Change History¶
Since the original release 1.0 of the WebAssembly specification, a number of proposals for extensions have been integrated. The following sections provide an overview of what has changed.
Release XX¶
Threads and Atomics¶
Added a new concept of shared memory, denoting a memory which can be accessed by multiple host-created threads, and atomic instructions for synchronising shared memory concurrency 1.
New memory type: \(\href{../syntax/types.html#syntax-memtype}{\mathsf{shared}}\)
New memory instructions: \(\href{../syntax/instructions.html#syntax-instr-atomic-memory}{\mathsf{atomic.load}}\), \(\href{../syntax/instructions.html#syntax-instr-atomic-memory}{\mathsf{atomic.store}}\), \(\href{../syntax/instructions.html#syntax-instr-atomic-memory}{\mathsf{atomic.rmw}}\), \(\href{../syntax/instructions.html#syntax-instr-atomic-memory}{\mathsf{memory.atomic.notify}}\), \(\href{../syntax/instructions.html#syntax-instr-atomic-memory}{\mathsf{memory.atomic.wait}}\), \(\href{../syntax/instructions.html#syntax-instr-atomic-memory}{\mathsf{atomic.fence}}\)
Note
The proposal does not add a mechanism for creating threads from within WebAssembly. Thread creation must be handled by the host.
Release 2.0¶
Sign extension instructions¶
Added new numeric instructions for performing sign extension within integer representations 2.
New numeric instructions: \(\mathsf{i}\mathit{nn}\mathsf{.}\href{../syntax/instructions.html#syntax-instr-numeric}{\mathsf{extend}}\mathit{N}\mathsf{\_s}\)
Non-trapping float-to-int conversions¶
Added new conversion instructions that avoid trapping when converting a floating-point number to an integer 3.
New numeric instructions: \(\mathsf{i}\mathit{nn}\mathsf{.}\href{../syntax/instructions.html#syntax-instr-numeric}{\mathsf{trunc}}\mathsf{\_sat\_f}\mathit{mm}\mathsf{\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\)
Multiple values¶
Generalized the result type of blocks and functions to allow for multiple values; in addition, introduced the ability to have block parameters 4.
Function types allow more than one result
Block types can be arbitrary function types
Reference types¶
Added \(\href{../syntax/types.html#syntax-reftype}{\mathsf{funcref}}\) and \(\href{../syntax/types.html#syntax-reftype}{\mathsf{externref}}\) as new value types and respective instructions 5.
New value types: reference types \(\href{../syntax/types.html#syntax-reftype}{\mathsf{funcref}}\) and \(\href{../syntax/types.html#syntax-reftype}{\mathsf{externref}}\)
New reference instructions: \(\href{../syntax/instructions.html#syntax-instr-ref}{\mathsf{ref{.}null}}\), \(\href{../syntax/instructions.html#syntax-instr-ref}{\mathsf{ref{.}func}}\), \(\href{../syntax/instructions.html#syntax-instr-ref}{\mathsf{ref{.}is\_null}}\)
Enrich parametric instruction: \(\href{../syntax/instructions.html#syntax-instr-parametric}{\mathsf{select}}\) with optional type immediate
New declarative form of element segment
Table instructions¶
Added instructions to directly access and modify tables 5.
Table types allow any reference type as element type
New table instructions: \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.get}}\), \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.set}}\), \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.size}}\), \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.grow}}\)
Multiple tables¶
Added the ability to use multiple tables per module 5.
Table instructions take a table index immediate: \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.get}}\), \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.set}}\), \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.size}}\), \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.grow}}\), \(\href{../syntax/instructions.html#syntax-instr-control}{\mathsf{call\_indirect}}\)
Element segments take a table index
Bulk memory and table instructions¶
Added instructions that modify ranges of memory or table entries 5 6
New memory instructions: \(\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{memory.fill}}\), \(\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{memory.init}}\), \(\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{memory.copy}}\), \(\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{data.drop}}\)
New table instructions: \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.fill}}\), \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.init}}\), \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{table.copy}}\), \(\href{../syntax/instructions.html#syntax-instr-table}{\mathsf{elem.drop}}\)
New passive form of data segment
New passive form of element segment
New data count section in binary format
Active data and element segments boundaries are no longer checked at compile time but may trap instead
Vector instructions¶
Added vector type and instructions that manipulate multiple numeric values in parallel (also known as SIMD, single instruction multiple data) 7
New value type: \(\href{../syntax/types.html#syntax-valtype}{\mathsf{v128}}\)
New memory instructions: \(\mathsf{v128.}\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{load}}\), \(\mathsf{v128.}\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{load}}{}\!N\!\mathsf{x}\!M\!\mathsf{\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{v128.}\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{load}}{}N\mathsf{\_zero}\), \(\mathsf{v128.}\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{load}}{}N\mathsf{\_splat}\), \(\mathsf{v128.}\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{load}}{}N\mathsf{\_lane}\), \(\mathsf{v128.}\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{store}}\), \(\mathsf{v128.}\href{../syntax/instructions.html#syntax-instr-memory}{\mathsf{store}}{}N\mathsf{\_lane}\)
New constant vector instruction: \(\mathsf{v128.}\href{../syntax/instructions.html#syntax-instr-vec}{\mathsf{const}}\)
New unary vector instructions: \(\mathsf{v128.not}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.abs}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.neg}\), \(\mathsf{i8x16.popcnt}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.abs}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.neg}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.sqrt}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.ceil}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.floor}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.trunc}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.nearest}\)
New binary vector instructions: \(\mathsf{v128.and}\), \(\mathsf{v128.andnot}\), \(\mathsf{v128.or}\), \(\mathsf{v128.xor}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.add}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.sub}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.mul}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.add\_sat\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.sub\_sat\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.min\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.max\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.shl}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.shr\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.add}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.extmul\_}\href{../syntax/instructions.html#syntax-half}{\mathit{half}}\mathsf{\_i}\!N'\!\mathsf{x}\!M'\!\mathsf{\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i16x8.q15mulr\_sat\_s}\), \(\mathsf{i32x4.dot\_i16x8\_s}\), \(\mathsf{i16x8.extadd\_pairwise\_i8x16\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i32x4.extadd\_pairwise\_i16x8\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i8x16.avgr\_u}\), \(\mathsf{i16x8.avgr\_u}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.sub}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.mul}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.div}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.min}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.max}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.pmin}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.pmax}\)
New ternary vector instruction: \(\mathsf{v128.bitselect}\)
New test vector instructions: \(\mathsf{v128.any\_true}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.all\_true}\)
New relational vector instructions: \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.eq}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.ne}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.lt\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.gt\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.le\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.ge\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.eq}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.ne}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.lt}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.gt}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.le}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.ge}\)
New conversion vector instructions:\(\mathsf{i32x4.trunc\_sat\_f32x4\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i32x4.trunc\_sat\_f64x2\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\mathsf{\_zero}\), \(\mathsf{f32x4.convert\_i32x4\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{f32x4.demote\_f64x2\_zero}\), \(\mathsf{f64x2.convert\_low\_i32x4\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{f64x2.promote\_low\_f32x4}\)
New lane access vector instructions: \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.extract\_lane\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}^?\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.replace\_lane}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.extract\_lane}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.replace\_lane}\)
New lane splitting/combining vector instructions: \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.extend\_}\href{../syntax/instructions.html#syntax-half}{\mathit{half}}\mathsf{\_i}\!N'\!\mathsf{x}\!M'\!\mathsf{\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i8x16.narrow\_i16x8\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\), \(\mathsf{i16x8.narrow\_i32x4\_}\href{../syntax/instructions.html#syntax-sx}{\mathit{sx}}\)
New byte reordering vector instructions: \(\mathsf{i8x16.shuffle}\), \(\mathsf{i8x16.swizzle}\)
New injection/projection vector instructions: \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.splat}\), \(\mathsf{f}\!N\!\mathsf{x}\!M\!\mathsf{.splat}\), \(\mathsf{i}\!N\!\mathsf{x}\!M\!\mathsf{.bitmask}\)
- 1
https://github.com/WebAssembly/spec/tree/main/proposals/threads/
- 2
https://github.com/WebAssembly/spec/tree/main/proposals/sign-extension-ops/
- 3
https://github.com/WebAssembly/spec/tree/main/proposals/nontrapping-float-to-int-conversion/
- 4
https://github.com/WebAssembly/spec/tree/main/proposals/multi-value/
- 5(1,2,3,4)
https://github.com/WebAssembly/spec/tree/main/proposals/reference-types/
- 6
https://github.com/WebAssembly/spec/tree/main/proposals/bulk-memory-operations/
- 7
https://github.com/WebAssembly/spec/tree/main/proposals/simd/