C++20 by ax3l · Pull Request #6606 · BLAST-WarpX/warpx

ax3l · 2026-02-24T07:30:48Z

Switch to C++20. AMReX is making the switch soon (AMReX-Codes/amrex#4840) and is preparing CI in AMReX-Codes/amrex#4976 so we can also generally pull up the requirements.

wait for 26.03 release
patch toml11 / use clang 14+ (for CI, we can do Use most recent main branch of toml11 openPMD/openPMD-api#1874 )
merge for the 26.05 release cycle

ax3l · 2026-02-24T16:56:17Z

GCC 12 in C++20 mode might have a false positive here (cc @lucafedeli88 ?):

                 from /home/runner/work/warpx/warpx/Source/ablastr/utils/TextMsg.H:11,
                 from /home/runner/work/warpx/warpx/Source/ablastr/utils/TextMsg.cpp:8:
In static member function ‘static constexpr std::char_traits<char>::char_type* std::char_traits<char>::copy(char_type*, const char_type*, std::size_t)’,
    inlined from ‘static constexpr void std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::_S_copy(_CharT*, const _CharT*, size_type) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:431:21,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Allocator>& std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::_M_replace(size_type, size_type, const _CharT*, size_type) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.tcc:532:22,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Alloc>& std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::replace(size_type, size_type, const _CharT*, size_type) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:2179:19,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Alloc>& std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::insert(size_type, const _CharT*) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:1936:22,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Allocator> std::operator+(const _CharT*, __cxx11::basic_string<_CharT, _Traits, _Allocator>&&) [with _CharT = char; _Traits = char_traits<char>; _Alloc = allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:3549:36,
    inlined from ‘void ablastr::utils::TextMsg::Assert(const char*, const char*, int, const std::string&)’ at /home/runner/work/warpx/warpx/Source/ablastr/utils/TextMsg.cpp:79:38:
/usr/include/c++/12/bits/char_traits.h:435:56: error: ‘void* __builtin_memcpy(void*, const void*, long unsigned int)’ accessing 9223372036854775810 or more bytes at offsets [2, 9223372036854775807] and 1 may overlap up to 9223372036854775813 bytes at offset -3 [-Werror=restrict]
  435 |         return static_cast<char_type*>(__builtin_memcpy(__s1, __s2, __n));
      |                                        ~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~

lucafedeli88 · 2026-02-25T02:02:21Z

GCC 12 in C++20 mode might have a false positive here (cc @lucafedeli88 ?):

                 from /home/runner/work/warpx/warpx/Source/ablastr/utils/TextMsg.H:11,
                 from /home/runner/work/warpx/warpx/Source/ablastr/utils/TextMsg.cpp:8:
In static member function ‘static constexpr std::char_traits<char>::char_type* std::char_traits<char>::copy(char_type*, const char_type*, std::size_t)’,
    inlined from ‘static constexpr void std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::_S_copy(_CharT*, const _CharT*, size_type) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:431:21,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Allocator>& std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::_M_replace(size_type, size_type, const _CharT*, size_type) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.tcc:532:22,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Alloc>& std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::replace(size_type, size_type, const _CharT*, size_type) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:2179:19,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Alloc>& std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::insert(size_type, const _CharT*) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:1936:22,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Allocator> std::operator+(const _CharT*, __cxx11::basic_string<_CharT, _Traits, _Allocator>&&) [with _CharT = char; _Traits = char_traits<char>; _Alloc = allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:3549:36,
    inlined from ‘void ablastr::utils::TextMsg::Assert(const char*, const char*, int, const std::string&)’ at /home/runner/work/warpx/warpx/Source/ablastr/utils/TextMsg.cpp:79:38:
/usr/include/c++/12/bits/char_traits.h:435:56: error: ‘void* __builtin_memcpy(void*, const void*, long unsigned int)’ accessing 9223372036854775810 or more bytes at offsets [2, 9223372036854775807] and 1 may overlap up to 9223372036854775813 bytes at offset -3 [-Werror=restrict]
  435 |         return static_cast<char_type*>(__builtin_memcpy(__s1, __s2, __n));
      |                                        ~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~

ah, that's really weird...

lucafedeli88 · 2026-03-07T10:01:58Z

GCC 12 in C++20 mode might have a false positive here (cc @lucafedeli88 ?):

                 from /home/runner/work/warpx/warpx/Source/ablastr/utils/TextMsg.H:11,
                 from /home/runner/work/warpx/warpx/Source/ablastr/utils/TextMsg.cpp:8:
In static member function ‘static constexpr std::char_traits<char>::char_type* std::char_traits<char>::copy(char_type*, const char_type*, std::size_t)’,
    inlined from ‘static constexpr void std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::_S_copy(_CharT*, const _CharT*, size_type) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:431:21,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Allocator>& std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::_M_replace(size_type, size_type, const _CharT*, size_type) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.tcc:532:22,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Alloc>& std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::replace(size_type, size_type, const _CharT*, size_type) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:2179:19,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Alloc>& std::__cxx11::basic_string<_CharT, _Traits, _Alloc>::insert(size_type, const _CharT*) [with _CharT = char; _Traits = std::char_traits<char>; _Alloc = std::allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:1936:22,
    inlined from ‘constexpr std::__cxx11::basic_string<_CharT, _Traits, _Allocator> std::operator+(const _CharT*, __cxx11::basic_string<_CharT, _Traits, _Allocator>&&) [with _CharT = char; _Traits = char_traits<char>; _Alloc = allocator<char>]’ at /usr/include/c++/12/bits/basic_string.h:3549:36,
    inlined from ‘void ablastr::utils::TextMsg::Assert(const char*, const char*, int, const std::string&)’ at /home/runner/work/warpx/warpx/Source/ablastr/utils/TextMsg.cpp:79:38:
/usr/include/c++/12/bits/char_traits.h:435:56: error: ‘void* __builtin_memcpy(void*, const void*, long unsigned int)’ accessing 9223372036854775810 or more bytes at offsets [2, 9223372036854775807] and 1 may overlap up to 9223372036854775813 bytes at offset -3 [-Werror=restrict]
  435 |         return static_cast<char_type*>(__builtin_memcpy(__s1, __s2, __n));
      |                                        ~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~

ah, that's really weird...

I've tried compiling this branch on my machine (GCC 15.2.0) and the issue does not show up. So it may be a false positive.

lucafedeli88 · 2026-03-11T10:59:39Z

@ax3l , I am increasingly convinced that the g++12 warning is a false positive (see google/googletest#4570). Maybe we can write a workaround if we need to use that version of the compiler. FYI, with v12.5.0 I don't see the issue on my machine. Which specific version of g++ are you using ?

WeiqunZhang · 2026-03-31T17:40:19Z

I think for the failed cuda checks, we just need to switch from 11.7 to 12.2.

Switch to C++20. AMReX is making the switch so we can also generally pull up the requirements.

ax3l · 2026-04-23T17:47:57Z

 # Old GCCs: std::filesystem as a separate library
 try_compile(CXX_HAS_STD_FS "${WarpX_BINARY_DIR}/temp"
    "${WarpX_SOURCE_DIR}/cmake/has_filesystem.cpp"
-    CMAKE_FLAGS -DCMAKE_CXX_STANDARD=17 -DCMAKE_CXX_STANDARD_REQUIRED=ON
+    CMAKE_FLAGS -DCMAKE_CXX_STANDARD=20 -DCMAKE_CXX_STANDARD_REQUIRED=ON


The old GCC we supported here does probably not support C++20 anyway. But we can remove this another time, does not hurt.

ax3l · 2026-04-23T17:49:14Z

    endforeach()
    foreach(warpx_tgt IN LISTS _ALL_TARGETS)
-        target_compile_features(${warpx_tgt} PUBLIC cuda_std_17)
+        target_compile_features(${warpx_tgt} PUBLIC cuda_std_20)


Probably worth bumping minimum CMake eventually:
(Source: Google/Gemini. Might need validation when we act on it.)

Co-authored-by: Edoardo Zoni <ezoni@lbl.gov>

EZoni · 2026-04-23T20:28:33Z

@lucafedeli88

After merging the Adastra update from #6591, we need to re-update it here (see merge conflict). Do you want to test it manually first or can we go ahead with updating the script trivially?

lucafedeli88 · 2026-04-23T21:36:10Z

@lucafedeli88

After merging the Adastra update from #6591, we need to re-update it here (see merge conflict). Do you want to test it manually first or can we go ahead with updating the script trivially?

@EZoni , I should have fixed the merge conflict. I am 99.9% sure that it will work as it is. To be sure, I will test it tomorrow.

lucafedeli88 · 2026-04-23T21:41:45Z

@ax3l , this may be useful to fix the clang-tidy CI test : #6800 .

lucafedeli88 · 2026-04-24T08:39:34Z

@ax3l , I've ported my clang-tidy fix to your PR (see #6803) in order to test that everything is OK. Feel free to merge the changes in .clang-tidy , github/workflows/clang_tidy.yml, Docs/source/developers/how_to_run_clang_tidy.rst, and Tools/Linter/runClangTidy.sh into your PR.
In essence the changes are:

bump clang-tidy from v17 to v18 because the clang-diagnostic-unknown-pragmas issue that you are observing is a false positive. Updating the tool makes it go away.
disable the new checks in v18 and the checks relevant to C++20 (we may want to enable many of them in the future, but certainly not in this PR)
update the documentation and a tool to take the version change into account

lucafedeli88

LGTM. Thanks for this!

ax3l · 2026-04-24T17:06:37Z

Thank you both for your help! 🎉 ✨

ax3l requested a review from a team February 24, 2026 07:30

ax3l added the install label Feb 24, 2026

ax3l commented Feb 24, 2026

View reviewed changes

Comment thread Docs/source/install/dependencies.rst Outdated

ax3l commented Feb 24, 2026

View reviewed changes

Comment thread Docs/source/developers/faq.rst Outdated

ax3l mentioned this pull request Feb 24, 2026

Moving to C++20 AMReX-Codes/amrex#4840

Closed

RemiLehe changed the title ~~C++20~~ [WIP] C++20 Feb 24, 2026

ax3l requested a review from WeiqunZhang March 31, 2026 04:04

ax3l changed the title ~~[WIP] C++20~~ C++20 Mar 31, 2026

ax3l force-pushed the topic-cxx20 branch from 61ebf21 to 515fee3 Compare March 31, 2026 04:06

ax3l assigned WeiqunZhang Mar 31, 2026

EZoni mentioned this pull request Apr 22, 2026

Dependencies: weekly update #6790

Merged

1 task

EZoni reviewed Apr 22, 2026

View reviewed changes

Comment thread .github/workflows/dependencies/nvcc.sh Outdated

ax3l changed the title ~~C++20~~ C++20 CI Apr 23, 2026

ax3l changed the title ~~C++20 CI~~ C++20 Apr 23, 2026

ax3l and others added 4 commits April 23, 2026 10:44

C++20

59e551f

Switch to C++20. AMReX is making the switch so we can also generally pull up the requirements.

Use CUDA 12.2

a490c5e

Update nvcc.sh

ee81494

Remove SENSEI

1307940

ax3l force-pushed the topic-cxx20 branch from 61f608c to 1307940 Compare April 23, 2026 17:45

ax3l commented Apr 23, 2026

View reviewed changes

WeiqunZhang approved these changes Apr 23, 2026

View reviewed changes

nvcc: simplify patch

6362962

Co-authored-by: Edoardo Zoni <ezoni@lbl.gov>

ax3l force-pushed the topic-cxx20 branch from 652b239 to 6362962 Compare April 23, 2026 17:55

EZoni reviewed Apr 23, 2026

View reviewed changes

Comment thread Docs/source/install/dependencies.rst

Merge branch 'development' into topic-cxx20

2956310

lucafedeli88 mentioned this pull request Apr 24, 2026

[DO NOT MERGE] Test clang-tidy fix to be merged with PR 6606 (C++20) #6803

Closed

Port clang-tidy version update from BLAST-WarpX#6803

b27c641

EZoni requested review from EZoni and lucafedeli88 April 24, 2026 16:23

lucafedeli88 approved these changes Apr 24, 2026

View reviewed changes

EZoni approved these changes Apr 24, 2026

View reviewed changes

ax3l enabled auto-merge (squash) April 24, 2026 17:07

ax3l merged commit 44f7bc1 into BLAST-WarpX:development Apr 24, 2026
47 of 48 checks passed

ax3l deleted the topic-cxx20 branch April 25, 2026 15:56

lucafedeli88 mentioned this pull request Apr 29, 2026

[WIP] clang-tidy CI test : enable performance-enum-size check #6822

Closed

Conversation

ax3l commented Feb 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ax3l commented Feb 24, 2026

Uh oh!

lucafedeli88 commented Feb 25, 2026

Uh oh!

lucafedeli88 commented Mar 7, 2026

Uh oh!

lucafedeli88 commented Mar 11, 2026

Uh oh!

WeiqunZhang commented Mar 31, 2026

Uh oh!

Uh oh!

ax3l Apr 23, 2026

Choose a reason for hiding this comment

Uh oh!

ax3l Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

EZoni commented Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lucafedeli88 commented Apr 23, 2026

Uh oh!

lucafedeli88 commented Apr 23, 2026

Uh oh!

lucafedeli88 commented Apr 24, 2026

Uh oh!

lucafedeli88 left a comment

Choose a reason for hiding this comment

Uh oh!

ax3l commented Apr 24, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

ax3l commented Feb 24, 2026 •

edited

Loading

ax3l Apr 23, 2026 •

edited

Loading

EZoni commented Apr 23, 2026 •

edited

Loading