tequilaOS/platform_hardware_interfaces
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

NNAPI HAL: Remove priority from prepareModelFromCache_1_3

Browse source 
prepareModelFromCache_1_3 uses model cache and data cache that
represent a prepared model. Any argument that contributes in a
meaningful way to the prepared model is implicitly included in
this model cache and data cache. For example, "model" and
"executionPreference" appear in prepareModel_1_3 but not in
prepareModelFromCache_1_3 because they are implicitly included
in the model cache and data cache. In a similar way, because
it could affect the resulting model, "priority" should be
removed from prepareModelFromCache_1_3.

Fixes: 148802784
Test: mma
Test: VtsHalNeuralnetworksV1_3TargetTest
Change-Id: I518e493ead8aa97220b16370cce8523b425c378c
Merged-In: I518e493ead8aa97220b16370cce8523b425c378c
(cherry picked from commit c90250ff57)
This commit is contained in:
Michael Butler 2020-01-31 18:39:22 -08:00 committed by Slava Shklyaev
parent 124659e450
commit 68a868b2ac
3 changed files with 3 additions and 10 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
current.txt
View file

@ -625,7 +625,7 @@ bbeee9604128ede83ee755b67e73b5ad29e6e1dbac9ec41fea6ffe2745b0c50a android.hardwar
adb0efdf1462e9b2e742c0dcadd598666aac551f178be06e755bfcdf5797abd0 android.hardware.keymaster@4.1::IOperation
ac429fca0da4ce91218768ec31b64ded88251f8a26d8c4f27c06abdc5b1926d9 android.hardware.keymaster@4.1::types
65c16331e57f6dd68b3971f06f78fe9e3209afb60630c31705aa355f9a52bf0d android.hardware.neuralnetworks@1.3::IBuffer
d1f382d14e1384b907d5bb5780df7f01934650d556fedbed2f15a90773c657d6 android.hardware.neuralnetworks@1.3::IDevice
9b41dd49e2dcc2ecb4243d03f8421d72494ada5cf2945bff88f0019eeca56923 android.hardware.neuralnetworks@1.3::IDevice
4167dc3ad35e9cd0d2057d4868c7675ae2c3c9d05bbd614c1f5dccfa5fd68797 android.hardware.neuralnetworks@1.3::IExecutionCallback
2fa3679ad7c94b5e88724adcd560c561041068a4ca565c63830e68101988746a android.hardware.neuralnetworks@1.3::IFencedExecutionCallback
237b23b126a66f3432658020fed78cdd06ba6297459436fe6bae0ba753370833 android.hardware.neuralnetworks@1.3::IPreparedModel

9
neuralnetworks/1.3/IDevice.hal
View file

@ -260,11 +260,6 @@ interface IDevice extends @1.2::IDevice {
* the model, the callback object must be invoked with the appropriate
* ErrorStatus value and nullptr for the IPreparedModel.
*
* The model is prepared with a priority. This priority is relative to other
* prepared models owned by the same client. Higher priority executions may
* use more compute resources than lower priority executions, and may
* preempt or starve lower priority executions.
*
* prepareModelFromCache_1_3 can be called with an optional deadline. If the
* model is not able to prepared before the provided deadline, the model
* preparation must be aborted, and either {@link
@ -284,8 +279,6 @@ interface IDevice extends @1.2::IDevice {
* used with different shapes of inputs on different (possibly concurrent)
* executions.
*
* @param priority The priority of the prepared model relative to other
* prepared models owned by the client.
* @param deadline The time by which the model must be prepared. If the
* model cannot be prepared by the deadline, the preparation must be
* aborted.
@ -318,7 +311,7 @@ interface IDevice extends @1.2::IDevice {
* met
* - RESOURCE_EXHAUSTED_* if the task was aborted by the driver
*/
prepareModelFromCache_1_3(Priority priority, OptionalTimePoint deadline,
prepareModelFromCache_1_3(OptionalTimePoint deadline,
vec<handle> modelCache, vec<handle> dataCache,
uint8_t[Constant:BYTE_SIZE_OF_CACHE_TOKEN] token,
IPreparedModelCallback callback)

2
neuralnetworks/1.3/vts/functional/CompilationCachingTests.cpp
View file

@ -370,7 +370,7 @@ class CompilationCachingTestBase : public testing::Test {
sp<PreparedModelCallback> preparedModelCallback = new PreparedModelCallback();
hidl_array<uint8_t, sizeof(mToken)> cacheToken(mToken);
Return<ErrorStatus> prepareLaunchStatus = kDevice->prepareModelFromCache_1_3(
kDefaultPriority, {}, modelCache, dataCache, cacheToken, preparedModelCallback);
{}, modelCache, dataCache, cacheToken, preparedModelCallback);
ASSERT_TRUE(prepareLaunchStatus.isOk());
if (static_cast<ErrorStatus>(prepareLaunchStatus) != ErrorStatus::NONE) {
*preparedModel = nullptr;