Enabling c++20 on linux by jchen351 · Pull Request #17816 · microsoft/onnxruntime

jchen351 · 2023-10-06T16:55:57Z

Description

Enabling c++20 on Linux
Currently blocking issue:

Eigen using deprecated '[=]' lambda expression.

Motivation and Context

We want the latest and the greatest features from c++20.

depends on

change is_pod tp is_trivial #21071
Update OpenVino CI Ubuntu to 22.04 #21127
Update QNN pipeline pool #21482
Update NDK version in use-android-ndk.yml to 27

…ERTY LINK_FLAGS " -s ASSERTIONS=2 -s SAFE_HEAP=1 -s STACK_OVERFLOW_CHECK=2")"

…ING PROPERTY LINK_FLAGS " -s ASSERTIONS=2 -s SAFE_HEAP=1 -s STACK_OVERFLOW_CHECK=2")"" This reverts commit c72c40e.

snnn · 2023-10-06T17:51:38Z

We cannot do it for Linux CUDA build yet. Because the compiler we use is GCC 8, which is too low.

wschin · 2023-10-06T18:25:22Z

Just out of curiosity. Why do you need C++20? Would it crash onnxruntime build in my old conda environment?

snnn · 2023-10-06T20:28:31Z

For your second question, no. A new C++ standard consists of two things:

Core language features like constexpr/constinit, which do not need runtime support.
New std header files and functions like , which do not need a new runtime library.

The first one doesn't have impact on compatibility. The second one has impacts on non-Windows/non-Linux systems like macOS/iOS. We can avoid the issue by not using the headers that are not supported by the target system.

snnn · 2024-05-24T02:56:30Z

Wait this #20786 . The PR will update all GCCs to >=11. Now we still have GCC 8.

snnn · 2024-06-03T18:18:59Z

The GCC version is updated.

snnn · 2024-06-03T21:37:51Z

Just out of curiosity. Why do you need C++20? Would it crash onnxruntime build in my old conda environment?

We can use the new C++ features to write better code, and it shouldn't impact our compatibility at runtime. The language features are listed in https://gcc.gnu.org/projects/cxx-status.html in "C++20 Language Features" section

jchen351 · 2024-06-03T23:09:24Z

 template <typename InputType, typename ThresholdType, typename OutputType>
 class TreeAggregatorMax : public TreeAggregator<InputType, ThresholdType, OutputType> {
 public:
-  TreeAggregatorMax<InputType, ThresholdType, OutputType>(size_t n_trees,


All other similar class like TreeAggregatorMin and TreeAggregatorSum does not include type qualifier.

jchen351 · 2024-06-19T18:16:30Z

depends on #21071

…nux_c++20 # Conflicts: # onnxruntime/contrib_ops/cuda/math/cufft_plan_cache.h # orttraining/orttraining/training_ops/cuda/nn/conv_shared.h # orttraining/orttraining/training_ops/rocm/nn/conv_grad.cc

#pragma clang diagnostic pop #endif

# Conflicts: # onnxruntime/core/providers/cpu/ml/tree_ensemble_aggregator.h

edgchen1 · 2024-09-23T18:27:15Z

 class OrtMutex {
+#if defined(__clang__) && __cplusplus >= 202002L
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Wdeprecated-pragma"


I think this can be fixed - see this patch: e34c91c

edgchen1 · 2024-09-23T18:33:58Z

    int n = batch_size * sequence_length;
    concurrency::ThreadPool::TryBatchParallelFor(
+#if __cplusplus >= 202002L
+        context->GetOperatorThreadPool(), n, [=, this, &failed](ptrdiff_t index) {


see this core guideline:
https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#f54-when-writing-a-lambda-that-captures-this-or-any-class-data-member-dont-use--default-capture

I think they have a point. at least, we should consider how to make the code easier to understand.

snnn · 2024-10-02T04:13:57Z

Now we are using NDK 27?

# Conflicts: # include/onnxruntime/core/platform/ort_mutex.h # onnxruntime/core/providers/tensorrt/tensorrt_execution_provider.cc # onnxruntime/core/session/inference_session.cc

github-actions

You can commit the suggested changes from lintrunner.

github-actions · 2025-03-31T22:08:45Z

    NodeComputeInfo compute_info;
-    compute_info.create_state_func = [=](ComputeContext* context, FunctionState* state) {
+    compute_info.create_state_func = [=,this](ComputeContext* context, FunctionState* state) {
      std::unique_ptr<MIGraphXFuncState> p = std::make_unique<MIGraphXFuncState>();


Suggested change

NodeComputeInfo compute_info;

compute_info.create_state_func = [=](ComputeContext* context, FunctionState* state) {

compute_info.create_state_func = [=,this](ComputeContext* context, FunctionState* state) {

std::unique_ptr<MIGraphXFuncState> p = std::make_unique<MIGraphXFuncState>();

NodeComputeInfo compute_info;

compute_info.create_state_func = [=, this](ComputeContext* context, FunctionState* state) {

std::unique_ptr<MIGraphXFuncState> p = std::make_unique<MIGraphXFuncState>();

github-actions

You can commit the suggested changes from lintrunner.

[input_ids_data, word_embedding_length, sequence_length, position_ids_data, broadcast_position_ids, position_embedding_data, word_embedding_data, segment_ids_data, segment_embedding_data, position_embedding_length, segment_embedding_length, output_data, embedding_sum_data, gamma_data, beta_data, epsilon_value, hidden_size, &failed](ptrdiff_t index) {

github-actions

You can commit the suggested changes from lintrunner.

github-actions · 2025-04-02T01:18:05Z

+        context->GetOperatorThreadPool(), n,
+    [input_ids_data,
+     word_embedding_length,
+     sequence_length,
+     position_ids_data,
+     broadcast_position_ids,
+     position_embedding_data,
+     word_embedding_data,
+     segment_ids_data,
+     segment_embedding_data,
+     position_embedding_length,
+     segment_embedding_length,
+     output_data,
+     embedding_sum_data,
+     gamma_data,
+     beta_data,
+     epsilon_value,
+     hidden_size,
+     &failed](ptrdiff_t index) {
+#else


Suggested change

context->GetOperatorThreadPool(), n,

[input_ids_data,

word_embedding_length,

sequence_length,

position_ids_data,

broadcast_position_ids,

position_embedding_data,

word_embedding_data,

segment_ids_data,

segment_embedding_data,

position_embedding_length,

segment_embedding_length,

output_data,

embedding_sum_data,

gamma_data,

beta_data,

epsilon_value,

hidden_size,

&failed](ptrdiff_t index) {

#else

context->GetOperatorThreadPool(), n,

[input_ids_data,

word_embedding_length,

sequence_length,

position_ids_data,

broadcast_position_ids,

position_embedding_data,

word_embedding_data,

segment_ids_data,

segment_embedding_data,

position_embedding_length,

segment_embedding_length,

output_data,

embedding_sum_data,

gamma_data,

beta_data,

epsilon_value,

hidden_size,

&failed](ptrdiff_t index) {

#else

jchen351 added 2 commits October 6, 2023 09:55

Trying out c++20 on linux

81fbc65

Enabling "set_property(TARGET onnxruntime_test_all APPEND_STRING PROP…

c72c40e

…ERTY LINK_FLAGS " -s ASSERTIONS=2 -s SAFE_HEAP=1 -s STACK_OVERFLOW_CHECK=2")"

jchen351 requested a review from a team October 6, 2023 17:33

Revert "Enabling "set_property(TARGET onnxruntime_test_all APPEND_STR…

512cd4d

…ING PROPERTY LINK_FLAGS " -s ASSERTIONS=2 -s SAFE_HEAP=1 -s STACK_OVERFLOW_CHECK=2")"" This reverts commit c72c40e.

jchen351 added 4 commits October 13, 2023 13:26

Merge branch 'main' into Cjian/linux_c++20

e9fd1ef

ignore "-Werror=deprecated" for Eigen CXX11 Tensor

a3a69d0

-Wdeprecated-declarations

2e795b4

Merge branch 'refs/heads/main' into Cjian/linux_c++20

1a7efec

Merge remote-tracking branch 'origin/main' into Cjian/linux_c++20

c46ab4a

Merge branch 'refs/heads/main' into Cjian/linux_c++20

11f8d5b

snnn reviewed Jun 3, 2024

View reviewed changes

Comment thread include/onnxruntime/core/common/eigen_common_wrapper.h Outdated

jchen351 added 3 commits June 3, 2024 18:19

#pragma GCC diagnostic ignored "-Wdeprecated"

107788f

#pragma GCC diagnostic ignored "-Wdeprecated"

7646254

TreeAggregatorMin

661a465

jchen351 commented Jun 3, 2024

View reviewed changes

snnn reviewed Jun 3, 2024

View reviewed changes

Comment thread include/onnxruntime/core/common/eigen_common_wrapper.h Outdated

jchen351 added 5 commits June 3, 2024 20:08

#pragma GCC diagnostic ignored "-Wdeprecated"

450d994

[=

168beb8

qembed_layer_norm.cc:97:50

3736658

Merge branch 'refs/heads/main' into Cjian/linux_c++20

89d8e59

is_trivial

2d6abd5

jchen351 added 2 commits June 19, 2024 11:18

Merge branch 'refs/heads/main' into Cjian/linux_c++20

c66de02

Merge remote-tracking branch 'refs/remotes/origin/main' into Cjian/li…

31cb371

…nux_c++20 # Conflicts: # onnxruntime/contrib_ops/cuda/math/cufft_plan_cache.h # orttraining/orttraining/training_ops/cuda/nn/conv_shared.h # orttraining/orttraining/training_ops/rocm/nn/conv_grad.cc

jchen351 added 13 commits July 22, 2024 15:22

if(UNIX)

7386423

if(UNIX)

9750089

-Wdeprecated-pragma

42ee12f

-Wdeprecated-pragma

d837bfa

google nsync

a6b56f3

#if defined(

78c7b4a

#if defined(__clang__) && __cplusplus >= 202002L

4f2ebed

#pragma clang diagnostic pop #endif

[=, this]

7de09f7

maxThreadsPerBlock

524decf

#ifdef __GNUC__

0fe8720

deprecated

6eb70d5

Merge branch 'refs/heads/main' into Cjian/linux_c++20

fa8a8cc

Changing ostringstream to wostringstream

1471383

fs-eire mentioned this pull request Aug 30, 2024

[WIP] WebGPU EP [skip ci] #21904

Closed

7 tasks

jchen351 added 4 commits August 30, 2024 13:08

Using wostringstream only on Windows

fb7ee24

change ifndef to ifdef

fc63b36

Merge branch 'refs/heads/main' into Cjian/linux_c++20

3c31842

Merge branch 'refs/heads/main' into Cjian/linux_c++20

9e30f73

# Conflicts: # onnxruntime/core/providers/cpu/ml/tree_ensemble_aggregator.h

edgchen1 reviewed Sep 23, 2024

View reviewed changes

Merge branch 'main' into Cjian/linux_c++20

bed798e

# Conflicts: # include/onnxruntime/core/platform/ort_mutex.h # onnxruntime/core/providers/tensorrt/tensorrt_execution_provider.cc # onnxruntime/core/session/inference_session.cc

github-actions Bot reviewed Mar 31, 2025

View reviewed changes

jchen351 added 2 commits April 1, 2025 10:22

Update namespace timestamp_ns = std::chrono; for linux

2fab000

return static_cast<int>(flags) & static_cast<int>(MissingTrack::kTrue);

782ebaa

github-actions Bot reviewed Apr 1, 2025

View reviewed changes

Comment thread onnxruntime/core/providers/cpu/ml/tree_ensemble_aggregator.h

jchen351 added 2 commits April 1, 2025 18:03

Merge with main

e21fe01

github-actions Bot reviewed Apr 2, 2025

View reviewed changes

snnn closed this Jul 3, 2025

Uh oh!

Conversation

jchen351 commented Oct 6, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Motivation and Context

depends on

Uh oh!

snnn commented Oct 6, 2023

Uh oh!

wschin commented Oct 6, 2023

Uh oh!

snnn commented Oct 6, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

snnn commented May 24, 2024

Uh oh!

snnn commented Jun 3, 2024

Uh oh!

snnn commented Jun 3, 2024

Uh oh!

Uh oh!

jchen351 Jun 3, 2024

Choose a reason for hiding this comment

Uh oh!

Uh oh!

jchen351 commented Jun 19, 2024

Uh oh!

edgchen1 Sep 23, 2024

Choose a reason for hiding this comment

Uh oh!

edgchen1 Sep 23, 2024

Choose a reason for hiding this comment

Uh oh!

snnn commented Oct 2, 2024

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions Bot Mar 31, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions Bot Apr 2, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

jchen351 commented Oct 6, 2023 •

edited

Loading

snnn commented Oct 6, 2023 •

edited

Loading