Keil Logo

Technical Support

On-Line Manuals

Compiler User Guide

Preface Overview of the Compiler Getting Started with the Compiler Compiler Features Compiler Coding Practices Compiler Diagnostic Messages Using the Inline and Embedded Assemblers of the AR Compiler Command-line Options -Aopt --allow_fpreg_for_nonfpdata, --no_allow_fpreg_for_ --allow_null_this, --no_allow_null_this --alternative_tokens, --no_alternative_tokens --anachronisms, --no_anachronisms --apcs=qualifier...qualifier --arm --arm_only --asm --asm_dir=directory_name --autoinline, --no_autoinline --bigend --bitband --branch_tables, --no_branch_tables --brief_diagnostics, --no_brief_diagnostics --bss_threshold=num -c -C --c90 --c99 --code_gen, --no_code_gen --comment_section, --no_comment_section --compatible=name --compile_all_input, --no_compile_all_input --conditionalize, --no_conditionalize --cpp --cpp11 --cpp_compat --cpu=list --cpu=name --create_pch=filename -Dname[(parm-list)][=def] --data_reorder, --no_data_reorder --debug, --no_debug --debug_macros, --no_debug_macros --default_extension=ext --dep_name, --no_dep_name --depend=filename --depend_dir=directory_name --depend_format=string --depend_single_line, --no_depend_single_line --depend_system_headers, --no_depend_system_header --depend_target=target --diag_error=tag[,tag,...] --diag_remark=tag[,tag,...] --diag_style=arm|ide|gnu compiler option --diag_suppress=tag[,tag,...] --diag_suppress=optimizations --diag_warning=tag[,tag,...] --diag_warning=optimizations --dollar, --no_dollar --dwarf2 --dwarf3 -E --echo --emit_frame_directives, --no_emit_frame_directive --enum_is_int --errors=filename --exceptions, --no_exceptions --exceptions_unwind, --no_exceptions_unwind --execute_only --extended_initializers, --no_extended_initializer --feedback=filename --float_literal_pools, --no_float_literal_pools --force_new_nothrow, --no_force_new_nothrow --forceinline --fp16_format=format --fpmode=model --fpu=list --fpu=name --friend_injection, --no_friend_injection -g --global_reg=reg_name[,reg_name,...] --gnu --guiding_decls, --no_guiding_decls --help -Idir[,dir,...] --ignore_missing_headers --implicit_include, --no_implicit_include --implicit_include_searches, --no_implicit_include --implicit_key_function, --no_implicit_key_functio --implicit_typename, --no_implicit_typename --info=totals --inline, --no_inline --integer_literal_pools, --no_integer_literal_pool --interface_enums_are_32_bit --interleave -Jdir[,dir,...] --kandr_include -Lopt --library_interface=lib --library_type=lib --liclinger=seconds --link_all_input, --no_link_all_input --list --list_dir=directory_name --list_macros --littleend --locale=lang_country --long_long --loop_optimization_level=opt --loose_implicit_cast --lower_ropi, --no_lower_ropi --lower_rwpi, --no_lower_rwpi -M --md --message_locale=lang_country[.codepage] --min_array_alignment=opt --mm --multibyte_chars, --no_multibyte_chars --multifile, --no_multifile --multiply_latency=cycles --narrow_volatile_bitfields --nonstd_qualifier_deduction, --no_nonstd_qualifie -o filename -Onum --old_specializations, --no_old_specializations --old_style_preprocessing --omf_browse --ool_section_name, --no_ool_section_name -Ospace -Otime --output_dir=directory_name -P --parse_templates, --no_parse_templates --pch --pch_dir=dir --pch_messages, --no_pch_messages --pch_verbose, --no_pch_verbose --pending_instantiations=n --phony_targets --pointer_alignment=num --preinclude=filename --preprocess_assembly --preprocessed --protect_stack, --no_protect_stack --reassociate_saturation, --no_reassociate_saturat --reduce_paths, --no_reduce_paths --remarks --remove_unneeded_entities, --no_remove_unneeded_e --restrict, --no_restrict --retain=option --rtti, --no_rtti --rtti_data, --no_rtti_data -S --share_inlineable_strings, --no_share_inlineable_ --show_cmdline --signed_bitfields, --unsigned_bitfields --signed_chars, --unsigned_chars --split_ldm --split_sections --strict, --no_strict --strict_warnings --string_literal_pools, --no_string_literal_pools --sys_include --thumb --trigraphs, --no_trigraphs --type_traits_helpers, --no_type_traits_helpers -Uname --unaligned_access, --no_unaligned_access --use_frame_pointer, --no_use_frame_pointer --use_pch=filename --using_std, --no_using_std --version_number --vfe, --no_vfe --via=filename --vla, --no_vla --vsn -W --wchar, --no_wchar --wchar16 --wchar32 --whole_program --wrap_diagnostics, --no_wrap_diagnostics Language Extensions Compiler-specific Features C and C++ Implementation Details What is Semihosting? Via File Syntax Summary Table of GNU Language Extensions Standard C Implementation Definition Standard C++ Implementation Definition C and C++ Compiler Implementation Limits

--fp16_format=format

7.67 --fp16_format=format

Enables the use of half-precision floating-point numbers as an optional extension to the VFPv3 architecture. If a format is not specified, use of the __fp16 data type is faulted by the compiler.

Syntax

--fp16_format=format
Where format is one of:
alternative
An alternative to ieee that provides additional range, but has no NaN or infinity values.
ieee
Half-precision binary floating-point format defined by IEEE 754r, a revision to the IEEE 754 standard.
none
This is the default setting. It is equivalent to not specifying a format and means that the compiler faults use of the __fp16 data type.

Restrictions

The following restrictions apply when you use the __fp16 data type:
  • When used in a C or C++ expression, an __fp16 type is promoted to single precision. Subsequent promotion to double precision can occur if required by one of the operands.
  • A single precision value can be converted to __fp16. A double precision value is converted to single precision and then to __fp16, that could involve double rounding. This reflects the lack of direct double-to-16-bit conversion in the ARM architecture.
  • When using fpmode=fast, no floating-point exceptions are raised when converting to and from half-precision floating-point format.
  • Function formal arguments cannot be of type __fp16. However, pointers to variables of type __fp16 can be used as function formal argument types.
  • __fp16 values can be passed as actual function arguments. In this case, they are converted to single-precision values.
  • __fp16 cannot be specified as the return type of a function. However, a pointer to an __fp16 type can be used as a return type.
  • An __fp16 value is converted to a single-precision or double-precision value when used as a return value for a function that returns a float or double.
Related reference
Non-ConfidentialPDF file icon PDF versionARM DUI0375H
Copyright © 2007, 2008, 2011, 2012, 2014-2016 ARM. All rights reserved. 
  Arm logo
Important information

This site uses cookies to store information on your computer. By continuing to use our site, you consent to our cookies.

Change Settings

Privacy Policy Update

Arm’s Privacy Policy has been updated. By continuing to use our site, you consent to Arm’s Privacy Policy. Please review our Privacy Policy to learn more about our collection, use and transfers
of your data.