Bandwidth-optimal all-to-all exchanges in fat tree networks

Prisacari, Bogdan; Rodríguez, Germán; Minkenberg, Cyriel; Hoefler, Torsten

doi:10.1145/2464996.2465434

Cited by 37 publications

(24 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By default, collective operations are included in the communication graph by breaking down each collective operation to point-to-point messages [33], [34]. Some MPI implementations optimize their collective operations based on a default placement or without regard to placement [35].…”

Section: E Collective Operationsmentioning

confidence: 99%

APHiD: Hierarchical Task Placement to Enable a Tapered Fat Tree Topology for Lower Power and Cost in HPC Networks

Ibrahim

Shalf

Wilke

et al. 2017

2017 17th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGRID)

View full text Add to dashboard Cite

Abstract-The power and procurement cost of bandwidth in system-wide networks has forced a steady drop in the byte/flop ratio. This trend of computation becoming faster relative to the network is expected to hold. In this paper, we explore how cost-oriented task placement enables reducing the cost of system-wide networks by enabling high performance even on tapered topologies where more bandwidth is provisioned at lower levels. We describe APHiD, an efficient hierarchical placement algorithm that uses new techniques to improve the quality of heuristic solutions and reduces the demand on high-level, expensive bandwidth in hierarchical topologies. We apply APHiD to a tapered fat-tree, demonstrating that APHiD maintains application scalability even for severely tapered network configurations. Using simulation, we show that for tapered networks APHiD improves performance by more than 50% over random placement and even 15% in some cases over costlier, state-of-the-art placement algorithms.

show abstract

Section: E Collective Operationsmentioning

confidence: 99%

APHiD: Hierarchical Task Placement to Enable a Tapered Fat Tree Topology for Lower Power and Cost in HPC Networks

Ibrahim

Shalf

Wilke

et al. 2017

2017 17th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGRID)

View full text Add to dashboard Cite

show abstract

“…This exchange pattern is similar to the commonly used linear exchange pattern [29], except for the fact that it tries to optimize the traffic for full-duplex links. …”

Section: )mentioning

confidence: 99%

Fail-in-Place Network Design: Interaction Between Topology, Routing Algorithm and Failures

Domke

Hoefler

Matsuoka

2014

SC14: International Conference for High Performance Computing, Networking, Storage and Analysis

Self Cite

View full text Add to dashboard Cite

Abstract-The growing system size of high performance computers results in a steady decrease of the mean time between failures. Exchanging network components often requires whole system downtime which increases the cost of failures. In this work, we study a fail-in-place strategy where broken network elements remain untouched. We show, that a fail-in-place strategy is feasible for todays networks and the degradation is manageable, and provide guidelines for the design. Our network failure simulation toolchain allows system designers to extrapolate the performance degradation based on expected failure rates, and it can be used to evaluate the current state of a system. In a case study of real-world HPC systems, we will analyze the performance degradation throughout the systems lifetime under the assumption that faulty network components are not repaired, which results in a recommendation to change the used routing algorithm to improve the network performance as well as the fail-in-place characteristic.

show abstract

“…This work was originally triggered by a very specific example, described in detail in Prisacari et al [2013]. There, we tackled the problem of optimizing the communication pattern of the all-to-all exchange collective operation, encountered in many parallel programming models, for XGFT networks with less than full bisection bandwidth.…”

Section: Motivationmentioning

confidence: 99%

“…In a recent work [Prisacari et al 2013], we addressed all-to-all optimization in the context of fat tree networks. We considered a case where each leaf node of an XGFT network is assigned a single task.…”

Section: Practical Use Casementioning

confidence: 99%

“…The cause was the fact that the routing algorithms were not able to sustain the theoretically ideal workload performance. Prisacari et al [2013] also derived the minimum amount of bandwidth required at each fat tree layer to ensure contention-free traffic in each exchange phase. The higher the layer, the more potential for bandwidth reduction exists compared to the full-bisection design.…”

Section: Practical Use Casementioning

confidence: 99%

See 1 more Smart Citation

Fast pattern-specific routing for fat tree networks

Prisacari

Rodríguez

Minkenberg

et al. 2013

TACO

Self Cite

View full text Add to dashboard Cite

In the context of eXtended Generalized Fat Tree (XGFT) topologies, widely used in HPC and datacenter network designs, we propose a generic method, based on Integer Linear Programming (ILP), to efficiently determine optimal routes for arbitrary workloads. We propose a novel approach that combines ILP with dynamic programming, effectively reducing the time to solution. Specifically, we divide the network into smaller subdomains optimized using a custom ILP formulation that ensures global optimality of local solutions. Local solutions are then combined into an optimal global solution using dynamic programming. Finally, we demonstrate through a series of extensive benchmarks that our approach scales in practice to networks interconnecting several thousands of nodes, using a single-threaded, freely available linear programming solver on commodity hardware, with the potential for higher scalability by means of commercial, parallel solvers.

show abstract

Bandwidth-optimal all-to-all exchanges in fat tree networks

Cited by 37 publications

References 29 publications

APHiD: Hierarchical Task Placement to Enable a Tapered Fat Tree Topology for Lower Power and Cost in HPC Networks

APHiD: Hierarchical Task Placement to Enable a Tapered Fat Tree Topology for Lower Power and Cost in HPC Networks

Fail-in-Place Network Design: Interaction Between Topology, Routing Algorithm and Failures

Fast pattern-specific routing for fat tree networks

Contact Info

Product

Resources

About