Improving Flow-Insensitive Solutions for Non-Separable Dataflow Problems

Flow-insensitive solutions to dataflow problems have been known to be highly scalable; however also hugely imprecise. For non-separable dataflow problems this solution is further degraded due to spurious facts generated as a result of dependence among the dataflow facts. We propose an improvement to the standard flow-insensitive analysis by creating a generalized version of the dominator relation that reduces the number of spurious facts generated. In addition, the solution obtained contains extra information to facilitate the extraction of a better solution at any program point, very close to the flow-sensitive solution. To improve the solution further, we propose the use of an intra-block variable renaming scheme. We illustrate these concepts using two classic non-separable dataflow problems --- points-to analysis and constant propagation.

Minimizing buffer requirements under rate-optimal schedule in regular dataflow networks

Large-grain synchronous dataflow graphs or multi-rate graphs have the distinct feature that the nodes of the dataflow graph fire at different rates. Such multi-rate large-grain dataflow graphs have been widely regarded as a powerful programming model for DSP applications. In this paper we propose a method to minimize buffer storage requirement in constructing rate-optimal compile-time (MBRO) schedules for multi-rate dataflow graphs. We demonstrate that the constraints to minimize buffer storage while executing at the optimal computation rate (i.e. the maximum possible computation rate without storage constraints) can be formulated as a unified linear programming problem in our framework. A novel feature of our method is that in constructing the rate-optimal schedule, it directly minimizes the memory requirement by choosing the schedule time of nodes appropriately. Lastly, a new circular-arc interval graph coloring algorithm has been proposed to further reduce the memory requirement by allowing buffer sharing among the arcs of the multi-rate dataflow graph. We have constructed an experimental testbed which implements our MBRO scheduling algorithm as well as (i) the widely used periodic admissible parallel schedules (also known as block schedules) proposed by Lee and Messerschmitt (IEEE Transactions on Computers, vol. 36, no. 1, 1987, pp. 24-35), (ii) the optimal scheduling buffer allocation (OSBA) algorithm of Ning and Gao (Conference Record of the Twentieth Annual ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, Charleston, SC, Jan. 10-13, 1993, pp. 29-42), and (iii) the multi-rate software pipelining (MRSP) algorithm (Govindarajan and Gao, in Proceedings of the 1993 International Conference on Application Specific Array Processors, Venice, Italy, Oct. 25-27, 1993, pp. 77-88). Schedules generated for a number of random dataflow graphs and for a set of DSP application programs using the different scheduling methods are compared. The experimental results have demonstrated a significant improvement (10-20%) in buffer requirements for the MBRO schedules compared to the schedules generated by the other three methods, without sacrificing the computation rate. The MBRO method also gives a 20% average improvement in computation rate compared to Lee's Block scheduling method.

Dataflow analysis for data-race free programs

Memory models for shared-memory concurrent programming languages typically guarantee sequential consistency (SC) semantics for datarace-free (DRF) programs, while providing very weak or no guarantees for non-DRF programs. In effect programmers are expected to write only DRF programs, which are then executed with SC semantics. With this in mind, we propose a novel scalable solution for dataflow analysis of concurrent programs, which is proved to be sound for DRF programs with SC semantics. We use the synchronization structure of the program to propagate dataflow information among threads without requiring to consider all interleavings explicitly. Given a dataflow analysis that is sound for sequential programs and meets certain criteria, our technique automatically converts it to an analysis for concurrent programs.

PolyGLoT: A Polyhedral Loop Transformation Framework for a Graphical Dataflow Language

Polyhedral techniques for program transformation are now used in several proprietary and open source compilers. However, most of the research on polyhedral compilation has focused on imperative languages such as C, where the computation is specified in terms of statements with zero or more nested loops and other control structures around them. Graphical dataflow languages, where there is no notion of statements or a schedule specifying their relative execution order, have so far not been studied using a powerful transformation or optimization approach. The execution semantics and referential transparency of dataflow languages impose a different set of challenges. In this paper, we attempt to bridge this gap by presenting techniques that can be used to extract polyhedral representation from dataflow programs and to synthesize them from their equivalent polyhedral representation. We then describe PolyGLoT, a framework for automatic transformation of dataflow programs which we built using our techniques and other popular research tools such as Clan and Pluto. For the purpose of experimental evaluation, we used our tools to compile LabVIEW, one of the most widely used dataflow programming languages. Results show that dataflow programs transformed using our framework are able to outperform those compiled otherwise by up to a factor of seventeen, with a mean speed-up of 2.30x while running on an 8-core Intel system.

Compiler/Runtime Framework for Dynamic Dataflow Parallelization of Tiled Programs

Task-parallel languages are increasingly popular. Many of them provide expressive mechanisms for intertask synchronization. For example, OpenMP 4.0 will integrate data-driven execution semantics derived from the StarSs research language. Compared to the more restrictive data-parallel and fork-join concurrency models, the advanced features being introduced into task-parallelmodels in turn enable improved scalability through load balancing, memory latency hiding, mitigation of the pressure on memory bandwidth, and, as a side effect, reduced power consumption. In this article, we develop a systematic approach to compile loop nests into concurrent, dynamically constructed graphs of dependent tasks. We propose a simple and effective heuristic that selects the most profitable parallelization idiom for every dependence type and communication pattern. This heuristic enables the extraction of interband parallelism (cross-barrier parallelism) in a number of numerical computations that range from linear algebra to structured grids and image processing. The proposed static analysis and code generation alleviates the burden of a full-blown dependence resolver to track the readiness of tasks at runtime. We evaluate our approach and algorithms in the PPCG compiler, targeting OpenStream, a representative dataflow task-parallel language with explicit intertask dependences and a lightweight runtime. Experimental results demonstrate the effectiveness of the approach.

RETHROTTLE: Execution Throttling in the REDEFINE SoC Architecture

REDEFINE is a reconfigurable SoC architecture that provides a unique platform for high performance and low power computing by exploiting the synergistic interaction between coarse grain dynamic dataflow model of computation (to expose abundant parallelism in applications) and runtime composition of efficient compute structures (on the reconfigurable computation resources). We propose and study the throttling of execution in REDEFINE to maximize the architecture efficiency. A feature specific fast hybrid (mixed level) simulation framework for early in design phase study is developed and implemented to make the huge design space exploration practical. We do performance modeling in terms of selection of important performance criteria, ranking of the explored throttling schemes and investigate effectiveness of the design space exploration using statistical hypothesis testing. We find throttling schemes which give appreciable (24.8%) overall performance gain in the architecture and 37% resource usage gain in the throttling unit simultaneously.

Modelling and analysis of the variance in parallelism in parallel computations

In this paper, we introduce an analytical technique based on queueing networks and Petri nets for making a performance analysis of dataflow computations when executed on the Manchester machine. This technique is also applicable for the analysis of parallel computations on multiprocessors. We characterize the parallelism in dataflow computations through a four-parameter characterization, namely, the minimum parallelism, the maximum parallelism, the average parallelism and the variance in parallelism. We observe through detailed investigation of our analytical models that the average parallelism is a good characterization of the dataflow computations only as long as the variance in parallelism is small. However, significant difference in performance measures will result when the variance in parallelism is comparable to or higher than the average parallelism.

A Flat Concurrent Prolog compiler for PARAM

We describe a compiler for the Flat Concurrent Prolog language on a message passing multiprocessor architecture. This compiler permits symbolic and declarative programming in the syntax of Guarded Horn Rules, The implementation has been verified and tested on the 64-node PARAM parallel computer developed by C-DAC (Centre for the Development of Advanced Computing, India), Flat Concurrent Prolog (FCP) is a logic programming language designed for concurrent programming and parallel execution, It is a process oriented language, which embodies dataflow synchronization and guarded-command as its basic control mechanisms. An identical algorithm is executed on every processor in the network, We assume regular network topologies like mesh, ring, etc, Each node has a local memory, The algorithm comprises of two important parts: reduction and communication, The most difficult task is to integrate the solutions of problems that arise in the implementation in a coherent and efficient manner. We have tested the efficacy of the compiler on various benchmark problems of the ICOT project that have been reported in the recent book by Evan Tick, These problems include Quicksort, 8-queens, and Prime Number Generation, The results of the preliminary tests are favourable, We are currently examining issues like indexing and load balancing to further optimize our compiler.

Null Dereference Verification via Over-approximated Weakest Pre-conditions Analysis

Null dereferences are a bane of programming in languages such as Java. In this paper we propose a sound, demand-driven, inter-procedurally context-sensitive dataflow analysis technique to verify a given dereference as safe or potentially unsafe. Our analysis uses an abstract lattice of formulas to find a pre-condition at the entry of the program such that a null-dereference can occur only if the initial state of the program satisfies this pre-condition. We use a simplified domain of formulas, abstracting out integer arithmetic, as well as unbounded access paths due to recursive data structures. For the sake of precision we model aliasing relationships explicitly in our abstract lattice, enable strong updates, and use a limited notion of path sensitivity. For the sake of scalability we prune formulas continually as they get propagated, reducing to true conjuncts that are less likely to be useful in validating or invalidating the formula. We have implemented our approach, and present an evaluation of it on a set of ten real Java programs. Our results show that the set of design features we have incorporated enable the analysis to (a) explore long, inter-procedural paths to verify each dereference, with (b) reasonable accuracy, and (c) very quick response time per dereference, making it suitable for use in desktop development environments.

Two techniques to improve the precision of a demand-driven null-dereference verification approach

The problem addressed in this paper is sound, scalable, demand-driven null-dereference verification for Java programs. Our approach consists conceptually of a base analysis, plus two major extensions for enhanced precision. The base analysis is a dataflow analysis wherein we propagate formulas in the backward direction from a given dereference, and compute a necessary condition at the entry of the program for the dereference to be potentially unsafe. The extensions are motivated by the presence of certain ``difficult'' constructs in real programs, e.g., virtual calls with too many candidate targets, and library method calls, which happen to need excessive analysis time to be analyzed fully. The base analysis is hence configured to skip such a difficult construct when it is encountered by dropping all information that has been tracked so far that could potentially be affected by the construct. Our extensions are essentially more precise ways to account for the effect of these constructs on information that is being tracked, without requiring full analysis of these constructs. The first extension is a novel scheme to transmit formulas along certain kinds of def-use edges, while the second extension is based on using manually constructed backward-direction summary functions of library methods. We have implemented our approach, and applied it on a set of real-life benchmarks. The base analysis is on average able to declare about 84% of dereferences in each benchmark as safe, while the two extensions push this number up to 91%. (C) 2014 Elsevier B.V. All rights reserved.

PLAStiCC: Predictive Look-Ahead Scheduling for Continuous dataflows on Clouds

Scalable stream processing and continuous dataflow systems are gaining traction with the rise of big data due to the need for processing high velocity data in near real time. Unlike batch processing systems such as MapReduce and workflows, static scheduling strategies fall short for continuous dataflows due to the variations in the input data rates and the need for sustained throughput. The elastic resource provisioning of cloud infrastructure is valuable to meet the changing resource needs of such continuous applications. However, multi-tenant cloud resources introduce yet another dimension of performance variability that impacts the application's throughput. In this paper we propose PLAStiCC, an adaptive scheduling algorithm that balances resource cost and application throughput using a prediction-based lookahead approach. It not only addresses variations in the input data rates but also the underlying cloud infrastructure. In addition, we also propose several simpler static scheduling heuristics that operate in the absence of accurate performance prediction model. These static and adaptive heuristics are evaluated through extensive simulations using performance traces obtained from Amazon AWS IaaS public cloud. Our results show an improvement of up to 20% in the overall profit as compared to the reactive adaptation algorithm.