8147
2021
eng
277
288
conferenceobject
0
--
--
--
Combining XOR and Partner Checkpointing for Resilient Multilevel Checkpoint/Restart
2021 IEEE International Parallel and Distributed Processing Symposium (IPDPS)
10.1109/IPDPS49936.2021.00036
yes
2021-02-07
publish
Masoud Gholami
Florian Schintke
Florian Schintke
Distributed Algorithms and Supercomputing
Schintke, Florian
2013-FFMK
2014-BBDC-WP12
Gholami Estahbanati, Masoud
7068
2018
eng
poster
0
--
--
--
Modeling Checkpoint Schedules for Concurrent HPC Applications
CoSaS 2018 International Symposium on Computational Science at Scale
yes
Masoud Gholami
Florian Schintke
Florian Schintke
Thorsten Schütt
Alexander Reinefeld
Distributed Algorithms and Supercomputing
Distributed Data Management
Reinefeld, Alexander
Schintke, Florian
Schütt, Thorsten
2013-FFMK
Gholami Estahbanati, Masoud
7539
2019
eng
143
152
conferenceobject
0
--
--
--
Multilevel Checkpoint/Restart for Large Computational Jobs on Distributed Computing Resources
2019 IEEE 38th Symposium on Reliable Distributed Systems (SRDS)
10.1109/SRDS47363.2019.00025
yes
Masoud Gholami
Florian Schintke
Florian Schintke
Distributed Algorithms and Supercomputing
Schintke, Florian
2013-FFMK
Gholami Estahbanati, Masoud
8712
2022
eng
109
119
2
22
article
Springer Berlin Heidelberg
0
--
--
--
IOSIG: Declarative I/O-Stream Properties Using Pragmas
Datenbank-Spektrum
10.1007/s13222-022-00419-w
yes
2022-05-27
publish
https://link.springer.com/content/pdf/10.1007/s13222-022-00419-w.pdf
Masoud Gholami
Schintke
Florian Schintke
Distributed Algorithms and Supercomputing
Schintke, Florian
Gholami Estahbanati, Masoud
2020-FONDA
7025
2018
eng
44:1
44:10
conferenceobject
0
--
--
--
Checkpoint Scheduling for Shared Usage of Burst-Buffers in Supercomputers
User-defined and system-level checkpointing have contrary properties. While user-defined checkpoints are smaller and simpler to recover, system-level checkpointing better knows the global system's state and parameters like the expected mean time to failure (MTTF) per node. Both approaches lead to non-optimal checkpoint time, intervals, sizes, or I/O bandwidth when concurrent checkpoints conflict and compete for it.
We combine user-defined and system-level checkpointing to exploit the benefits and avoid the drawbacks of each other. Thus, applications frequently offer to create checkpoints. The system accepts such offers according to the current status and implied costs to recalculate from the last checkpoint or denies them, i.e., immediately lets continue the application without checkpoint creation. To support this approach, we develop economic models for multi-application checkpointing on shared I/O resources that are dedicated for checkpointing (e.g. burst-buffers) by defining an appropriate goal function and solving a global optimization problem.
Using our models, the checkpoints of applications on a supercomputer are scheduled to effectively use the available I/O bandwidth and minimize the failure overhead (checkpoint creations plus recalculations). Our simulations show an overall reduction in failure overhead of all nodes of up to 30% for a typical supercomputer workload (HLRN). We can also derive the most cost effective burst-buffer bandwidth for a given node's MTTF and application workload.
Proceedings of the 47th International Conference on Parallel Processing Companion; SRMPDS 2018: The 14th International Workshop on Scheduling and Resource Management for Parallel and Distributed Systems
10.1145/3229710.3229755
yes
Masoud Gholami
Florian Schintke
Florian Schintke
Thorsten Schütt
Distributed Algorithms and Supercomputing
Schintke, Florian
Schütt, Thorsten
2013-FFMK
Gholami Estahbanati, Masoud
7699
2019
eng
483
516
conferenceobject
Springer
0
--
--
--
FFMK: A Fast and Fault-Tolerant Microkernel-Based System for Exascale Computing
Software for Exascale Computing - SPPEXA 2016-2019
10.1007/978-3-030-47956-5_16
yes
2019-10-07
Carsten Weinhold
Florian Schintke
Adam Lackorzynski
Jan Bierbaum
Martin Küttler
Maksym Planeta
Hannes Weisbach
Matthias Hille
Hermann Härtig
Alexander Margolin
Dror Sharf
Ely Levy
Pavel Gak
Amnon Barak
Masoud Gholami
Florian Schintke
Thorsten Schütt
Alexander Reinefeld
Matthias Lieber
Wolfgang Nagel
Distributed Algorithms and Supercomputing
Reinefeld, Alexander
Schintke, Florian
Schütt, Thorsten
2013-FFMK
Gholami Estahbanati, Masoud
9221
2023
eng
1
22
article
0
--
--
--
CAWL: A Cache-aware Write Performance Model of Linux Systems
arXiv
10.48550/arXiv.2306.05701
2306.05701
no
publish
Masoud Gholami
Florian Schintke
Florian Schintke
Distributed Algorithms and Supercomputing
Schintke, Florian
Gholami Estahbanati, Masoud
2020-FONDA