Automatic Insertion of Copy Annotation in Data-Parallel Programs

Mendonca, Gleison Souza Diniz; Guimarães, Breno C. F.; Alves, Péricles; Pereira, Fernando Magno Quintão; Pereira, Marcio Machado; Araújo, Guido

doi:10.1109/sbac-pad.2016.13

Cited by 12 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compile-time setting of accesses: Many research efforts target automatic compile-time parallelization, e.g., Cetus [24], DawnCC [25], [26], AutoPar [27], Pluto [28] and TaskMiner [29]. These primarily target data parallelism, and only TaskMiner is specific to tasks.…”

Section: Related Workmentioning

confidence: 99%

Automatic aggregation of subtask accesses for nested OpenMP-style tasks

Shaaban

Aguilar

Beltrán

et al. 2022

2022 IEEE 34th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD)

View full text Add to dashboard Cite

Task-based programming is a high performance and productive model to express parallelism. Tasks encapsulate work to be executed across multiple cores or offloaded to GPUs, FPGAs, other accelerators or other nodes. In order to maintain parallelism and afford maximum freedom to the scheduler, the task dependency graph should be created in parallel and well in advance of task execution. A key limitation with OpenMP and OmpSs-2 tasking is that a task cannot be created until all its accesses and its descendents' accesses are known. Current approaches to work around this limitation either stop task creation and execution using a taskwait or they substitute "fake" accesses known as sentinels. This paper proposes the auto clause, which indicates that the task may create subtasks that access unspecified memory regions or it may allocate and return memory at addresses that are of course not yet known. Unlike approaches using taskwaits, there is no interruption to the concurrent creation and execution of tasks, maintaining parallelism and the scheduler's ability to optimize load balance and data locality. Unlike existing approaches using sentinels, all tasks can be given a precise specification of their own data accesses, so that a single mechanism is used to control task ordering, program data transfers on distributed memory and optimize data locality, e.g. on NUMA systems. The auto clause also provides an incremental path to develop programs with nested tasks, by removing the need for every parent task to have a complete specification of the accesses of its descendent tasks. This is redundant information that can be time consuming and error-prone to describe. We present a straightforward runtime implementation that achieves a 1.4 times speedup for n-body with OmpSs-2@Cluster task offloading to 32 nodes and <4% slowdown for three benchmarks with task offloading to 8 nodes. All code is open source.

show abstract

Section: Related Workmentioning

confidence: 99%

Automatic aggregation of subtask accesses for nested OpenMP-style tasks

Shaaban

Aguilar

Beltrán

et al. 2022

2022 IEEE 34th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD)

View full text Add to dashboard Cite

show abstract

“…All these are written in standard C, without any adaptations for a Graphics Processing Unit (GPU). To compile these programs to the Tegra's GPU, we have used DawnCC [Mendonca et al (2016)] 4 to annotate them with OpenACC directives. Each benchmark has only one core loop, which implements the bulk of the processing.…”

Section: Static Scheduling Of Code On Heterogeneous Devices -Manual Amentioning

confidence: 99%

“…In the previous discussion, we have manually annotated a series of simple benchmarks with OpenACC directives, and have compared them when running on the CPU, or on the GPU. This time we repeat this experiment, but now using an automatic parallelizer, dawn-cc [Mendonca et al (2016)]. This tool is available through an on-line server 5 which reads C sources, and produces C sources annotated with either OpenACC or OpenMP directives.…”

Section: Static Scheduling Of Code On Heterogeneous Devices -Manual Amentioning

confidence: 99%

JetsonLEAP: A framework to measure power on a heterogeneous system-on-a-chip device

Bessa

Gull

Quintão

et al. 2019

Science of Computer Programming

View full text Add to dashboard Cite

Computer science marches towards energy-aware practices. This trend impacts not only the design of computer architectures, but also the design of programs. However, developers still lack affordable and accurate technology to measure energy consumption in computing systems. The goal of this paper is to mitigate such problem. To this end, we introduce JetsonLEAP, a framework that supports the implementation of energy-aware programs. Jet-sonLEAP consists of an embedded hardware, in our case, the Nvidia Tegra TK1 System-on-a-chip device, a circuit to control the flow of energy, of our own design, plus a library to instrument program parts. We discuss two different circuit setups. The most precise setup lets us reliably measure the energy spent by 225,000 instructions, the least precise, although more affordable setup, gives us a window of 975,000 instructions. To probe the precision of our system, we use it in tandem with a high-precision, high-cost acquisition system, and show that results do not differ in any significant way from those that we get using our simpler apparatus. Our entire infrastructure -board, power meter and both circuits -can be reproduced with about $500.00. To demonstrate the efficacy of our framework, we have used it to measure the energy consumed by programs running on ARM cores, on the GPU, and on a remote server. Furthermore, we have studied the impact of OpenACC directives on the energy efficiency of high-performance applications.

show abstract