Performance Models for Split-Execution Computing Systems

Humble, Travis S.; McCaskey, Alexander; Schrock, Jonathan; Seddiqi, Hadayat; Britt, Keith A.; Imam, Neena

doi:10.1109/ipdpsw.2016.113

Cited by 5 publications

(5 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Data is archived locally and post-processed to generate the results presented in the next section. The results presented were collected in August and September of 2016 [27].…”

Section: Methodsmentioning

confidence: 99%

“…These steps include the time required to initialize the programmable magnetic memory (PMM) that is used as the control lines into the super-cooled processor. Technical details about the electronic programming process are available in the relevant literature [44], but it suffices to note that these steps contribute a near constant time cost to the total for the execution model [27].…”

Section: Experimental Validation Of Cam Recallmentioning

confidence: 99%

“…Notwithstanding the search for quantum speedups, the fact that the D-Wave QPU computes the solution to any problem accurately, and with high probability, raises the question as to whether the idealized principles of adiabatic quantum optimization are warranted when designing applications and algorithms for this processor. This is an important point in understanding how the D-Wave QPU programming and execution models integrate future applications into existing computing systems [25][26][27].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Recall Performance for Content-Addressable Memory Using Adiabatic Quantum Optimization

Schrock

McCaskey

Hamilton

et al. 2017

Entropy

Self Cite

View full text Add to dashboard Cite

A content-addressable memory (CAM) stores key-value associations such that the key is recalled by providing its associated value. While CAM recall is traditionally performed using recurrent neural network models, we show how to solve this problem using adiabatic quantum optimization. Our approach maps the recurrent neural network to a commercially available quantum processing unit by taking advantage of the common underlying Ising spin model. We then assess the accuracy of the quantum processor to store key-value associations by quantifying recall performance against an ensemble of problem sets. We observe that different learning rules from the neural network community influence recall accuracy but performance appears to be limited by potential noise in the processor. The strong connection established between quantum processors and neural network problems supports the growing intersection of these two ideas.

show abstract

“…Data is archived locally and post-processed to generate the results presented in the next section. The results presented were collected in August and September of 2016 [27].…”

Section: Methodsmentioning

confidence: 99%

Section: Experimental Validation Of Cam Recallmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Recall Performance for Content-Addressable Memory Using Adiabatic Quantum Optimization

Schrock

McCaskey

Hamilton

et al. 2017

Entropy

Self Cite

View full text Add to dashboard Cite

show abstract

“…Under this graph-theoretic formulation, the compilation process reduces to the NP-hard problem of minor embedding the problem graph into the hardware graph. In practice, this step represents a limitation bottleneck for the end-to-end program performance because existing embedding algorithms take orders of magnitude longer to execute than the quantum annealer itself [22]. Furthermore, no efficient universal embedding algorithm exists, with past algorithms addressing specific classes of problem instance (e.g.…”

Section: Introductionmentioning

confidence: 99%

Optimizing adiabatic quantum program compilation using a graph-theoretic framework

Goodrich

Sullivan

Humble

2018

Quantum Inf Process

Self Cite

View full text Add to dashboard Cite

Adiabatic quantum computing has evolved in recent years from a theoretical field into an immensely practical area, a change partially sparked by D-Wave System's quantum annealing hardware. These multimillion-dollar quantum annealers offer the potential to solve optimization problems millions of times faster than classical heuristics, prompting researchers at Google, NASA and Lockheed Martin to study how these computers can be applied to complex real-world problems such as NASA rover missions. Unfortunately, compiling (embedding) an optimization problem into the annealing hardware is itself a difficult optimization problem and a major bottleneck currently preventing widespread adoption. Additionally, while finding a single embedding is difficult, no generalized method is known for tuning embeddings to use minimal hardware resources. To address these barriers, we introduce a graph-theoretic framework for developing structured embedding algorithms. Using this framework, we introduce a biclique virtual hardware layer to provide a simplified interface to the physical hardware. Additionally, we exploit bipartite structure in quantum programs using odd cycle transversal (OCT) decompositions. By coupling an OCT-based embedding algorithm with new, generalized reduction methods, we develop a new baseline for embedding a wide range of optimization problems into fault-free D-Wave annealing hardware. To encourage the reuse and extension of these techniques, we provide an implementation of the framework and embedding algorithms.

show abstract

“…However, the method is not guaranteed to succeed and has a worst case complexity that scales as O(n 9 ) with the input graph order n (though average case behavior appears to be O(n 3 )). The CMR embedding algorithm represents a significant portion of the time needed for a quantum annealing workflow, and for even modest problem sizes it can far exceed the time required for executing a quantum annealing schedule [15]. Problem instances represented by large but incompletely connected input graphs must use embeddings that are both resource efficient and time efficient in order to ensure fast and correct solutions.…”

Section: Introductionmentioning

confidence: 99%

Identifying the minor set cover of dense connected bipartite graphs via random matching edge sets

Hamilton

Humble

2017

Quantum Inf Process

Self Cite

View full text Add to dashboard Cite

Using quantum annealing to solve an optimization problem requires minor embeddings of a logic graph into a known hardware graph. In an effort to reduce the complexity of the minor embedding problem, we introduce the minor set cover (MSC) of a known graph G: a subset of graph minors which contain any remaining minor of the graph as a subgraph. Any graph that can be embedded into G will be embeddable into a member of the MSC. Focusing on embedding into the hardware graph of commercially available quantum annealers, we establish the MSC for a particular known virtual hardware, which is a complete bipartite graph. We show that the complete bipartite graph KN,N has a MSC of N minors, from which KN+1 is identified as the largest clique minor of KN,N . The case of determining the largest clique minor of hardware with faults is briefly discussed but remains an open question.

show abstract

Performance Models for Split-Execution Computing Systems

Cited by 5 publications

References 31 publications

Recall Performance for Content-Addressable Memory Using Adiabatic Quantum Optimization

Recall Performance for Content-Addressable Memory Using Adiabatic Quantum Optimization

Optimizing adiabatic quantum program compilation using a graph-theoretic framework

Identifying the minor set cover of dense connected bipartite graphs via random matching edge sets

Contact Info

Product

Resources

About