OSDN Git Service

NNAPI Concurrent Query Management -- HAL and VTS update
authorMichael Butler <butlermichael@google.com>
Fri, 1 Sep 2017 17:59:38 +0000 (10:59 -0700)
committerDavid Gross <dgross@google.com>
Fri, 1 Sep 2017 18:07:56 +0000 (11:07 -0700)
commit0e2ac1bb885d7ba77e80a082526fd1b510190a72
tree70cbf2c272e0d7d56b6ee1499fe7ed2f0b4cf3f4
parent5d143d2a0ff7c8d908aab13727b587003896d031
NNAPI Concurrent Query Management -- HAL and VTS update

The NNAPI requires requests on a model to be asynchronously
processed. This CL implements a basic Event that can later
be used to block the runtime thread until the asynchronous
request has completed.

Bug: 63905942
Test: VtsHalNeuralnetworksV1_0TargetTest (32-bit, 64-bit) with sample driver enabled by cherry-pick
      frameworks/ml/nn/runtime/test with and without sample driver enabled
Change-Id: Ie27a574aaaac312e7cbb731750f9c06278357a1c
neuralnetworks/1.0/Android.bp
neuralnetworks/1.0/IEvent.hal [new file with mode: 0644]
neuralnetworks/1.0/IPreparedModel.hal
neuralnetworks/1.0/types.hal
neuralnetworks/1.0/vts/functional/Android.bp
neuralnetworks/1.0/vts/functional/Event.cpp [new file with mode: 0644]
neuralnetworks/1.0/vts/functional/Event.h [new file with mode: 0644]
neuralnetworks/1.0/vts/functional/VtsHalNeuralnetworksV1_0TargetTest.cpp
neuralnetworks/1.0/vts/functional/VtsHalNeuralnetworksV1_0TargetTest.h