Mapping adaptive resonance theory onto ring and mesh architectures

In recent years, parallel computers have been attracting attention for simulating artificial neural networks (ANN). This is due to the inherent parallelism in ANN. This work is aimed at studying ways of parallelizing adaptive resonance theory (ART), a popular neural network algorithm. The core computations of ART are separated and different strategies of parallelizing ART are discussed. We present mapping strategies for ART 2-A neural network onto ring and mesh architectures. The required parallel architecture is simulated using a parallel architectural simulator, PROTEUS and parallel programs are written using a superset of C for the algorithms presented. A simulation-based scalability study of the algorithm-architecture match is carried out. The various overheads are identified in order to suggest ways of improving the performance. Our main objective is to find out the performance of the ART2-A network on different parallel architectures. (C) 1999 Elsevier Science B.V. All rights reserved.

Parallel Computation of 2D Morse-Smale Complexes

The Morse-Smale complex is a useful topological data structure for the analysis and visualization of scalar data. This paper describes an algorithm that processes all mesh elements of the domain in parallel to compute the Morse-Smale complex of large two-dimensional data sets at interactive speeds. We employ a reformulation of the Morse-Smale complex using Forman's Discrete Morse Theory and achieve scalability by computing the discrete gradient using local accesses only. We also introduce a novel approach to merge gradient paths that ensures accurate geometry of the computed complex. We demonstrate that our algorithm performs well on both multicore environments and on massively parallel architectures such as the GPU.

Generating Efficient Data Movement Code for Heterogeneous Architectures with Distributed-Memory

Programming for parallel architectures that do not have a shared address space is extremely difficult due to the need for explicit communication between memories of different compute devices. A heterogeneous system with CPUs and multiple GPUs, or a distributed-memory cluster are examples of such systems. Past works that try to automate data movement for distributed-memory architectures can lead to excessive redundant communication. In this paper, we propose an automatic data movement scheme that minimizes the volume of communication between compute devices in heterogeneous and distributed-memory systems. We show that by partitioning data dependences in a particular non-trivial way, one can generate data movement code that results in the minimum volume for a vast majority of cases. The techniques are applicable to any sequence of affine loop nests and works on top of any choice of loop transformations, parallelization, and computation placement. The data movement code generated minimizes the volume of communication for a particular configuration of these. We use a combination of powerful static analyses relying on the polyhedral compiler framework and lightweight runtime routines they generate, to build a source-to-source transformation tool that automatically generates communication code. We demonstrate that the tool is scalable and leads to substantial gains in efficiency. On a heterogeneous system, the communication volume is reduced by a factor of 11X to 83X over state-of-the-art, translating into a mean execution time speedup of 1.53X. On a distributed-memory cluster, our scheme reduces the communication volume by a factor of 1.4X to 63.5X over state-of-the-art, resulting in a mean speedup of 1.55X. In addition, our scheme yields a mean speedup of 2.19X over hand-optimized UPC codes.

Parallel Flow-Sensitive Pointer Analysis by Graph-Rewriting

Precise pointer analysis is a problem of interest to both the compiler and the program verification community. Flow-sensitivity is an important dimension of pointer analysis that affects the precision of the final result computed. Scaling flow-sensitive pointer analysis to millions of lines of code is a major challenge. Recently, staged flow-sensitive pointer analysis has been proposed, which exploits a sparse representation of program code created by staged analysis. In this paper we formulate the staged flow-sensitive pointer analysis as a graph-rewriting problem. Graph-rewriting has already been used for flow-insensitive analysis. However, formulating flow-sensitive pointer analysis as a graph-rewriting problem adds additional challenges due to the nature of flow-sensitivity. We implement our parallel algorithm using Intel Threading Building Blocks and demonstrate considerable scaling (upto 2.6x) for 8 threads on a set of 10 benchmarks. Compared to the sequential implementation of staged flow-sensitive analysis, a single threaded execution of our implementation performs better in 8 of the benchmarks.

Analyzing Cache Performance Bottlenecks of STM Applications and addressing them with Compiler's help

Software transactional memory (STM) is a promising programming paradigm for shared memory multithreaded programs as an alternative to traditional lock based synchronization. However adoption of STM in mainstream software has been quite low due to its considerable overheads and its poor cache/memory performance. In this paper, we perform a detailed study of the cache behavior of STM applications and quantify the impact of different STM factors on the cache misses experienced by the applications. Based on our analysis, we propose a compiler driven Lock-Data Colocation (LDC), targeted at reducing the cache overheads on STM. We show that LDC is effective in improving the cache behavior of STM applications by reducing the dcache miss latency and improving execution time performance.

NUcache: A Multicore Cache Organization Based on Next-Use Distance

In this work, we propose a new organization for the last level shared cache of a rnulticore system. Our design is based on the observation that the Next-Use distance, measured in terms of intervening misses between the eviction of a line and its next use, for lines brought in by a given delinquent PC falls within a predictable range of values. We exploit this correlation to improve the performance of shared caches in multi-core architectures by proposing the NUcache organization.

Region Based Structure Layout Optimization by Selective Data Copying

As the gap between processor and memory continues to grow Memory performance becomes a key performance bottleneck for many applications. Compilers therefore increasingly seek to modify an application’s data layout to improve cache locality and cache reuse. Whole program Structure Layout [WPSL] transformations can significantly increase the spatial locality of data and reduce the runtime of programs that use link-based data structures, by increasing the cache line utilization. However, in production compilers WPSL transformations do not realize the entire performance potential possible due to a number of factors. Structure layout decisions made on the basis of whole program aggregated affinity/hotness of structure fields, can be sub optimal for local code regions. WPSL is also restricted in applicability in production compilers for type unsafe languages like C/C++ due to the extensive legality checks and field sensitive pointer analysis required over the entire application. In order to overcome the issues associated with WPSL, we propose Region Based Structure Layout (RBSL) optimization framework, using selective data copying. We describe our RBSL framework, implemented in the production compiler for C/C++ on HP-UX IA-64. We show that acting in complement to the existing and mature WPSL transformation framework in our compiler, RBSL improves application performance in pointer intensive SPEC benchmarks ranging from 3% to 28% over WPSL

Two-level Mapping Based Cache Index Selection for Packet Forwarding Engines

Packet forwarding is a memory-intensive application requiring multiple accesses through a trie structure. The efficiency of a cache for this application critically depends on the placement function to reduce conflict misses. Traditional placement functions use a one-level mapping that naively partitions trie-nodes into cache sets. However, as a significant percentage of trie nodes are not useful, these schemes suffer from a non-uniform distribution of useful nodes to sets. This in turn results in increased conflict misses. Newer organizations such as variable associativity caches achieve flexibility in placement at the expense of increased hit-latency. This makes them unsuitable for L1 caches.We propose a novel two-level mapping framework that retains the hit-latency of one-level mapping yet incurs fewer conflict misses. This is achieved by introducing a secondlevel mapping which reorganizes the nodes in the naive initial partitions into refined partitions with near-uniform distribution of nodes. Further as this remapping is accomplished by simply adapting the index bits to a given routing table the hit-latency is not affected. We propose three new schemes which result in up to 16% reduction in the number of misses and 13% speedup in memory access time. In comparison, an XOR-based placement scheme known to perform extremely well for general purpose architectures, can obtain up to 2% speedup in memory access time.

Making STMs Cache Friendly with Compiler Transformations

Software transactional memory (STM) is a promising programming paradigm for shared memory multithreaded programs. In order for STMs to be adopted widely for performance critical software, understanding and improving the cache performance of applications running on STM becomes increasingly crucial, as the performance gap between processor and memory continues to grow. In this paper, we present the most detailed experimental evaluation to date, of the cache behavior of STM applications and quantify the impact of the different STM factors on the cache misses experienced by the applications. We find that STMs are not cache friendly, with the data cache stall cycles contributing to more than 50% of the execution cycles in a majority of the benchmarks. We find that on an average, misses occurring inside the STM account for 62% of total data cache miss latency cycles experienced by the applications and the cache performance is impacted adversely due to certain inherent characteristics of the STM itself. The above observations motivate us to propose a set of specific compiler transformations targeted at making the STMs cache friendly. We find that STM's fine grained and application unaware locking is a major contributor to its poor cache behavior. Hence we propose selective Lock Data co-location (LDC) and Redundant Lock Access Removal (RLAR) to address the lock access misses. We find that even transactions that are completely disjoint access parallel, suffer from costly coherence misses caused by the centralized global time stamp updates and hence we propose the Selective Per-Partition Time Stamp (SPTS) transformation to address this. We show that our transformations are effective in improving the cache behavior of STM applications by reducing the data cache miss latency by 20.15% to 37.14% and improving execution time by 18.32% to 33.12% in five of the 8 STAMP applications.

Row-buffer reorganization: simultaneously improving performance and reducing energy in DRAMs

In this paper, based on the temporal and spatial locality characteristics of memory accesses in multicores, we propose a re-organization of the existing single large row buffer in a DRAM bank into multiple smaller row-buffers. The proposed configuration helps improve the row hit rates and also brings down the energy required for row-activations. The major contribution of this work is proposing such a reorganization without requiring any significant changes to the existing widely accepted DRAM specifications. Our proposed reorganization improves performance by 35.8%, 14.5% and 21.6% in quad, eight and sixteen core workloads along with a 42%, 28% and 31% reduction in DRAM energy. Additionally, we introduce a Need Based Allocation scheme for buffer management that shows additional performance improvement.

Fast and efficient automatic memory management for GPUs using compiler-assisted runtime coherence scheme

Exploiting the performance potential of GPUs requires managing the data transfers to and from them efficiently which is an error-prone and tedious task. In this paper, we develop a software coherence mechanism to fully automate all data transfers between the CPU and GPU without any assistance from the programmer. Our mechanism uses compiler analysis to identify potential stale accesses and uses a runtime to initiate transfers as necessary. This allows us to avoid redundant transfers that are exhibited by all other existing automatic memory management proposals. We integrate our automatic memory manager into the X10 compiler and runtime, and find that it not only results in smaller and simpler programs, but also eliminates redundant memory transfers. Tested on eight programs ported from the Rodinia benchmark suite it achieves (i) a 1.06x speedup over hand-tuned manual memory management, and (ii) a 1.29x speedup over another recently proposed compiler--runtime automatic memory management system. Compared to other existing runtime-only and compiler-only proposals, it also transfers 2.2x to 13.3x less data on average.

Implementation of Art1 and Art2 Artificial Neural Networks on Ring and Mesh Architectures

The Artificial Neural Networks (ANNs) are being used to solve a variety of problems in pattern recognition, robotic control, VLSI CAD and other areas. In most of these applications, a speedy response from the ANNs is imperative. However, ANNs comprise a large number of artificial neurons, and a massive interconnection network among them. Hence, implementation of these ANNs involves execution of computer-intensive operations. The usage of multiprocessor systems therefore becomes necessary. In this article, we have presented the implementation of ART1 and ART2 ANNs on ring and mesh architectures. The overall system design and implementation aspects are presented. The performance of the algorithm on ring, 2-dimensional mesh and n-dimensional mesh topologies is presented. The parallel algorithm presented for implementation of ART1 is not specific to any particular architecture. The parallel algorithm for ARTE is more suitable for a ring architecture.

An optimal parallel algorithm for sparse matrix bandwidth reduction on a linear array

A simple and efficient algorithm for the bandwidth reduction of sparse symmetric matrices is proposed. It involves column-row permutations and is well-suited to map onto the linear array topology of the SIMD architectures. The efficiency of the algorithm is compared with the other existing algorithms. The interconnectivity and the memory requirement of the linear array are discussed and the complexity of its layout area is derived. The parallel version of the algorithm mapped onto the linear array is then introduced and is explained with the help of an example. The optimality of the parallel algorithm is proved by deriving the time complexities of the algorithm on a single processor and the linear array.

Compiler-assisted power optimization for clustered VLIW architectures

Clustered VLIW architectures solve the scalability problem associated with flat VLIW architectures by partitioning the register file and connecting only a subset of the functional units to a register file. However, inter-cluster communication in clustered architectures leads to increased leakage in functional components and a high number of register accesses. In this paper, we propose compiler scheduling algorithms targeting two previously ignored power-hungry components in clustered VLIW architectures, viz., instruction decoder and register file. We consider a split decoder design and propose a new energy-aware instruction scheduling algorithm that provides 14.5% and 17.3% benefit in the decoder power consumption on an average over a purely hardware based scheme in the context of 2-clustered and 4-clustered VLIW machines. In the case of register files, we propose two new scheduling algorithms that exploit limited register snooping capability to reduce extra register file accesses. The proposed algorithms reduce register file power consumption on an average by 6.85% and 11.90% (10.39% and 17.78%), respectively, along with performance improvement of 4.81% and 5.34% (9.39% and 11.16%) over a traditional greedy algorithm for 2-clustered (4-clustered) VLIW machine. (C) 2010 Elsevier B.V. All rights reserved.