Provable GPU Data-Races in Static Race Detection

03/24/2022
by   Dennis Liew, et al.
0

We extend the theory behind the Faial tool-chain, which can soundly prove that CUDA programs (aka, kernels) are data-race free using specialized behavioral types called memory access protocols (MAPs). In this paper we extend the theory of MAPs to characterize kernels for which alarms can be identified as true alarms. We introduce a core calculus for CUDA, which we named BabyCUDA, and a behavioral type system for it. We show that if a BabyCUDA program can be assigned a MAP, then any alarm raised by Faial for this program is a true alarm.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
04/27/2020

Static Race Detection and Mutex Safety and Liveness for Go Programs (extended version)

Go is a popular concurrent programming language thanks to its ability to...
research
12/27/2019

LLOV: A Fast Static Data-Race Checker for OpenMP Programs

In the era of Exascale computing, writing efficient parallel programs is...
research
04/30/2019

Behavioral Program Logic and LAGC Semantics without Continuations (Technical Report)

We present Behavioral Program Logic (BPL), a dynamic logic for trace pro...
research
11/24/2021

Predictive Data Race Detection for GPUs

The high degree of parallelism and relatively complicated synchronizatio...
research
09/06/2020

A Thread-Local Semantics and Efficient Static Analyses for Race Free Programs

Data race free (DRF) programs constitute an important class of concurren...
research
06/08/2022

Comparative Analysis of Dynamic Data Race Detection Techniques

The consequences of data races can be potentially very problematic [1], ...
research
11/17/2020

GPURepair: Automated Repair of GPU Kernels

This paper presents a tool for repairing errors in GPU kernels written i...

Please sign up or login with your details

Forgot password? Click here to reset