EDG over clang emits bfloat16 numbers with bf16 suffix, which clang does not support

[ilazaric]

Issue Summary

Clang seems to support __bf16 type, but not the bf16 suffix.
EDG likes to emit bfloat16 constants with the bf16 suffix.

Reproducing Source Code

__bf16 foo = 0;

Command-line Options

eccp --clang_version 210000 repro.cpp

Type of Issue

[Back End] There is a problem with the generated C code.

Additional Details

Generated C code:

/* Translated by the Edison Design Group C++/C front end (version 7.0) */
/* Mon Oct  5 03:00:05 2026 */
extern int __EDGCPFE__7_0;
void *memcpy(void *,const void *,unsigned long);
void *memset(void *,int,unsigned long);

# 1 "bf.cpp"
extern __bf16 foo; __bf16 foo = (0.0bf16);

0.0bf16 is not supported by clang.

This issue can also be observed in EDG codebase/build too:

$ podman run --rm -e EDG_BASE="$PWD/bases/docker/dev-env/clang" -w "$PWD" -v "$PWD":"$PWD" docker.io/edgcpp/ci-runner:latest cmake --preset linux-clang-release -DEDG_CPP_RT_LIBS=linux_x86_64
[...]
$ podman run --rm -e EDG_BASE="$PWD/bases/docker/dev-env/clang" -w "$PWD" -v "$PWD":"$PWD" docker.io/edgcpp/ci-runner:latest cmake --build build/clang-release/ --target "$PWD/build/clang-release/lib_src/edg-cpp-rt-cache/c99_complex.o"
[...]
[86/86] Generating ../../edg-cpp-rt-cache/c99_complex.o
FAILED: [code=129] lib_src/edg-cpp-rt-cache/c99_complex.o /home/ilazaric/repos/ALL/submodules/edg-compiler/build/clang-release/lib_src/edg-cpp-rt-cache/c99_complex.o 
cd /home/ilazaric/repos/ALL/submodules/edg-compiler/build/clang-release/lib_src/cmake/default && /home/ilazaric/repos/ALL/submodules/edg-compiler/build/clang-release/bin/eccp --building_runtime --c++23 -g -DCHECKING=1 -DDEBUG=1 -DUSE_STDINT_HEADER -DSYSTEM_RUNTIME_HAS_IA64_THREAD_ATEXIT=0 --no_standard_includes --sys_include=/home/ilazaric/repos/ALL/submodules/edg-compiler/include_c++ --sys_include=/home/ilazaric/repos/ALL/submodules/edg-compiler/include_c99 --sys_include=/usr/lib/gcc/x86_64-redhat-linux/15/include --sys_include=/usr/local/include --sys_include=/usr/include --g++ --c_to_obj_option -fPIC --diag_suppress lossy_conversion -o /home/ilazaric/repos/ALL/submodules/edg-compiler/build/clang-release/lib_src/edg-cpp-rt-cache/c99_complex.o -c /home/ilazaric/repos/ALL/submodules/edg-compiler/lib_src/c99_complex.c
eccp: diagnostics generated from compilation of c99_complex.int.c:
/home/ilazaric/repos/ALL/submodules/edg-compiler/lib_src/c99_complex.c:254:187: error: invalid suffix 'bf16' on floating constant
  254 | _Complex_bfloat16 __c99_ibfloat16_to_cbfloat16( _EDG_bfloat16_t __372_1_j) { auto struct _Complex_bfloat16 __T198811768; auto _Complex_bfloat16 __372_1_r; (((__372_1_r._Vals))[0]) = (0.0bf16); (((__372_1_r._Vals))[1]) = ((_EDG_bfloat16_t)__372_1_j); { __T198811768 = __372_1_r; return __T198811768; } 
      |                                                                                                                                                                                           ^
/home/ilazaric/repos/ALL/submodules/edg-compiler/lib_src/c99_complex.c:255:243: error: invalid suffix 'bf16' on floating constant
  255 | _Complex_bfloat16 __c99_bfloat16_to_cbfloat16( _EDG_bfloat16_t __373_1_j) { auto struct _Complex_bfloat16 __T198816936; auto _Complex_bfloat16 __373_1_r; (((__373_1_r._Vals))[0]) = ((_EDG_bfloat16_t)__373_1_j); (((__373_1_r._Vals))[1]) = (0.0bf16); { __T198816936 = __373_1_r; return __T198816936; } 
      |                                                                                                                                                                                                                                                   ^
2 errors generated.
eccp: end of diagnostics from compilation of c99_complex.int.c
eccp: clang compilation of c99_complex.int.c returned an exit status of 1
ninja: build stopped: subcommand failed.

Reproducing (Standard) Configuration(s)

linux-clang-release

Reproducing (Non-standard) Configuration

No response

I acknowledge that:

• I have checked open issues and existing discussions before filing this report.
• The reproducing source code that I have provided does not depend on headers that I have not included with my submission.
• I have reduced the reproducing source code to the best of my ability.
• I have selected at least one standard configuration or uploaded a non-standard configuration dump.
• I understand code shared via this form is NOT confidential.
• I have permission to share the reproducing source code and grant permission for the project to use the source code (or derivatives of the source code) to create tests.
• I agree that any tests created using the reproducing source code (or deriatives of it) will be licensed under project license unless I included a LICENSE.txt file in my .zip or .tar.gz upload. If I included a LICENSE.txt file, any tests created using the reproducing source code will be governed by the terms of the license I provided instead.

[daveedvdv-nvidia]

Any suggestions? Should we render it as ((__bf16)<fp-literal>) when targeting Clang?

[ilazaric]

This approach sounds good to me, as float can represent bloat16 exactly.