General data structure expansion for multi-threading

Yu, Hongtao; Ko, Hou-Jen; Li, Zhiyuan

doi:10.1145/2491956.2462182

Cited by 6 publications

(6 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An effective technique was developed by Yu et al for data privatization on CPUs [Yu et al 2013]. This section shows how LD adapts their technique for use on GPUs, which have a different task model and memory architecture.…”

Section: Gpu Data Privatizationmentioning

confidence: 99%

“…The two transformations, variable expansion and access redirection, are implemented in Clang based on scout [Krzikalla 2011] using algorithms in Algorithm 1 and Figure 6. They are adapted from the technique by [Yu et al 2013]. To adapt their technique for use in GPU programs, we need the type promotion rules in Figure 6.…”

Section: Access Redirectionmentioning

confidence: 99%

“…The main novelty of LD is a use of privatization, not the technique itself. For more implementation details, please refer to Yu et al [2013].…”

Section: Access Redirectionmentioning

confidence: 99%

“…Prior privatization techniques [Gu et al 1997;Li 1992] were developed for automatic parallelization of scientific code on CPUs. Recently, Yu et al [2013] gave a general technique for data structure expansion for loop parallelization on CPUs. Our solution is for GPUs and has several differences.…”

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

Ld

Chen

et al. 2017

ACM Trans. Archit. Code Optim.

View full text Add to dashboard Cite

Data race detection has become an important problem in GPU programming. Previous designs of CPU racechecking tools are mainly task parallel and incur high overhead on GPUs due to access instrumentation, especially when monitoring many thousands of threads routinely used by GPU programs. This article presents a novel data-parallel solution designed and optimized for the GPU architecture. It includes compiler support and a set of runtime techniques. It uses value-based checking, which detects the races reported in previous work, finds new races, and supports race-free deterministic GPU execution. More important, race checking is massively data parallel and does not introduce divergent branching or atomic synchronization. Its slowdown is less than 5× for over half of the tests and 10× on average, which is orders of magnitude more efficient than the cuda-memcheck tool by Nvidia and the methods that use fine-grained access instrumentation.

show abstract

Section: Gpu Data Privatizationmentioning

confidence: 99%

Section: Access Redirectionmentioning

confidence: 99%

“…The main novelty of LD is a use of privatization, not the technique itself. For more implementation details, please refer to Yu et al [2013].…”

Section: Access Redirectionmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Ld

Chen

et al. 2017

ACM Trans. Archit. Code Optim.

View full text Add to dashboard Cite

show abstract

“…Other techniques looked at automatic software parallelization using expansion [29,41], data duplication and reductions [15,16,28,18], and privatization [38].…”

Section: Duplication Expansion and Privatizationmentioning

confidence: 99%

Flexible Support for Fast Parallel Commutative Updates

Balaji¹,

Tirumala²,

Lucia³

2017

Preprint

View full text Add to dashboard Cite

Privatizing data is a useful strategy for increasing parallelism in a shared memory multithreaded program. Independent cores can compute independently on duplicates of shared data, combining their results at the end of their computations. Conventional approaches to privatization, however, rely on explicit static or dynamic memory allocation for duplicated state, increasing memory footprint and contention for cache resources, especially in shared caches. In this work, we describe CCache, a system for on-demand privatization of data manipulated by commutative operations. CCache garners the benefits of privatization, without the increase in memory footprint or cache occupancy. Each core in CCache dynamically privatizes commutatively manipulated data, operating on a copy. Periodically or at the end of its computation, the core merges its value with the value resident in memory, and when all cores have merged, the in-memory copy contains the up-to-date value. We describe a low-complexity architectural implementation of CCache that extends a conventional multicore to support on-demand privatization without using additional memory for private copies. We evaluate CCache on several high-value applications, including random access key-value store, clustering, breadth first search and graph ranking, showing speedups upto 3.2X.

show abstract

$${\mathrm{DS}}_{\mathrm{spirit}}$$ DS spirit : a data dependence and stride reference patterns profiling infrastructure

Shu

2016

J Supercomput

View full text Add to dashboard Cite

General data structure expansion for multi-threading

Cited by 6 publications

References 34 publications

Ld

Ld

Flexible Support for Fast Parallel Commutative Updates

$${\mathrm{DS}}_{\mathrm{spirit}}$$ DS spirit : a data dependence and stride reference patterns profiling infrastructure

Contact Info

Product

Resources

About