common/runtime_breakdown_utils.sh

@@ -25,6 +25,16 @@ create_breakdown_csv() {
 
     local csv_file="${folder_path}/${breakdown_file_name}"
 
+    # Idempotent: if the file already exists, leave it alone. Multi-instance
+    # workloads (e.g. feedsim run-feedsim-multi.sh) have several processes
+    # racing to call create_breakdown_csv concurrently — truncating after the
+    # first writer would silently drop entries already logged by the earlier
+    # instance. benchpress's copymove hook (is_move: true) clears the file
+    # between iterations, so stale data is not a risk.
+    if [ -f "$csv_file" ]; then
+        return 0
+    fi
+
     # Create CSV file with headers
     if echo "operation_name,PID,timestamp_type,timestamp,sub_operation_name" > "$csv_file"; then
         echo "Created breakdown CSV file: $csv_file"

packages/feedsim/run.sh

@@ -574,6 +574,12 @@ main() {
     fi
 
     create_breakdown_csv "$BREAKDOWN_FOLDER"
+    # Phase timeline (per-instance, shared CSV in $BREAKDOWN_FOLDER):
+    #   preprocessing   = run.sh start → driver launch (server bring-up, graph/model load, warmup)
+    #   main_benchmark  = the experiment_time window (logged by search_qps.sh around `sleep $experiment_time`)
+    #   postprocessing  = driver exit → run.sh exit (queue drain, leaf shutdown)
+    # perfpub reads only main_benchmark entries by default; pre/post are informational.
+    log_preprocessing_start "$BREAKDOWN_FOLDER" "$$"
 
     set -u  # Enable unbound variables check from here onwards
 
@@ -822,6 +828,9 @@ main() {
         no_retry_args="-N"
     fi
 
+    # Preprocessing complete; search_qps.sh will own the main_benchmark phase.
+    log_preprocessing_end "$BREAKDOWN_FOLDER" "$$"
+
     if [ -z "$fixed_qps" ] && [ "$auto_driver_threads" != "1" ]; then
         benchreps_tell_state "before search_qps"
         # shellcheck disable=SC2086
@@ -876,8 +885,10 @@ main() {
         benchreps_tell_state "after fixed_qps_exp"
     fi
 
+    log_postprocessing_start "$BREAKDOWN_FOLDER" "$$"
     sleep "$queue_drain_time" # wait for queue to drain
     kill -SIGINT $LEAF_PID || true > /dev/null # SIGINT so exits cleanly
+    log_postprocessing_end "$BREAKDOWN_FOLDER" "$$"
 }
 
 main "$@"