build: Update C++ standard from c++17 to c++20

CMakeLists.txt

@@ -36,7 +36,7 @@ option(TRITON_CORE_HEADERS_ONLY "Build only headers and stub" ON)
 #
 # Specifying min required C++ standard
 #
-set(TRITON_MIN_CXX_STANDARD 17 CACHE STRING "The minimum C++ standard which features are requested to build this target.")
+set(TRITON_MIN_CXX_STANDARD 20 CACHE STRING "The minimum C++ standard which features are requested to build this target.")
 
 #
 # Triton Server API

python/tritonserver/CMakeLists.txt

@@ -1,4 +1,4 @@
-# Copyright (c) 2023-2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# Copyright (c) 2023-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
 # modification, are permitted provided that the following conditions
@@ -58,7 +58,7 @@ target_link_libraries(
   triton-core-serverapi           # from repo-core
   triton-core-serverstub          # from repo-core
 )
-target_compile_features(python-bindings PRIVATE cxx_std_17)
+target_compile_features(python-bindings PRIVATE cxx_std_${TRITON_MIN_CXX_STANDARD})
 
 set_property(TARGET python-bindings PROPERTY OUTPUT_NAME triton_bindings)
 # Add Triton library default path in 'rpath' for runtime library lookup

src/CMakeLists.txt

@@ -133,7 +133,7 @@ set(
   infer_trace.cc
   instance_queue.cc
   label_provider.cc
-  memory.cc
+  triton_memory.cc
   metric_model_reporter.cc
   metrics.cc
   metric_family.cc
@@ -180,7 +180,7 @@ set(
   infer_trace.h
   instance_queue.h
   label_provider.h
-  memory.h
+  triton_memory.h
   metric_model_reporter.h
   metrics.h
   metric_family.h

src/backend_model.cc

@@ -682,7 +682,7 @@ TritonModel::PrepareInstances(
 
         // Note that the local variables should be captured by value
         creation_results.emplace_back(
-            std::async(launch_policy, [=, &instance_mu]() {
+            std::async(launch_policy, [=, this, &instance_mu]() {
               // The requested instance did not match an existing instance.
               // Create a new instance.
               std::shared_ptr<TritonModelInstance> new_instance;

src/backend_model_instance.h

@@ -1,4 +1,4 @@
-// Copyright 2020-2022, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+// Copyright 2020-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 //
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions
@@ -33,13 +33,13 @@
 #include <thread>
 
 #include "constants.h"
-#include "memory.h"
 #include "metric_model_reporter.h"
 #include "model_config.pb.h"
 #include "model_config_utils.h"
 #include "server_message.h"
 #include "status.h"
 #include "triton/common/sync_queue.h"
+#include "triton_memory.h"
 
 namespace triton { namespace core {
 

src/infer_request.h

@@ -34,12 +34,12 @@
 #include "infer_response.h"
 #include "infer_stats.h"
 #include "infer_trace.h"
-#include "memory.h"
 #include "response_allocator.h"
 #include "sequence_state.h"
 #include "status.h"
 #include "triton/common/logging.h"
 #include "triton/common/model_config.h"
+#include "triton_memory.h"
 #include "tritonserver_apis.h"
 
 namespace triton { namespace core {

src/sequence_state.cc

@@ -29,9 +29,9 @@
 #include <cstdint>
 
 #include "cuda_utils.h"
-#include "memory.h"
 #include "model_config_utils.h"
 #include "triton/common/logging.h"
+#include "triton_memory.h"
 
 namespace triton { namespace core {
 

src/sequence_state.h

@@ -28,9 +28,9 @@
 #include <map>
 #include <memory>
 
-#include "memory.h"
 #include "status.h"
 #include "triton/common/model_config.h"
+#include "triton_memory.h"
 
 #pragma once
 

src/test/CMakeLists.txt

@@ -75,13 +75,13 @@ set(
 set(
   MEMORY_SRCS
   ../buffer_attributes.cc
-  ../memory.cc
+  ../triton_memory.cc
 )
 
 set(
   MEMORY_HDRS
   ../buffer_attributes.h
-  ../memory.h
+  ../triton_memory.h
 )
 
 #

src/test/memory_test.cc

@@ -1,4 +1,4 @@
-// Copyright 2020-2022, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+// Copyright 2020-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 //
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions
@@ -23,8 +23,6 @@
 // OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-#include "memory.h"
-
 #include <cuda_runtime_api.h>
 
 #include "cuda_block_manager.h"
@@ -33,6 +31,7 @@
 #include "gtest/gtest.h"
 #include "pinned_memory_manager.h"
 #include "triton/core/tritonserver.h"
+#include "triton_memory.h"
 
 namespace tc = triton::core;
 

src/test/response_cache_test.cc

@@ -1,4 +1,4 @@
-// Copyright 2021-2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+// Copyright 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 //
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions
@@ -29,8 +29,8 @@
 #include "cache_manager.h"
 #include "gtest/gtest-spi.h"
 #include "gtest/gtest.h"
-#include "memory.h"
 #include "triton/common/logging.h"
+#include "triton_memory.h"
 
 namespace tc = triton::core;
 

src/memory.cc → src/triton_memory.cc

@@ -1,4 +1,4 @@
-// Copyright 2018-2022, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+// Copyright 2018-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 //
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions
@@ -24,7 +24,7 @@
 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-#include "memory.h"
+#include "triton_memory.h"
 
 #include "pinned_memory_manager.h"
 #include "triton/common/logging.h"

src/memory.h → src/triton_memory.h

@@ -1,4 +1,4 @@
-// Copyright 2018-2022, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+// Copyright 2018-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 //
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions