Scenario 210

Creator	wujj123456
When	12 Aug 2024
Description	Per-app concurrency limit prevents work fetch and causes cores to be idle even when limit has not been reached
Input files	client_state.xml global_prefs.xml global_prefs_override.xml cc_config.xml projects/lhcathome.cern.ch_lhcathome/app_config.xml

Simulations

ID Click for details	Who	When	Parameters	Results	Comments
1	David Anderson	13 Aug 2024	Existing jobs only: no Round-robin only: no scheduler EDF sim: no Include empty projects: no REC half-life: 864000.000000 Simulation duration: 3600.000000 Time step: 60.000000	wasted fraction 0.000000 Idle fraction 0.882237 Share violation 1.407407 Monotony 0.015704 RPCs per job 0.500000
0	wujj123456	12 Aug 2024	Existing jobs only: no Round-robin only: no scheduler EDF sim: no Include empty projects: no REC half-life: 864000.000000 Simulation duration: 1800.000000 Time step: 60.000000	wasted fraction 0.000000 Idle fraction 0.967286 Share violation 1.407407 Monotony 0.008337 RPCs per job 0.000000	By wujj123456 (12 Aug 2024): From the log, LHC only started the two existing tasks, but the fetch logic never picked it to fetch more after multiple evaluations. This is when the system has shortfall of work, 12 cpus idling and the ATLAS app of LHC hasn't reached its concurrent limit of 4.

ID
Click for details

Who

When

Parameters

Results

Comments

David Anderson

13 Aug 2024

Existing jobs only: no
Round-robin only: no
scheduler EDF sim: no
Include empty projects: no
REC half-life: 864000.000000
Simulation duration: 3600.000000
Time step: 60.000000

wasted fraction 0.000000
Idle fraction 0.882237
Share violation 1.407407
Monotony 0.015704
RPCs per job 0.500000

wujj123456

12 Aug 2024

Existing jobs only: no
Round-robin only: no
scheduler EDF sim: no
Include empty projects: no
REC half-life: 864000.000000
Simulation duration: 1800.000000
Time step: 60.000000

wasted fraction 0.000000
Idle fraction 0.967286
Share violation 1.407407
Monotony 0.008337
RPCs per job 0.000000

By wujj123456 (12 Aug 2024):
From the log, LHC only started the two existing tasks, but the fetch logic never picked it to fetch more after multiple evaluations. This is when the system has shortfall of work, 12 cpus idling and the ATLAS app of LHC hasn't reached its concurrent limit of 4.

Copyright © 2025 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.