Graph representation learning resurges as a trending research subject owing to the widespread use of deep learning for Euclidean data, which inspire various creative designs of neural networks in the non-Euclidean domain, particularly graphs. With the success of these graph neural networks (GNN) in the static setting, we approach further practical scenarios where the graph dynamically evolves. Existing approaches typically resort to node embeddings and use a recurrent neural network (RNN, broadly speaking) to regulate the embeddings and learn the temporal dynamics. These methods require the knowledge of a node in the full time span (including both training and testing) and are less applicable to the frequent change of the node set. In some extreme scenarios, the node sets at different time steps may completely differ. To resolve this challenge, we propose EvolveGCN, which adapts the graph convolutional network (GCN) model along the temporal dimension without resorting to node embeddings. The proposed approach captures the dynamism of the graph sequence through using an RNN to evolve the GCN parameters. Two architectures are considered for the parameter evolution. We evaluate the proposed approach on tasks including link prediction, edge classification, and node classification. The experimental results indicate a generally higher performance of EvolveGCN compared with related approaches. The code is available at https://github.com/IBM/EvolveGCN.

The miniaturization of semiconductor transistors has driven the growth in computer performance for more than 50 years. As miniaturization approaches its limits, bringing an end to Moore’s law, performance gains will need to come from software, algorithms, and hardware. We refer to these technologies as the “Top” of the computing stack to distinguish them from the traditional technologies at the “Bottom”: semiconductor physics and silicon-fabrication technology. In the post-Moore era, the Top will provide substantial performance gains, but these gains will be opportunistic, uneven, and sporadic, and they will suffer from the law of diminishing returns. Big system components offer a promising context for tackling the challenges of working at the Top.

We have developed a multithreaded implementation of breadth-first search (BFS) of a sparse graph using the Cilk++ extensions to C++. Our PBFS program on a single processor runs as quickly as a standard C++ breadth-first search implementation. PBFS achieves high work-efficiency by using a novel implementation of a multiset data structure, called a "bag," in place of the FIFO queue usually employed in serial breadth-first search algorithms. For a variety of benchmark input graphs whose diameters are significantly smaller than the number of vertices -a condition met by many real-world graphs -PBFS demonstrates good speedup with the number of processing cores.Since PBFS employs a nonconstant-time "reducer" -a "hyperobject" feature of Cilk++ -the work inherent in a PBFS execution depends nondeterministically on how the underlying work-stealing scheduler load-balances the computation. We provide a general method for analyzing nondeterministic programs that use reducers. PBFS also is nondeterministic in that it contains benign races which affect its performance but not its correctness. Fixing these races with mutual-exclusion locks slows down PBFS empirically, but it makes the algorithm amenable to analysis. In particular, we show that for a graph G = (V, E) with diameter D and bounded outdegree, this data-race-free version of PBFS algorithm runs in time O((V + E)/P + D lg 3 (V /D)) on P processors, which means that it attains near-perfect linear speedup if P ≪ (V + E)/D lg 3 (V /D).

Pipeline parallelism organizes a parallel program as a linear sequence of s stages. Each stage processes elements of a data stream, passing each processed data element to the next stage, and then taking on a new element before the subsequent stages have necessarily completed their processing. Pipeline parallelism is used especially in streaming applications that perform video, audio, and digital signal processing. Three out of 13 benchmarks in PARSEC, a popular software benchmark suite designed for shared-memory multiprocessors, can be expressed as pipeline parallelism.Whereas most concurrency platforms that support pipeline parallelism use a "construct-and-run" approach, this paper investigates "on-the-fly" pipeline parallelism, where the structure of the pipeline emerges as the program executes rather than being specified a priori. On-the-fly pipeline parallelism allows the number of stages to vary from iteration to iteration and dependencies to be data dependent. We propose simple linguistics for specifying on-the-fly pipeline parallelism and describe a provably efficient scheduling algorithm, the PIPER algorithm, which integrates pipeline parallelism into a work-stealing scheduler, allowing pipeline and fork-join parallelism to be arbitrarily nested. The PIPER algorithm automatically throttles the parallelism, precluding "runaway" pipelines. Given a pipeline computation with T 1 work and T ∞ span (critical-path length), PIPER executes the computation on P processors in T P ≤ T 1 /P + O(T ∞ + lg P) expected time. PIPER also limits stack space, ensuring that it does not grow unboundedly with running time.We have incorporated on-the-fly pipeline parallelism into a Cilkbased work-stealing runtime system. Our prototype Cilk-P implementation exploits optimizations such as lazy enabling and dependency folding. We have ported the three PARSEC benchmarks that exhibit pipeline parallelism to run on Cilk-P. One of these, x264, cannot readily be executed by systems that support only construct-and-run pipeline parallelism. Benchmark results indicate that Cilk-P has low serial overhead and good scalability. On x264, for example, Cilk-P exhibits a speedup of 13.87 over its respective serial counterpart when running on 16 processors.

This paper introduces the largest-log-degree-first (LLF) and smallest-log-degree-last (SLL) ordering heuristics for parallel greedy graph-coloring algorithms, which are inspired by the largest-degree-first (LF) and smallest-degree-last (SL) serial heuristics, respectively. We show that although LF and SL, in practice, generate colorings with relatively small numbers of colors, they are vulnerable to adversarial inputs for which any parallelization yields a poor parallel speedup. In contrast, LLF and SLL allow for provably good speedups on arbitrary inputs while, in practice, producing colorings of competitive quality to their serial analogs.We applied LLF and SLL to the parallel greedy coloring algorithm introduced by Jones and Plassmann, referred to here as JP. Jones and Plassman analyze the variant of JP that processes the vertices of a graph in a random order, and show that on an O(1)-degree graph G = (V, E), this JP-R variant has an expected parallel running time of O(lgV / lg lgV ) in a PRAM model. We improve this bound to show, using work-span analysis, that JP-R, augmented to handle arbitrary-degree graphs, colors a graph G = (V, E) with degree ∆ using Θ(V + E) work and O(lgV + lg ∆ · min{ √ E, ∆ + lg ∆ lgV / lg lgV }) expected span. We prove that JP-LLF and JP-SLL-JP using the LLF and SLL heuristics, respectivelyexecute with the same asymptotic work as JP-R and only logarithmically more span while producing higher-quality colorings than JP-R in practice.We engineered an efficient implementation of JP for modern shared-memory multicore computers and evaluated its performance on a machine with 12 Intel Core-i7 (Nehalem) processor cores. Our implementation of JP-LLF achieves a geometric-mean speedup of 7.83 on eight real-world graphs and a geometric-mean speedup of 8.08 on ten synthetic graphs, while our implementation using SLL achieves a geometric-mean speedup of 5.36 on these real-world graphs and a geometric-mean speedup of 7.02 on these synthetic graphs. Furthermore, on one processor, JP-LLF is slightly faster than a well-engineered serial greedy algorithm using LF, and likewise, JP-SLL is slightly faster than the greedy algorithm using SL.

No abstract

Existing concurrency platforms for dynamic multithreading do not provide repeatable parallel random-number generators. This paper proposes that a mechanism called pedigrees be built into the runtime system to enable efficient deterministic parallel randomnumber generation. Experiments with the open-source MIT Cilk runtime system show that the overhead for maintaining pedigrees is negligible. Specifically, on a suite of 10 benchmarks, the relative overhead of Cilk with pedigrees to the original Cilk has a geometric mean of less than 1%.We persuaded Intel to modify its commercial C/C++ compiler, which provides the Cilk Plus concurrency platform, to include pedigrees, and we built a library implementation of a deterministic parallel random-number generator called DOTMIX that compresses the pedigree and then "RC6-mixes" the result. The statistical quality of DOTMIX is comparable to that of the popular Mersenne twister, but somewhat slower than a nondeterministic parallel version of this efficient and high-quality serial random-number generator. The cost of calling DOTMIX depends on the "spawn depth" of the invocation. For a naive Fibonacci calculation with n = 40 that calls DOTMIX in every node of the computation, this "price of determinism" is about a factor of 2.3 in running time over the nondeterministic Mersenne twister, but for more realistic applications with less intense use of random numbers -such as a maximalindependent-set algorithm, a practical samplesort program, and a Monte Carlo discrete-hedging application from QuantLib -the observed "price" was at most 21%, and sometimes much less. Moreover, even if overheads were several times greater, applications using DOTMIX should be amply fast for debugging purposes, which is a major reason for desiring repeatability.

A data-graph computation -popularized by such programming systems as Galois, Pregel, GraphLab, PowerGraph, and GraphChi -is an algorithm that performs local updates on the vertices of a graph. During each round of a data-graph computation, an update function atomically modifies the data associated with a vertex as a function of the vertex's prior data and that of adjacent vertices. A dynamic data-graph computation updates only an active subset of the vertices during a round, and those updates determine the set of active vertices for the next round.This paper introduces PRISM, a chromatic-scheduling algorithm for executing dynamic data-graph computations. PRISM uses a vertex-coloring of the graph to coordinate updates performed in a round, precluding the need for mutual-exclusion locks or other nondeterministic data synchronization. A multibag data structure is used by PRISM to maintain a dynamic set of active vertices as an unordered set partitioned by color. We analyze PRISM using work-span analysis. Let G = (V, E) be a degree-∆ graph colored with χ colors, and suppose that Q ⊆ V is the set of active vertices in a round. Define size(Q) = |Q| + v∈Q deg(v), which is proportional to the space required to store the vertices of Q using a sparsegraph layout. We show that a P-processor execution of PRISM performs updates in Q using O(χ(lg(Q/χ) + lg ∆) + lg P) span and Θ(size(Q) + χ + P) work. These theoretical guarantees are matched by good empirical performance. We modified GraphLab to incorporate PRISM and studied seven application benchmarks on a 12-core multicore machine. PRISM executes the benchmarks 1.2-2.1 times faster than GraphLab's nondeterministic lock-based scheduler while providing deterministic behavior. This paper also presents PRISM-R, a variation of PRISM that executes dynamic data-graph computations deterministically even when updates modify global variables with associative operations. PRISM-R satisfies the same theoretical bounds as PRISM, but its implementation is more involved, incorporating a multivector data structure to maintain an ordered set of vertices partitioned by color.