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

Commit 163e122

Browse files
ngzhianbinji
ngzhian
and
binji
authored
Specify text format for SIMD (#336)
* Specify text format for SIMD * Reword description of SIMD memory instructions Co-authored-by: Ben Smith <binjimin@gmail.com>
1 parent 3e652c3 commit 163e122

File tree

5 files changed

+305
-1
lines changed

5 files changed

+305
-1
lines changed

‎document/core/exec/numerics.rst‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -126,6 +126,7 @@ When a number is stored into :ref:`memory <syntax-mem>`, it is converted into a
126126

127127
.. math::
128128
\begin{array}{lll@{\qquad}l}
129+
\bytes_t(i^\ast) &=& (\bytes_t(i))^\ast \\[1ex]
129130
\bytes_t(i) &=& \littleendian(\bits_t(i)) \\[1ex]
130131
\littleendian(\epsilon) &=& \epsilon \\
131132
\littleendian(d^8~{d'}^\ast~) &=& \littleendian({d'}^\ast)~\ibits_8^{-1}(d^8) \\

‎document/core/syntax/instructions.rst‎

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -308,6 +308,8 @@ SIMD instructions provide basic operations over :ref:`values <syntax-value>` of
308308
\K{max} \\
309309
\end{array}
310310
311+
.. _syntax-simd-shape:
312+
311313
SIMD instructions have a naming convention involving a prefix that
312314
determines how their operands will be interpreted.
313315
This prefix describes the *shape* of the operand,

‎document/core/text/conventions.rst‎

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -67,7 +67,8 @@ In order to distinguish symbols of the textual syntax from symbols of the abstra
6767
\text{i32} &\Rightarrow& \I32 \\ &&|&
6868
\text{i64} &\Rightarrow& \I64 \\ &&|&
6969
\text{f32} &\Rightarrow& \F32 \\ &&|&
70-
\text{f64} &\Rightarrow& \F64 \\
70+
\text{f64} &\Rightarrow& \F64 \\ &&|&
71+
\text{v128} &\Rightarrow& \V128 \\
7172
\end{array}
7273
7374
The :ref:`textual grammar <text-limits>` for :ref:`limits <syntax-limits>` is defined as follows:

‎document/core/text/instructions.rst‎

Lines changed: 298 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -447,6 +447,304 @@ Numeric Instructions
447447
\end{array}
448448
449449
450+
.. index:: simd instruction
451+
pair: text format; instruction
452+
.. _text-instr-simd:
453+
454+
SIMD Instructions
455+
~~~~~~~~~~~~~~~~~~~~
456+
457+
SIMD memory instructions have optional offset and alignment immediates, like the :ref:`memory instructions <text-memarg>`.
458+
459+
.. math::
460+
\begin{array}{llclll}
461+
\production{instruction} & \Tplaininstr_I &::=& \dots \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\ &&|&
462+
\text{v128.load}~~m{:}\Tmemarg_{16} &\Rightarrow& \V128.\LOAD~m \\ &&|&
463+
\text{i16x8.load8x8\_s}~~m{:}\Tmemarg_8 &\Rightarrow& \I16X8.\LOAD\K{8x8\_s}~m \\ &&|&
464+
\text{i16x8.load8x8\_u}~~m{:}\Tmemarg_8 &\Rightarrow& \I16X8.\LOAD\K{8x8\_u}~m \\ &&|&
465+
\text{i32x4.load16x4\_s}~~m{:}\Tmemarg_8 &\Rightarrow& \I32X4.\LOAD\K{16x4\_s}~m \\ &&|&
466+
\text{i32x4.load16x4\_u}~~m{:}\Tmemarg_8 &\Rightarrow& \I32X4.\LOAD\K{16x4\_u}~m \\ &&|&
467+
\text{i64x2.load32x2\_s}~~m{:}\Tmemarg_8 &\Rightarrow& \I64X2.\LOAD\K{32x2\_s}~m \\ &&|&
468+
\text{i64x2.load32x2\_u}~~m{:}\Tmemarg_8 &\Rightarrow& \I64X2.\LOAD\K{32x2\_u}~m \\ &&|&
469+
\text{i8x16.load\_splat}~~m{:}\Tmemarg_1 &\Rightarrow& \I8X16.\LOAD\K{\_splat}~m \\ &&|&
470+
\text{i16x8.load\_splat}~~m{:}\Tmemarg_2 &\Rightarrow& \I16X8.\LOAD\K{\_splat}~m \\ &&|&
471+
\text{i32x4.load\_splat}~~m{:}\Tmemarg_4 &\Rightarrow& \I32X4.\LOAD\K{\_splat}~m \\ &&|&
472+
\text{i64x2.load\_splat}~~m{:}\Tmemarg_8 &\Rightarrow& \I64X2.\LOAD\K{\_splat}~m \\ &&|&
473+
\text{v128.store}~~m{:}\Tmemarg_{16} &\Rightarrow& \V128.\STORE~m \\
474+
\end{array}
475+
476+
SIMD const instructions have a mandatory :ref:`shape <syntax-simd-shape>` descriptor, which determines how the following values are parsed.
477+
478+
.. math::
479+
\begin{array}{llclll}
480+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
481+
\text{v128.const}~~\text{i8x16}~~(n{:}\Ti8)^{16} &\Rightarrow& \V128.\VCONST~\bytes_{i128}^{-1}(\bytes_{i8}(n^{16})) \\ &&|&
482+
\text{v128.const}~~\text{i16x8}~~(n{:}\Ti16)^{8} &\Rightarrow& \V128.\VCONST~\bytes_{i128}^{-1}(\bytes_{i16}(n^8)) \\ &&|&
483+
\text{v128.const}~~\text{i32x4}~~(n{:}\Ti32)^{4} &\Rightarrow& \V128.\VCONST~\bytes_{i128}^{-1}(\bytes_{i32}(n^4)) \\ &&|&
484+
\text{v128.const}~~\text{i64x2}~~(n{:}\Ti64)^{2} &\Rightarrow& \V128.\VCONST~\bytes_{i128}^{-1}(\bytes_{i64}(n^2)) \\ &&|&
485+
\text{v128.const}~~\text{f32x4}~~(z{:}\Tf32)^{4} &\Rightarrow& \V128.\VCONST~\bytes_{i128}^{-1}(\bytes_{f32}(z^4)) \\ &&|&
486+
\text{v128.const}~~\text{f64x2}~~(z{:}\Tf64)^{2} &\Rightarrow& \V128.\VCONST~\bytes_{i128}^{-1}(\bytes_{f64}(z^2)) \\ &&|&
487+
\text{i8x16.shuffle}~~(laneidx{:}\Tu8)^{16} &\Rightarrow& \I8X16.\SHUFFLE~laneidx^{16} \\ &&|&
488+
\text{i8x16.swizzle} &\Rightarrow& \I8X16.\SWIZZLE\\
489+
\end{array}
490+
491+
.. math::
492+
\begin{array}{llclll}
493+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
494+
\text{i8x16.splat} &\Rightarrow& \I8X16.\SPLAT\\ &&|&
495+
\text{i16x8.splat} &\Rightarrow& \I16X8.\SPLAT\\ &&|&
496+
\text{i32x4.splat} &\Rightarrow& \I32X4.\SPLAT\\ &&|&
497+
\text{i64x2.splat} &\Rightarrow& \I64X2.\SPLAT\\ &&|&
498+
\text{f32x4.splat} &\Rightarrow& \F32X4.\SPLAT\\ &&|&
499+
\text{f64x2.splat} &\Rightarrow& \F64X2.\SPLAT\\
500+
\end{array}
501+
502+
.. math::
503+
\begin{array}{llclll}
504+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
505+
\text{i8x16.extract\_lane\_s}~~laneidx{:}\Tu8 &\Rightarrow& \I8X16.\EXTRACTLANE\K{\_s}~laneidx \\ &&|&
506+
\text{i8x16.extract\_lane\_u}~~laneidx{:}\Tu8 &\Rightarrow& \I8X16.\EXTRACTLANE\K{\_u}~laneidx \\ &&|&
507+
\text{i8x16.replace\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \I8X16.\REPLACELANE~laneidx \\ &&|&
508+
\text{i16x8.extract\_lane\_s}~~laneidx{:}\Tu8 &\Rightarrow& \I16X8.\EXTRACTLANE\K{\_s}~laneidx \\ &&|&
509+
\text{i16x8.extract\_lane\_u}~~laneidx{:}\Tu8 &\Rightarrow& \I16X8.\EXTRACTLANE\K{\_u}~laneidx \\ &&|&
510+
\text{i16x8.replace\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \I16X8.\REPLACELANE~laneidx \\ &&|&
511+
\text{i32x4.extract\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \I32X4.\EXTRACTLANE~laneidx \\ &&|&
512+
\text{i32x4.replace\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \I32X4.\REPLACELANE~laneidx \\ &&|&
513+
\text{i64x2.extract\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \I64X2.\EXTRACTLANE~laneidx \\ &&|&
514+
\text{i64x2.replace\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \I64X2.\REPLACELANE~laneidx \\ &&|&
515+
\text{f32x4.extract\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \F32X4.\EXTRACTLANE~laneidx \\ &&|&
516+
\text{f32x4.replace\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \F32X4.\REPLACELANE~laneidx \\ &&|&
517+
\text{f64x2.extract\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \F64X2.\EXTRACTLANE~laneidx \\ &&|&
518+
\text{f64x2.replace\_lane}~~laneidx{:}\Tu8 &\Rightarrow& \F64X2.\REPLACELANE~laneidx \\
519+
\end{array}
520+
521+
.. _text-virelop:
522+
523+
.. math::
524+
\begin{array}{llclll}
525+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
526+
\text{i8x16.eq} &\Rightarrow& \I8X16.\VEQ\\ &&|&
527+
\text{i8x16.ne} &\Rightarrow& \I8X16.\VNE\\ &&|&
528+
\text{i8x16.lt\_s} &\Rightarrow& \I8X16.\VLT\K{\_s}\\ &&|&
529+
\text{i8x16.lt\_u} &\Rightarrow& \I8X16.\VLT\K{\_u}\\ &&|&
530+
\text{i8x16.gt\_s} &\Rightarrow& \I8X16.\VGT\K{\_s}\\ &&|&
531+
\text{i8x16.gt\_u} &\Rightarrow& \I8X16.\VGT\K{\_u}\\ &&|&
532+
\text{i8x16.le\_s} &\Rightarrow& \I8X16.\VLE\K{\_s}\\ &&|&
533+
\text{i8x16.le\_u} &\Rightarrow& \I8X16.\VLE\K{\_u}\\ &&|&
534+
\text{i8x16.ge\_s} &\Rightarrow& \I8X16.\VGE\K{\_s}\\ &&|&
535+
\text{i8x16.ge\_u} &\Rightarrow& \I8X16.\VGE\K{\_u}\\
536+
\end{array}
537+
538+
.. math::
539+
\begin{array}{llclll}
540+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
541+
\text{i16x8.eq} &\Rightarrow& \I16X8.\VEQ\\ &&|&
542+
\text{i16x8.ne} &\Rightarrow& \I16X8.\VNE\\ &&|&
543+
\text{i16x8.lt\_s} &\Rightarrow& \I16X8.\VLT\K{\_s}\\ &&|&
544+
\text{i16x8.lt\_u} &\Rightarrow& \I16X8.\VLT\K{\_u}\\ &&|&
545+
\text{i16x8.gt\_s} &\Rightarrow& \I16X8.\VGT\K{\_s}\\ &&|&
546+
\text{i16x8.gt\_u} &\Rightarrow& \I16X8.\VGT\K{\_u}\\ &&|&
547+
\text{i16x8.le\_s} &\Rightarrow& \I16X8.\VLE\K{\_s}\\ &&|&
548+
\text{i16x8.le\_u} &\Rightarrow& \I16X8.\VLE\K{\_u}\\ &&|&
549+
\text{i16x8.ge\_s} &\Rightarrow& \I16X8.\VGE\K{\_s}\\ &&|&
550+
\text{i16x8.ge\_u} &\Rightarrow& \I16X8.\VGE\K{\_u}\\
551+
\end{array}
552+
553+
.. math::
554+
\begin{array}{llclll}
555+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
556+
\text{i32x4.eq} &\Rightarrow& \I32X4.\VEQ\\ &&|&
557+
\text{i32x4.ne} &\Rightarrow& \I32X4.\VNE\\ &&|&
558+
\text{i32x4.lt\_s} &\Rightarrow& \I32X4.\VLT\K{\_s}\\ &&|&
559+
\text{i32x4.lt\_u} &\Rightarrow& \I32X4.\VLT\K{\_u}\\ &&|&
560+
\text{i32x4.gt\_s} &\Rightarrow& \I32X4.\VGT\K{\_s}\\ &&|&
561+
\text{i32x4.gt\_u} &\Rightarrow& \I32X4.\VGT\K{\_u}\\ &&|&
562+
\text{i32x4.le\_s} &\Rightarrow& \I32X4.\VLE\K{\_s}\\ &&|&
563+
\text{i32x4.le\_u} &\Rightarrow& \I32X4.\VLE\K{\_u}\\ &&|&
564+
\text{i32x4.ge\_s} &\Rightarrow& \I32X4.\VGE\K{\_s}\\ &&|&
565+
\text{i32x4.ge\_u} &\Rightarrow& \I32X4.\VGE\K{\_u}\\
566+
\end{array}
567+
568+
.. _text-vfrelop:
569+
570+
.. math::
571+
\begin{array}{llclll}
572+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
573+
\text{f32x4.eq} &\Rightarrow& \F32X4.\VEQ\\ &&|&
574+
\text{f32x4.ne} &\Rightarrow& \F32X4.\VNE\\ &&|&
575+
\text{f32x4.lt} &\Rightarrow& \F32X4.\VLT\\ &&|&
576+
\text{f32x4.gt} &\Rightarrow& \F32X4.\VGT\\ &&|&
577+
\text{f32x4.le} &\Rightarrow& \F32X4.\VLE\\ &&|&
578+
\text{f32x4.ge} &\Rightarrow& \F32X4.\VGE\\
579+
\end{array}
580+
581+
.. math::
582+
\begin{array}{llclll}
583+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
584+
\text{f64x2.eq} &\Rightarrow& \F64X2.\VEQ\\ &&|&
585+
\text{f64x2.ne} &\Rightarrow& \F64X2.\VNE\\ &&|&
586+
\text{f64x2.lt} &\Rightarrow& \F64X2.\VLT\\ &&|&
587+
\text{f64x2.gt} &\Rightarrow& \F64X2.\VGT\\ &&|&
588+
\text{f64x2.le} &\Rightarrow& \F64X2.\VLE\\ &&|&
589+
\text{f64x2.ge} &\Rightarrow& \F64X2.\VGE\\
590+
\end{array}
591+
592+
.. _text-vsunop:
593+
.. _text-vsbinop:
594+
.. _text-vsternop:
595+
596+
.. math::
597+
\begin{array}{llclll}
598+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
599+
\text{v128.not} &\Rightarrow& \V128.\VNOT\\ &&|&
600+
\text{v128.and} &\Rightarrow& \V128.\VAND\\ &&|&
601+
\text{v128.andnot} &\Rightarrow& \V128.\VANDNOT\\ &&|&
602+
\text{v128.or} &\Rightarrow& \V128.\VOR\\ &&|&
603+
\text{v128.xor} &\Rightarrow& \V128.\VXOR\\ &&|&
604+
\text{v128.bitselect} &\Rightarrow& \V128.\BITSELECT
605+
\end{array}
606+
607+
.. _text-vtestop:
608+
.. _text-vshiftop:
609+
.. _text-viunop:
610+
.. _text-vibinop:
611+
.. _text-viminmaxop:
612+
.. _text-vsatbinop:
613+
614+
.. math::
615+
\begin{array}{llclll}
616+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
617+
\text{i8x16.abs} &\Rightarrow& \I8X16.\VABS\\ &&|&
618+
\text{i8x16.neg} &\Rightarrow& \I8X16.\VNEG\\ &&|&
619+
\text{i8x16.any\_true} &\Rightarrow& \I8X16.\ANYTRUE\\ &&|&
620+
\text{i8x16.all\_true} &\Rightarrow& \I8X16.\ALLTRUE\\ &&|&
621+
\text{i8x16.bitmask} &\Rightarrow& \I8X16.\BITMASK\\ &&|&
622+
\text{i8x16.narrow\_i16x8\_s} &\Rightarrow& \I8X16.\NARROW\K{\_i16x8\_s}\\ &&|&
623+
\text{i8x16.narrow\_i16x8\_u} &\Rightarrow& \I8X16.\NARROW\K{\_i16x8\_u}\\ &&|&
624+
\text{i8x16.shl} &\Rightarrow& \I8X16.\VSHL\\ &&|&
625+
\text{i8x16.shr\_s} &\Rightarrow& \I8X16.\VSHR\K{\_s}\\ &&|&
626+
\text{i8x16.shr\_u} &\Rightarrow& \I8X16.\VSHR\K{\_u}\\ &&|&
627+
\text{i8x16.add} &\Rightarrow& \I8X16.\VADD\\ &&|&
628+
\text{i8x16.add\_sat\_s} &\Rightarrow& \I8X16.\VADD\K{\_sat\_s}\\ &&|&
629+
\text{i8x16.add\_sat\_u} &\Rightarrow& \I8X16.\VADD\K{\_sat\_u}\\ &&|&
630+
\text{i8x16.sub} &\Rightarrow& \I8X16.\VSUB\\ &&|&
631+
\text{i8x16.sub\_sat\_s} &\Rightarrow& \I8X16.\VSUB\K{\_sat\_s}\\ &&|&
632+
\text{i8x16.sub\_sat\_u} &\Rightarrow& \I8X16.\VSUB\K{\_sat\_u}\\ &&|&
633+
\text{i8x16.min\_s} &\Rightarrow& \I8X16.\VMIN\K{\_s}\\ &&|&
634+
\text{i8x16.min\_u} &\Rightarrow& \I8X16.\VMIN\K{\_u}\\ &&|&
635+
\text{i8x16.max\_s} &\Rightarrow& \I8X16.\VMAX\K{\_s}\\ &&|&
636+
\text{i8x16.max\_u} &\Rightarrow& \I8X16.\VMAX\K{\_u}\\ &&|&
637+
\text{i8x16.avgr\_u} &\Rightarrow& \I8X16.\AVGR\K{\_u}\\
638+
\end{array}
639+
640+
.. math::
641+
\begin{array}{llclll}
642+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
643+
\text{i16x8.abs} &\Rightarrow& \I16X8.\VABS\\ &&|&
644+
\text{i16x8.neg} &\Rightarrow& \I16X8.\VNEG\\ &&|&
645+
\text{i16x8.any\_true} &\Rightarrow& \I16X8.\ANYTRUE\\ &&|&
646+
\text{i16x8.all\_true} &\Rightarrow& \I16X8.\ALLTRUE\\ &&|&
647+
\text{i16x8.bitmask} &\Rightarrow& \I16X8.\BITMASK\\ &&|&
648+
\text{i16x8.narrow\_i32x4\_s} &\Rightarrow& \I16X8.\NARROW\K{\_i32x4\_s}\\ &&|&
649+
\text{i16x8.narrow\_i32x4\_u} &\Rightarrow& \I16X8.\NARROW\K{\_i32x4\_u}\\ &&|&
650+
\text{i16x8.widen\_low\_i8x16\_s} &\Rightarrow& \I16X8.\WIDEN\K{\_low\_i8x16\_s}\\ &&|&
651+
\text{i16x8.widen\_high\_i8x16\_s} &\Rightarrow& \I16X8.\WIDEN\K{\_high\_i8x16\_s}\\ &&|&
652+
\text{i16x8.widen\_low\_i8x16\_u} &\Rightarrow& \I16X8.\WIDEN\K{\_low\_i8x16\_u}\\ &&|&
653+
\text{i16x8.widen\_high\_i8x16\_u} &\Rightarrow& \I16X8.\WIDEN\K{\_high\_i8x16\_u}\\ &&|&
654+
\text{i16x8.shl} &\Rightarrow& \I16X8.\VSHL\\ &&|&
655+
\text{i16x8.shr\_s} &\Rightarrow& \I16X8.\VSHR\K{\_s}\\ &&|&
656+
\text{i16x8.shr\_u} &\Rightarrow& \I16X8.\VSHR\K{\_u}\\ &&|&
657+
\text{i16x8.add} &\Rightarrow& \I16X8.\VADD\\ &&|&
658+
\text{i16x8.add\_sat\_s} &\Rightarrow& \I16X8.\VADD\K{\_sat\_s}\\ &&|&
659+
\text{i16x8.add\_sat\_u} &\Rightarrow& \I16X8.\VADD\K{\_sat\_u}\\ &&|&
660+
\text{i16x8.sub} &\Rightarrow& \I16X8.\VSUB\\ &&|&
661+
\text{i16x8.sub\_sat\_s} &\Rightarrow& \I16X8.\VSUB\K{\_sat\_s}\\ &&|&
662+
\text{i16x8.sub\_sat\_u} &\Rightarrow& \I16X8.\VSUB\K{\_sat\_u}\\ &&|&
663+
\text{i16x8.mul} &\Rightarrow& \I16X8.\VMUL\\ &&|&
664+
\text{i16x8.min\_s} &\Rightarrow& \I16X8.\VMIN\K{\_s}\\ &&|&
665+
\text{i16x8.min\_u} &\Rightarrow& \I16X8.\VMIN\K{\_u}\\ &&|&
666+
\text{i16x8.max\_s} &\Rightarrow& \I16X8.\VMAX\K{\_s}\\ &&|&
667+
\text{i16x8.max\_u} &\Rightarrow& \I16X8.\VMAX\K{\_u}\\ &&|&
668+
\text{i16x8.avgr\_u} &\Rightarrow& \I16X8.\AVGR\K{\_u}\\
669+
\end{array}
670+
671+
.. math::
672+
\begin{array}{llclll}
673+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
674+
\text{i32x4.abs} &\Rightarrow& \I32X4.\VABS\\ &&|&
675+
\text{i32x4.neg} &\Rightarrow& \I32X4.\VNEG\\ &&|&
676+
\text{i32x4.any\_true} &\Rightarrow& \I32X4.\ANYTRUE\\ &&|&
677+
\text{i32x4.all\_true} &\Rightarrow& \I32X4.\ALLTRUE\\ &&|&
678+
\text{i32x4.bitmask} &\Rightarrow& \I32X4.\BITMASK\\ &&|&
679+
\text{i32x4.widen\_low\_i16x8\_s} &\Rightarrow& \I32X4.\WIDEN\K{\_low\_i16x8\_s}\\ &&|&
680+
\text{i32x4.widen\_high\_i16x8\_s} &\Rightarrow& \I32X4.\WIDEN\K{\_high\_i16x8\_s}\\ &&|&
681+
\text{i32x4.widen\_low\_i16x8\_u} &\Rightarrow& \I32X4.\WIDEN\K{\_low\_i16x8\_u}\\ &&|&
682+
\text{i32x4.widen\_high\_i16x8\_u} &\Rightarrow& \I32X4.\WIDEN\K{\_high\_i16x8\_u}\\ &&|&
683+
\text{i32x4.shl} &\Rightarrow& \I32X4.\VSHL\\ &&|&
684+
\text{i32x4.shr\_s} &\Rightarrow& \I32X4.\VSHR\K{\_s}\\ &&|&
685+
\text{i32x4.shr\_u} &\Rightarrow& \I32X4.\VSHR\K{\_u}\\ &&|&
686+
\text{i32x4.add} &\Rightarrow& \I32X4.\VADD\\ &&|&
687+
\text{i32x4.sub} &\Rightarrow& \I32X4.\VSUB\\ &&|&
688+
\text{i32x4.mul} &\Rightarrow& \I32X4.\VMUL\\ &&|&
689+
\text{i32x4.min\_s} &\Rightarrow& \I32X4.\VMIN\K{\_s}\\ &&|&
690+
\text{i32x4.min\_u} &\Rightarrow& \I32X4.\VMIN\K{\_u}\\ &&|&
691+
\text{i32x4.max\_s} &\Rightarrow& \I32X4.\VMAX\K{\_s}\\ &&|&
692+
\text{i32x4.max\_u} &\Rightarrow& \I32X4.\VMAX\K{\_u}\\
693+
\end{array}
694+
695+
.. math::
696+
\begin{array}{llclll}
697+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
698+
\text{i64x2.neg} &\Rightarrow& \I64X2.\VNEG\\ &&|&
699+
\text{i64x2.shl} &\Rightarrow& \I64X2.\VSHL\\ &&|&
700+
\text{i64x2.shr\_s} &\Rightarrow& \I64X2.\VSHR\K{\_s}\\ &&|&
701+
\text{i64x2.shr\_u} &\Rightarrow& \I64X2.\VSHR\K{\_u}\\ &&|&
702+
\text{i64x2.add} &\Rightarrow& \I64X2.\VADD\\ &&|&
703+
\text{i64x2.sub} &\Rightarrow& \I64X2.\VSUB\\ &&|&
704+
\text{i64x2.mul} &\Rightarrow& \I64X2.\VMUL\\
705+
\end{array}
706+
707+
.. _text-vfunop:
708+
.. _text-vfbinop:
709+
710+
.. math::
711+
\begin{array}{llclll}
712+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
713+
\text{f32x4.abs} &\Rightarrow& \F32X4.\VABS\\ &&|&
714+
\text{f32x4.neg} &\Rightarrow& \F32X4.\VNEG\\ &&|&
715+
\text{f32x4.sqrt} &\Rightarrow& \F32X4.\VSQRT\\ &&|&
716+
\text{f32x4.add} &\Rightarrow& \F32X4.\VADD\\ &&|&
717+
\text{f32x4.sub} &\Rightarrow& \F32X4.\VSUB\\ &&|&
718+
\text{f32x4.mul} &\Rightarrow& \F32X4.\VMUL\\ &&|&
719+
\text{f32x4.div} &\Rightarrow& \F32X4.\VDIV\\ &&|&
720+
\text{f32x4.min} &\Rightarrow& \F32X4.\VMIN\\ &&|&
721+
\text{f32x4.max} &\Rightarrow& \F32X4.\VMAX\\
722+
\end{array}
723+
724+
.. math::
725+
\begin{array}{llclll}
726+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
727+
\text{f64x2.abs} &\Rightarrow& \F64X2.\VABS\\ &&|&
728+
\text{f64x2.neg} &\Rightarrow& \F64X2.\VNEG\\ &&|&
729+
\text{f64x2.sqrt} &\Rightarrow& \F64X2.\VSQRT\\ &&|&
730+
\text{f64x2.add} &\Rightarrow& \F64X2.\VADD\\ &&|&
731+
\text{f64x2.sub} &\Rightarrow& \F64X2.\VSUB\\ &&|&
732+
\text{f64x2.mul} &\Rightarrow& \F64X2.\VMUL\\ &&|&
733+
\text{f64x2.div} &\Rightarrow& \F64X2.\VDIV\\ &&|&
734+
\text{f64x2.min} &\Rightarrow& \F64X2.\VMIN\\ &&|&
735+
\text{f64x2.max} &\Rightarrow& \F64X2.\VMAX\\
736+
\end{array}
737+
738+
.. math::
739+
\begin{array}{llclll}
740+
\phantom{\production{instruction}} & \phantom{\Tplaininstr_I} &\phantom{::=}& \phantom{averylonginstructionnameforsimdtext} && \phantom{simdhasreallylonginstructionnames} \\[-2ex] &&|&
741+
\text{i32x4.trunc\_sat\_f32x4\_s} &\Rightarrow& \I32X4.\TRUNC\K{\_sat\_f32x4\_s}\\ &&|&
742+
\text{i32x4.trunc\_sat\_f32x4\_u} &\Rightarrow& \I32X4.\TRUNC\K{\_sat\_f32x4\_u}\\ &&|&
743+
\text{f32x4.convert\_i32x4\_s} &\Rightarrow& \F32X4.\CONVERT\K{\_i32x4\_s}\\ &&|&
744+
\text{f32x4.convert\_i32x4\_u} &\Rightarrow& \F32X4.\CONVERT\K{\_i32x4\_u}\\
745+
\end{array}
746+
747+
450748
.. index:: ! folded instruction, S-expression
451749
.. _text-foldedinstr:
452750

‎document/core/util/macros.def‎

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -652,6 +652,8 @@
652652
.. |Ts64| mathdef:: \xref{text/values}{text-int}{\TsX{\T{64}}}
653653

654654
.. |TiN| mathdef:: \xref{text/values}{text-int}{\TiX{N}}
655+
.. |Ti8| mathdef:: \xref{text/values}{text-int}{\TiX{\T{8}}}
656+
.. |Ti16| mathdef:: \xref{text/values}{text-int}{\TiX{\T{16}}}
655657
.. |Ti32| mathdef:: \xref{text/values}{text-int}{\TiX{\T{32}}}
656658
.. |Ti64| mathdef:: \xref{text/values}{text-int}{\TiX{\T{64}}}
657659

0 commit comments

Comments
 (0)