An integrated tutorial on InfiniBand, verbs, and MPI
P MacArthur, Q Liu, RD Russell… - … Surveys & Tutorials, 2017 - ieeexplore.ieee.org
This tutorial presents the details of the interconnection network utilized in many high
performance computing (HPC) systems today.“InfiniBand” is the hardware interconnect …
performance computing (HPC) systems today.“InfiniBand” is the hardware interconnect …
Throttling for bandwidth imbalanced data transfers
Techniques are disclosed to throttle bandwidth imbalanced data transfers. In some
examples, an example computer-implemented method may include splitting a payload of a …
examples, an example computer-implemented method may include splitting a payload of a …
Toward lower-diameter large-scale HPC and data center networks with co-packaged optics
We investigate the advantages of using co-packaged optics for building low-diameter, large-
scale high-performance computing (HPC) and data center networks. The increased escape …
scale high-performance computing (HPC) and data center networks. The increased escape …
[PDF][PDF] Solving hot spot contention using infiniband architecture congestion control
G Pfister, M Gusat, W Denzel, D Craddock… - Proceedings HP-IPC …, 2005 - researchgate.net
Since at least 1985 [1] it has been known that certain traffic patterns in multistage
interconnection networks, hot spots, can cause catastrophic congestion and loss of …
interconnection networks, hot spots, can cause catastrophic congestion and loss of …
Revisiting Congestion Control for Lossless Ethernet
Y Zhang, Q Meng, C Hu, F Ren - 21st USENIX Symposium on …, 2024 - usenix.org
Congestion control is a key enabler for lossless Ethernet at scale. In this paper, we revisit
this classic topic from a new perspective, ie, understanding and exploiting the intrinsic …
this classic topic from a new perspective, ie, understanding and exploiting the intrinsic …
A new proposal to deal with congestion in InfiniBand-based fat-trees
The overall performance of High-Performance Computing applications may depend largely
on the performance achieved by the network interconnecting the end-nodes; thus high …
on the performance achieved by the network interconnecting the end-nodes; thus high …
Exploration of congestion control techniques on dragonfly-class hpc networks through simulation
N McGlohon, CD Carothers… - … and Simulation of …, 2021 - ieeexplore.ieee.org
Ensuring optimal communication latency in High Performance Computing (HPC) networks is
of critical importance to the efficient operation of facilitated applications. Different application …
of critical importance to the efficient operation of facilitated applications. Different application …
Latency and throughput optimization in modern networks: a comprehensive survey
A Mirzaeinnia, M Mirzaeinia, A Rezgui - arXiv preprint arXiv:2009.03715, 2020 - arxiv.org
Modern applications are highly sensitive to communication delays and throughput. This
paper surveys major attempts on reducing latency and increasing the throughput. These …
paper surveys major attempts on reducing latency and increasing the throughput. These …
Noise injection techniques to expose subtle and unintended message races
Debugging intermittently occurring bugs within MPI applications is challenging, and
message races, a condition in which two or more sends race to match with a receive, are …
message races, a condition in which two or more sends race to match with a receive, are …
Hot-spot avoidance with multi-pathing over infiniband: An mpi perspective
Large scale InfiniBand clusters are becoming increasingly popular, as reflected by the TOP
500 supercomputer rankings. At the same time, fat tree has become a popular …
500 supercomputer rankings. At the same time, fat tree has become a popular …