This is an archive of the discontinued LLVM Phabricator instance.

Differential D97759

[doc] Fix description of _Float16
ClosedPublic

Authored by kito-cheng on Mar 2 2021, 1:48 AM.

Details

Reviewers
SjoerdMeijer
craig.topper
simon_tatham
stuij
Commits
rGb46a1b129f68: [doc] Fix description of _Float16
Summary

According to ISO/IEC TS 18661-3:2015[1] _FloatN is interchange floating point type, extended floating-point type is _FloatNx (e.g. _Float32x).

[1] http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2342.pdf

Diff Detail

Event Timeline

kito-cheng requested review of this revision.Mar 2 2021, 1:48 AM
kito-cheng created this revision.
Herald added a project: Restricted Project. · View Herald TranscriptMar 2 2021, 1:48 AM
Herald added a subscriber: llvm-commits. · View Herald Transcript
SjoerdMeijer accepted this revision.Mar 2 2021, 2:28 AM

Yep, correct.

This revision is now accepted and ready to land.Mar 2 2021, 2:28 AM
Harbormaster completed remote builds in B91526: Diff 327387.Mar 2 2021, 3:45 AM
This revision was landed with ongoing or failed builds.Mar 3 2021, 10:19 PM
Closed by commit rGb46a1b129f68: [doc] Fix description of _Float16 (authored by kito-cheng). · Explain Why
This revision was automatically updated to reflect the committed changes.
Kito Cheng <kito.cheng@sifive.com> added a commit: rGb46a1b129f68: [doc] Fix description of _Float16.
Herald added a subscriber: cfe-commits. · View Herald TranscriptMar 3 2021, 10:19 PM

Revision Contents

PathSize
clang/
docs/
2 lines

Diff 328023

clang/docs/LanguageExtensions.rst

Show First 20 Lines • Show All 536 Lines • ▼ Show 20 Lines
``__fp16`` is a storage and interchange format only. This means that values of ``__fp16`` is a storage and interchange format only. This means that values of
``__fp16`` are immediately promoted to (at least) ``float`` when used in arithmetic ``__fp16`` are immediately promoted to (at least) ``float`` when used in arithmetic
operations, so that e.g. the result of adding two ``__fp16`` values has type ``float``. operations, so that e.g. the result of adding two ``__fp16`` values has type ``float``.
The behavior of ``__fp16`` is specified by the ARM C Language Extensions (`ACLE <http://infocenter.arm.com/help/topic/com.arm.doc.ihi0053d/IHI0053D_acle_2_1.pdf>`_). The behavior of ``__fp16`` is specified by the ARM C Language Extensions (`ACLE <http://infocenter.arm.com/help/topic/com.arm.doc.ihi0053d/IHI0053D_acle_2_1.pdf>`_).
Clang uses the ``binary16`` format from IEEE 754-2008 for ``__fp16``, not the ARM Clang uses the ``binary16`` format from IEEE 754-2008 for ``__fp16``, not the ARM
alternative format. alternative format.
``_Float16`` is an extended floating-point type. This means that, just like arithmetic on ``_Float16`` is an interchange floating-point type. This means that, just like arithmetic on
``float`` or ``double``, arithmetic on ``_Float16`` operands is formally performed in the ``float`` or ``double``, arithmetic on ``_Float16`` operands is formally performed in the
``_Float16`` type, so that e.g. the result of adding two ``_Float16`` values has type ``_Float16`` type, so that e.g. the result of adding two ``_Float16`` values has type
``_Float16``. The behavior of ``_Float16`` is specified by ISO/IEC TS 18661-3:2015 ``_Float16``. The behavior of ``_Float16`` is specified by ISO/IEC TS 18661-3:2015
("Floating-point extensions for C"). As with ``__fp16``, Clang uses the ``binary16`` ("Floating-point extensions for C"). As with ``__fp16``, Clang uses the ``binary16``
format from IEEE 754-2008 for ``_Float16``. format from IEEE 754-2008 for ``_Float16``.
``_Float16`` arithmetic will be performed using native half-precision support ``_Float16`` arithmetic will be performed using native half-precision support
when available on the target (e.g. on ARMv8.2a); otherwise it will be performed when available on the target (e.g. on ARMv8.2a); otherwise it will be performed
▲ Show 20 Lines • Show All 3,237 Lines • Show Last 20 Lines