Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[CPU][ARM] Make f16 precision as default for CNN #22839

Merged
merged 4 commits into from
Mar 19, 2024
Merged

[CPU][ARM] Make f16 precision as default for CNN #22839

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
4d8fb9b
init
alvoron Feb 14, 2024
e9c9698
force f32 in smoke_CompileModel tests
alvoron Feb 15, 2024
1414716
force f32 precision in AutoBatching_Test
alvoron Feb 15, 2024
17f7269
Merge branch 'master' into alvoron_arm_f16_default
alvoron Mar 18, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 1 addition & 4 deletions src/plugins/intel_cpu/src/config.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -369,10 +369,7 @@ void Config::readProperties(const ov::AnyMap& prop, const ModelType modelType) {
if (executionMode == ov::hint::ExecutionMode::PERFORMANCE) {
inferencePrecision = ov::element::f32;
#if defined(OV_CPU_ARM_ENABLE_FP16)
// fp16 precision is used as default precision on ARM for non-convolution networks
// fp16 ACL convolution is slower than fp32
if (modelType != ModelType::CNN)
inferencePrecision = ov::element::f16;
inferencePrecision = ov::element::f16;
#else
if (mayiuse(avx512_core_bf16))
inferencePrecision = ov::element::bf16;
Expand Down
1 change: 1 addition & 0 deletions src/tests/functional/plugin/shared/include/behavior/ov_plugin/auto_batching_tests.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,6 +71,7 @@ class AutoBatching_Test : public OVPluginTestBase,

if (target_device.find("CPU") != std::string::npos) {
config.insert(ov::num_streams(static_cast<int32_t>(num_streams)));
config.insert(ov::hint::inference_precision(ov::element::f32));
}
// minimize timeout to reduce test time
config.insert(ov::auto_batch_timeout(1));
Expand Down
6 changes: 4 additions & 2 deletions src/tests/functional/plugin/shared/include/behavior/ov_plugin/core_threading.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -487,7 +487,8 @@ TEST_P(CoreThreadingTestsWithIter, smoke_CompileModel_Accuracy_SingleCore) {
}

auto getOutputBlob = [&](ov::Core& core) {
auto compiled_model = core.compile_model(model, target_device);
ov::AnyMap f32_precision_property = {{ov::hint::inference_precision.name(), ov::element::f32.to_string()}};
auto compiled_model = core.compile_model(model, target_device, f32_precision_property);
auto req = compiled_model.create_infer_request();
for (const auto& input : inputs) {
req.set_tensor(input.first, input.second);
Expand Down Expand Up @@ -530,7 +531,8 @@ TEST_P(CoreThreadingTestsWithIter, smoke_CompileModel_Accuracy_MultipleCores) {
}

auto getOutputBlob = [&](ov::Core& core) {
auto compiled_model = core.compile_model(model, target_device);
ov::AnyMap f32_precision_property = {{ov::hint::inference_precision.name(), ov::element::f32.to_string()}};
auto compiled_model = core.compile_model(model, target_device, f32_precision_property);
auto req = compiled_model.create_infer_request();
for (const auto& input : inputs) {
req.set_tensor(input.first, input.second);
Expand Down
Loading