Skip to content
This repository was archived by the owner on Dec 22, 2021. It is now read-only.

Add Changelog #522

Merged
dtig merged 4 commits into from
Oct 4, 2021
Merged

Add Changelog #522

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
40e7c08
Add Changelog
rossberg Sep 28, 2021
507c462
Comments
rossberg Sep 28, 2021
f8d4e73
Split narrow
rossberg Sep 29, 2021
d665f49
Regroup instructions
rossberg Sep 29, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
47 changes: 42 additions & 5 deletions document/core/appendix/changes.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -104,17 +104,54 @@ Added instructions that modify ranges of memory or table entries [#proposal-reft
* Active data and element segments boundaries are no longer checked at compile time but may trap instead


.. index:: instructions, SIMD, value type, vector type

Vector instructions
...................

Added vector type and instructions that manipulate multiple numeric values in parallel (also known as *SIMD*, single instruction multiple data) [#proposal-vectype]_
Copy link
Member

@dtig dtig Sep 29, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Would it make sense to add text here that's representative of the lane operations that extract and replace single lanes instead of multiple values?

Copy link
Member Author

@rossberg rossberg Sep 29, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hm, there is no such breakdown in the intro sentence of the other changelog sections either. The details are in the bullet list, this is just for the general direction.

Copy link
Member

@dtig dtig Oct 4, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ack.


* New :ref:`value type <syntax-valtype>`: |V128|

* New :ref:`memory instructions <syntax-instr-memory>`: :math:`\K{v128.}\LOAD`, :math:`\K{v128.}\LOAD{}\!N\!\K{x}\!M\!\K{\_}\sx`, :math:`\K{v128.}\LOAD{}N\K{\_zero}`, :math:`\K{v128.}\LOAD{}N\K{\_splat}`, :math:`\K{v128.}\LOAD{}N\K{\_lane}`, :math:`\K{v128.}\STORE`, :math:`\K{v128.}\STORE{}N\K{\_lane}`

* New constant :ref:`vector instruction <syntax-instr-vec>`: :math:`\K{v128.}\VCONST`

* New unary :ref:`vector instructions <syntax-instr-vec>`: :math:`\K{v128.not}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.abs}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.neg}`, :math:`\K{i8x16.popcnt}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.abs}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.neg}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.sqrt}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.ceil}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.floor}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.trunc}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.nearest}`

* New binary :ref:`vector instructions <syntax-instr-vec>`: :math:`\K{v128.and}`, :math:`\K{v128.andnot}`, :math:`\K{v128.or}`, :math:`\K{v128.xor}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.add}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.sub}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.mul}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.add\_sat\_}\sx`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.sub\_sat\_}\sx`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.min\_}\sx`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.max\_}\sx`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.shl}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.shr\_}\sx`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.add}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.extmul\_}\half\K{\_i}\!N'\!\K{x}\!M'\!\K{\_}\sx`, :math:`\K{i16x8.q15mulr\_sat\_s}`, :math:`\K{i32x4.dot\_i16x8\_s}`, :math:`\K{i16x8.extadd\_pairwise\_i8x16\_}\sx`, :math:`\K{i32x4.extadd\_pairwise\_i16x8\_}\sx`, :math:`\K{i8x16.avgr\_u}`, :math:`\K{i16x8.avgr\_u}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.sub}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.mul}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.div}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.min}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.max}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.pmin}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.pmax}`

* New ternary :ref:`vector instruction <syntax-instr-vec>`: :math:`\K{v128.bitselect}`

* New test :ref:`vector instructions <syntax-instr-vec>`: :math:`\K{v128.any\_true}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.all\_true}`

* New relational :ref:`vector instructions <syntax-instr-vec>`: :math:`\K{i}\!N\!\K{x}\!M\!\K{.eq}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.ne}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.lt\_}\sx`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.gt\_}\sx`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.le\_}\sx`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.ge\_}\sx`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.eq}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.ne}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.lt}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.gt}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.le}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.ge}`

* New conversion :ref:`vector instructions <syntax-instr-vec>`::math:`\K{i32x4.trunc\_sat\_f32x4\_}\sx`, :math:`\K{i32x4.trunc\_sat\_f64x2\_}\sx\K{\_zero}`, :math:`\K{f32x4.convert\_i32x4\_}\sx`, :math:`\K{f32x4.demote\_f64x2\_zero}`, :math:`\K{f64x2.convert\_low\_i32x4\_}\sx`, :math:`\K{f64x2.promote\_low\_f32x4}`

* New lane access :ref:`vector instructions <syntax-instr-vec>`: :math:`\K{i}\!N\!\K{x}\!M\!\K{.extract\_lane\_}\sx^?`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.replace\_lane}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.extract\_lane}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.replace\_lane}`

* New lane splitting/combining :ref:`vector instructions <syntax-instr-vec>`: :math:`\K{i}\!N\!\K{x}\!M\!\K{.extend\_}\half\K{\_i}\!N'\!\K{x}\!M'\!\K{\_}\sx`, :math:`\K{i8x16.narrow\_i16x8\_}\sx`, :math:`\K{i16x8.narrow\_i32x4\_}\sx`

* New byte reordering :ref:`vector instructions <syntax-instr-vec>`: :math:`\K{i8x16.shuffle}`, :math:`\K{i8x16.swizzle}`

* New injection/projection :ref:`vector instructions <syntax-instr-vec>`: :math:`\K{i}\!N\!\K{x}\!M\!\K{.splat}`, :math:`\K{f}\!N\!\K{x}\!M\!\K{.splat}`, :math:`\K{i}\!N\!\K{x}\!M\!\K{.bitmask}`


.. [#proposal-signext]
https://github.com/WebAssembly/spec/tree/master/proposals/sign-extension-ops/
https://github.com/WebAssembly/spec/tree/main/proposals/sign-extension-ops/

.. [#proposal-cvtsat]
https://github.com/WebAssembly/spec/tree/master/proposals/nontrapping-float-to-int-conversion/
https://github.com/WebAssembly/spec/tree/main/proposals/nontrapping-float-to-int-conversion/

.. [#proposal-multivalue]
https://github.com/WebAssembly/spec/tree/master/proposals/multi-value/
https://github.com/WebAssembly/spec/tree/main/proposals/multi-value/

.. [#proposal-reftype]
https://github.com/WebAssembly/spec/tree/master/proposals/reference-types/
https://github.com/WebAssembly/spec/tree/main/proposals/reference-types/

.. [#proposal-bulk]
https://github.com/WebAssembly/spec/tree/master/proposals/bulk-memory-operations/
https://github.com/WebAssembly/spec/tree/main/proposals/bulk-memory-operations/

.. [#proposal-vectype]
https://github.com/WebAssembly/spec/tree/main/proposals/simd/
4 changes: 2 additions & 2 deletions document/core/appendix/gen-index-instructions.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -569,8 +569,8 @@ def Instruction(name, opcode, type=None, validation=None, execution=None, operat
Instruction(r'\F64X2.\VPMAX', r'\hex{FD}~~\hex{F7}', r'[\V128~\V128] \to [\V128]', r'valid-vbinop', r'exec-vbinop', r'op-fpmax'),
Instruction(r'\I32X4.\TRUNC\K{\_sat\_f32x4\_s}', r'\hex{FD}~~\hex{F8}', r'[\V128] \to [\V128]', r'valid-vcvtop', r'exec-vcvtop', r'op-trunc_sat_s'),
Instruction(r'\I32X4.\TRUNC\K{\_sat\_f32x4\_u}', r'\hex{FD}~~\hex{F9}', r'[\V128] \to [\V128]', r'valid-vcvtop', r'exec-vcvtop', r'op-trunc_sat_u'),
Instruction(r'\F32X4.\CONVERT\K{\_i32x4\_s}', r'\hex{FD}~~\hex{FA}', r'[\V128] \to [\V128]', r'valid-vcvtop', r'exec-vcvtop', r'op-convert_s'),
Instruction(r'\F32X4.\CONVERT\K{\_i32x4\_u}', r'\hex{FD}~~\hex{FB}', r'[\V128] \to [\V128]', r'valid-vcvtop', r'exec-vcvtop', r'op-convert_u'),
Instruction(r'\F32X4.\VCONVERT\K{\_i32x4\_s}', r'\hex{FD}~~\hex{FA}', r'[\V128] \to [\V128]', r'valid-vcvtop', r'exec-vcvtop', r'op-convert_s'),
Instruction(r'\F32X4.\VCONVERT\K{\_i32x4\_u}', r'\hex{FD}~~\hex{FB}', r'[\V128] \to [\V128]', r'valid-vcvtop', r'exec-vcvtop', r'op-convert_u'),
Instruction(r'\I32X4.\VTRUNC\K{\_sat\_f64x2\_s\_zero}', r'\hex{FD}~~\hex{FC}', r'[\V128] \to [\V128]', r'valid-vcvtop', r'exec-vcvtop', r'op-trunc_sat_s'),
Instruction(r'\I32X4.\VTRUNC\K{\_sat\_f64x2\_u\_zero}', r'\hex{FD}~~\hex{FD}', r'[\V128] \to [\V128]', r'valid-vcvtop', r'exec-vcvtop', r'op-trunc_sat_u'),
Instruction(r'\F64X2.\VCONVERT\K{\_low\_i32x4\_s}', r'\hex{FD}~~\hex{FE}', r'[\V128] \to [\V128]', r'valid-vcvtop', r'exec-vcvtop', r'op-convert_s'),
Expand Down
4 changes: 2 additions & 2 deletions document/core/appendix/index-instructions.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -517,8 +517,8 @@ Instruction Binary Opcode T
:math:`\F64X2.\VPMAX` :math:`\hex{FD}~~\hex{F7}` :math:`[\V128~\V128] \to [\V128]` :ref:`validation <valid-vbinop>` :ref:`execution <exec-vbinop>`, :ref:`operator <op-fpmax>`
:math:`\I32X4.\TRUNC\K{\_sat\_f32x4\_s}` :math:`\hex{FD}~~\hex{F8}` :math:`[\V128] \to [\V128]` :ref:`validation <valid-vcvtop>` :ref:`execution <exec-vcvtop>`, :ref:`operator <op-trunc_sat_s>`
:math:`\I32X4.\TRUNC\K{\_sat\_f32x4\_u}` :math:`\hex{FD}~~\hex{F9}` :math:`[\V128] \to [\V128]` :ref:`validation <valid-vcvtop>` :ref:`execution <exec-vcvtop>`, :ref:`operator <op-trunc_sat_u>`
:math:`\F32X4.\CONVERT\K{\_i32x4\_s}` :math:`\hex{FD}~~\hex{FA}` :math:`[\V128] \to [\V128]` :ref:`validation <valid-vcvtop>` :ref:`execution <exec-vcvtop>`, :ref:`operator <op-convert_s>`
:math:`\F32X4.\CONVERT\K{\_i32x4\_u}` :math:`\hex{FD}~~\hex{FB}` :math:`[\V128] \to [\V128]` :ref:`validation <valid-vcvtop>` :ref:`execution <exec-vcvtop>`, :ref:`operator <op-convert_u>`
:math:`\F32X4.\VCONVERT\K{\_i32x4\_s}` :math:`\hex{FD}~~\hex{FA}` :math:`[\V128] \to [\V128]` :ref:`validation <valid-vcvtop>` :ref:`execution <exec-vcvtop>`, :ref:`operator <op-convert_s>`
:math:`\F32X4.\VCONVERT\K{\_i32x4\_u}` :math:`\hex{FD}~~\hex{FB}` :math:`[\V128] \to [\V128]` :ref:`validation <valid-vcvtop>` :ref:`execution <exec-vcvtop>`, :ref:`operator <op-convert_u>`
:math:`\I32X4.\VTRUNC\K{\_sat\_f64x2\_s\_zero}` :math:`\hex{FD}~~\hex{FC}` :math:`[\V128] \to [\V128]` :ref:`validation <valid-vcvtop>` :ref:`execution <exec-vcvtop>`, :ref:`operator <op-trunc_sat_s>`
:math:`\I32X4.\VTRUNC\K{\_sat\_f64x2\_u\_zero}` :math:`\hex{FD}~~\hex{FD}` :math:`[\V128] \to [\V128]` :ref:`validation <valid-vcvtop>` :ref:`execution <exec-vcvtop>`, :ref:`operator <op-trunc_sat_u>`
:math:`\F64X2.\VCONVERT\K{\_low\_i32x4\_s}` :math:`\hex{FD}~~\hex{FE}` :math:`[\V128] \to [\V128]` :ref:`validation <valid-vcvtop>` :ref:`execution <exec-vcvtop>`, :ref:`operator <op-convert_s>`
Expand Down
19 changes: 9 additions & 10 deletions document/core/syntax/instructions.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -268,7 +268,7 @@ Vector instructions (also known as *SIMD* instructions, single data multiple val
\K{i32x4.}\VTRUNC\K{\_sat\_f64x2\_}\sx\K{\_zero} \\&&|&
\K{f32x4.}\VCONVERT\K{\_i32x4\_}\sx ~|~
\K{f32x4.}\VDEMOTE\K{\_f64x2\_zero} \\&&|&
\K{f64x2.}\VCONVERT\K{\_low\_i32x4\_}sx ~|~
\K{f64x2.}\VCONVERT\K{\_low\_i32x4\_}\sx ~|~
\K{f64x2.}\VPROMOTE\K{\_low\_f32x4} \\&&|&
\dots \\
\production{vector bitwise unary operator} & \vvunop &::=&
Expand Down Expand Up @@ -312,8 +312,7 @@ Vector instructions (also known as *SIMD* instructions, single data multiple val
\K{sub\_sat\_}\sx \\
\production{vector integer shift operator} & \vishiftop &::=&
\K{shl} ~|~
\K{shr\_s} ~|~
\K{shr\_u} \\
\K{shr\_}\sx \\
\production{vector floating-point unary operator} & \vfunop &::=&
\K{abs} ~|~
\K{neg} ~|~
Expand Down Expand Up @@ -542,7 +541,7 @@ Instructions in this group are concerned with linear :ref:`memory <syntax-mem>`.
\begin{array}{llcl}
\production{memory immediate} & \memarg &::=&
\{ \OFFSET~\u32, \ALIGN~\u32 \} \\
\production{lane width} & \lanewidth &::=&
\production{lane width} & \X{ww} &::=&
8 ~|~ 16 ~|~ 32 ~|~ 64 \\
\production{instruction} & \instr &::=&
\dots \\&&|&
Expand All @@ -558,14 +557,14 @@ Instructions in this group are concerned with linear :ref:`memory <syntax-mem>`.
\K{i}\X{nn}\K{.}\STORE\K{8}~\memarg ~|~
\K{i}\X{nn}\K{.}\STORE\K{16}~\memarg ~|~
\K{i64.}\STORE\K{32}~\memarg \\&&|&
\K{v128.}\LOAD\K{8x8}\_\sx~\memarg ~|~
\K{v128.}\LOAD\K{16x4}\_\sx~\memarg ~|~
\K{v128.}\LOAD\K{32x2}\_\sx~\memarg \\&&|&
\K{v128.}\LOAD\lanewidth\K{\_splat}~\memarg \\&&|&
\K{v128.}\LOAD\K{8x8\_}\sx~\memarg ~|~
\K{v128.}\LOAD\K{16x4\_}\sx~\memarg ~|~
\K{v128.}\LOAD\K{32x2\_}\sx~\memarg \\&&|&
\K{v128.}\LOAD\K{32\_zero}~\memarg ~|~
\K{v128.}\LOAD\K{64\_zero}~\memarg \\&&|&
\K{v128.}\LOAD\lanewidth\K{\_lane}~\memarg~\laneidx ~|~
\K{v128.}\STORE\lanewidth\K{\_lane}~\memarg~\laneidx \\&&|&
\K{v128.}\LOAD\X{ww}\K{\_splat}~\memarg \\&&|&
\K{v128.}\LOAD\X{ww}\K{\_lane}~\memarg~\laneidx ~|~
\K{v128.}\STORE\X{ww}\K{\_lane}~\memarg~\laneidx \\&&|&
\MEMORYSIZE \\&&|&
\MEMORYGROW \\&&|&
\MEMORYFILL \\&&|&
Expand Down
1 change: 0 additions & 1 deletion document/core/util/macros.def
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -535,7 +535,6 @@
.. |sx| mathdef:: \xref{syntax/instructions}{syntax-sx}{\X{sx}}
.. |half| mathdef:: \xref{syntax/instructions}{syntax-half}{\X{half}}
.. |memarg| mathdef:: \xref{syntax/instructions}{syntax-memarg}{\X{memarg}}
.. |lanewidth| mathdef:: \xref{syntax/instructions}{syntax-lanewidth}{\X{lanewidth}}

.. |blocktype| mathdef:: \xref{syntax/instructions}{syntax-blocktype}{\X{blocktype}}

Expand Down