CUDA Preemption Experiments

Introduction

This set of experiments are designed to test how kernels behave when preempted. To test this we use CUDA streams with priority and cuda events to synchronize between streams.

How to use this repository

There are a number of experiments in this repository. Compile all of them at once by running make in the root directory. Each experiment can be compiled individually by cding to the experiment directory and running make.

Experiments

1. one-stream

We create one low-priority stream and one high-priority stream. The low-priority stream and high-priority stream first run separately and we time them using cuda events. Then, the low-priority stream is launched first and the high-priority stream is launched after, preempting it.

1. n-streams

We launch n low-priority background kernels and 1 high-priority kernel. This experiment proceeds similarly to the one-stream experiment.

1. priority-interrupt

We launch one long-running, low-priority kernel and preempt it n consecutive times with a high-priority kernel. We measure the runtime of all kernels.

1. priority-streams

We create n streams and launch n/m kernels with each of the m stream priority levels. On current architectures m=2 so this experiment is not that interesting.

Other files

common includes helper functions from the NVIDIA CUDA Samples.

Note

To obtain more precise timing measurements, use nvprof or the NVIDIA Visual Profiler.

This experiment was created by modifying the NVIDIA CUDA StreamPriorities sample code.