You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/connectors/ScoreP.md
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -21,7 +21,7 @@ ScoreP is particularly valuable for:
21
21
- Tracing communication and synchronization in distributed applications
22
22
- Performance tuning of HPC applications
23
23
24
-
Note: While ScoreP provides advanced tracing capabilities from a third-party library, it focuses on profiling and tracing rather than intelligent analysis or automatic tuning (unlike tools such as APEX).
24
+
Note: While ScoreP provides advanced tracing capabilities as a third-party library, it focuses on profiling and tracing rather than intelligent analysis or automatic tuning (unlike tools such as APEX).
Output is located in `timemory-output/<DATE-TIME>` unless `KOKKOS_ROOFLINE` is set. The `<DATE-TIME>` format can be customized via `TIMEMORY_TIME_FORMAT` (default: `"%F_%I.%M_%p"`).
131
+
Output is located in `timemory-output/<DATE-TIME>` by default. The `<DATE-TIME>` format can be customized via `TIMEMORY_TIME_FORMAT` (default: `"%F_%I.%M_%p"`).
132
+
133
+
When using roofline analysis with `KOKKOS_ROOFLINE=ON`, the output location is controlled by the `TIMEMORY_OUTPUT_PATH` environment variable as shown in the roofline example above.
0 commit comments