Changeset 29c6a7d for doc/theses/mike_brooks_MMath/plots/ListCommon.py

Timestamp:

Jul 26, 2025, 3:04:16 PM (2 months ago)

Author:

Michael Brooks <mlbrooks@…>

Branches:

master

Children:

f449d1e

Parents:

7806f91

Message:

Use plots-based data crunching for quick analysis of ad-hoc testing, supporting tailq data not being present

File:

• doc/theses/mike_brooks_MMath/plots/ListCommon.py (modified) (1 diff)

doc/theses/mike_brooks_MMath/plots/ListCommon.py

@@ -57 +57 @@
     byPeer = timings.groupby(['NumNodes', 'op', 'InterleaveFrac'])
     for [NumNodes, op, intrlFrac], peerGroup in byPeer:
-        baselineRows = peerGroup.groupby(['fx']).get_group(baseline_fx)
-        baselineDur = meanNoOutlr( baselineRows['mean_op_dur_ns'] )
+        grpfx = peerGroup.groupby(['fx'])
+        if baseline_fx in grpfx.groups:
+            baselineRows = grpfx.get_group(baseline_fx)
+            baselineDur = meanNoOutlr( baselineRows['mean_op_dur_ns'] )
+        else:
+            baselineDur = 1.0
         timings.loc[peerGroup.index, 'BaselineFxOpDurNs'] = baselineDur
     timings['OpDurRelFx'] = timings['mean_op_dur_ns'] / timings['BaselineFxOpDurNs']