Interleaving and Lock-Step Semantics for Analysis and Verification of GPU Kernels

Collingbourne, Peter; Donaldson, Alastair F.; Ketema, Jeroen; Qadeer, Shaz

doi:10.1007/978-3-642-37036-6_16

Cited by 36 publications

(29 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We make explicit how the current article significantly extends a previous conference version [Betts et al 2012] and complements other papers that have arisen from the GPUVerify project [Collingbourne et al 2013;Chong et al 2013;Bardsley et al 2014a;.…”

Section: Contributions In Relation To Prior Workmentioning

confidence: 57%

“…In addition to the original version of the paper [Betts et al 2012], we have published three technical articles that extend the GPUVerify verification method, each in a distinct manner [Collingbourne et al 2013;Chong et al 2013;. Each of these extensions is separate from, and complementary to, the new contributions made in the current article.…”

Section: Contributions In Relation To Prior Workmentioning

confidence: 95%

“…In related work, we have generalized the lock-step execution semantics used by our method to apply to unstructured control flow graphs, allowing the technique to be applied at the level of the intermediate representation used by a compiler [Collingbourne et al 2013].…”

Section: Contributions In Relation To Prior Workmentioning

confidence: 99%

See 2 more Smart Citations

The Design and Implementation of a Verification Technique for GPU Kernels

Betts

Chong

Donaldson

et al. 2015

ACM Trans. Program. Lang. Syst.

Self Cite

View full text Add to dashboard Cite

We present a technique for the formal verification of GPU kernels, addressing two classes of correctness properties: data races and barrier divergence. Our approach is founded on a novel formal operational semantics for GPU kernels termed synchronous, delayed visibility (SDV) semantics, which captures the execution of a GPU kernel by multiple groups of threads. The SDV semantics provides operational definitions for barrier divergence and for both inter-and intra-group data races. We build on the semantics to develop a method for reducing the task of verifying a massively parallel GPU kernel to that of verifying a sequential program. This completely avoids the need to reason about thread interleavings, and allows existing techniques for sequential program verification to be leveraged. We describe an efficient encoding of data race detection and propose a method for automatically inferring the loop invariants that are required for verification. We have implemented these techniques as a practical verification tool, GPUVerify, that can be applied directly to OpenCL and CUDA source code. We evaluate GPUVerify with respect to a set of 162 kernels drawn from public and commercial sources. Our evaluation demonstrates that GPUVerify is capable of efficient, automatic verification of a large number of real-world kernels.

show abstract

Section: Contributions In Relation To Prior Workmentioning

confidence: 57%

Section: Contributions In Relation To Prior Workmentioning

confidence: 95%

See 1 more Smart Citation

The Design and Implementation of a Verification Technique for GPU Kernels

Betts

Chong

Donaldson

et al. 2015

ACM Trans. Program. Lang. Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Collingbourne et al [19] presented a sound analysis technique for GPU-oriented languages like OpenCL and CUDA, which have concurrency models that are rather different from Java. One of the applications of their technique is to race detection.…”

Section: Related Workmentioning

confidence: 99%

Race directed scheduling of concurrent programs

Eslamimehr

Palsberg

2014

Proceedings of the 19th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming

View full text Add to dashboard Cite

Detection of data races in Java programs remains a difficult problem. The best static techniques produce many false positives, and also the best dynamic techniques leave room for improvement. We present a new technique called race directed scheduling that for a given race candidate searches for an input and a schedule that lead to the race. The search iterates a combination of concolic execution and schedule improvement, and turns out to find useful inputs and schedules efficiently. We use an existing technique to produce a manageable number of race candidates. Our experiments on 23 Java programs found 72 real races that were missed by the best existing dynamic techniques. Among those 72 races, 31 races were found with schedules that have between 1 million and 108 million events, which suggests that they are rare and hard-to-find races.

show abstract

“…The soundness of the two-thread abstraction is argued in [3], and of race analysis via a single schedule in [14,19].…”

Section: Race Analysis Using Gpuverifymentioning

confidence: 99%

Warps and Atomics: Beyond Barrier Synchronization in the Verification of GPU Kernels

Bardsley

Donaldson

2014

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. We describe the design and implementation of methods to support reasoning about data races in GPU kernels where constructs other than the standard barrier primitive are used for synchronization. At one extreme we consider kernels that exploit implicit, coarse-grained synchronization between threads in the same warp, a feature provided by many architectures. At the other extreme we consider kernels that reduce or avoid barrier synchronization through the use of atomic operations. We discuss design decisions associated with providing support for warps and atomics in GPUVerify, a formal verification tool for OpenCL and CUDA kernels. We evaluate the practical impact of these design decisions using a large set of benchmarks, showing that warps can be supported in a scalable manner, that a coarse abstraction suffices for efficient reasoning about most practical uses of atomic operations, and that a novel, refined abstraction captures an important design pattern where atomic operations are used to compute unique array indices. Our evaluation revealed two previously unknown bugs in publicly available benchmark suites.

show abstract

Interleaving and Lock-Step Semantics for Analysis and Verification of GPU Kernels

Cited by 36 publications

References 18 publications

The Design and Implementation of a Verification Technique for GPU Kernels

The Design and Implementation of a Verification Technique for GPU Kernels

Race directed scheduling of concurrent programs

Warps and Atomics: Beyond Barrier Synchronization in the Verification of GPU Kernels

Contact Info

Product

Resources

About