This is an experimental technology, part of the Harmony (ECMAScript 7) proposal.
Because this technology's specification has not stabilized, check the compatibility table for usage in various browsers. Also note that the syntax and behavior of an experimental technology is subject to change in future version of browsers as the spec changes.
The SIMD.float32x4
data type is a 128-bit vector divided into 4 lanes storing single precision floating point values.
Figure 1: SIMD.float32x4 in a 128-bit SIMD register.
Syntax
SIMD.float32x4(x, y, z, w);
Parameters
The input values are specified in double precision floating and are converted to single precision floating point values before being stored.
x
Optional- An double specifying the value of the first lane. Defaults to
NaN
. y
Optional- An double specifying the value of the second lane. Defaults to
NaN
. z
Optional- An double specifying the value of the third lane. Defaults to
NaN
. w
Optional- An double specifying the value of the fourth lane. Defaults to
NaN
.
Constructors
In addition to the simple constructors, the SIMD API provides the following constructor. Note that you can also convert from another SIMD data type to float32x4.
SIMD.float32x4.splat()
- Creates an float32x4 with all lanes set to a given value.
Operations
To actually do something with SIMD types, SIMD operations are needed that work on SIMD data types.
Checking SIMD types
SIMD.float32x4.check()
- Returns a new float32x4 if the parameter is a valid float32x4 data type. Throws a
TypeError
otherwise.
Accessing and mutating lanes
SIMD.float32x4.extractLane()
- Returns the value of the given lane.
SIMD.float32x4.replaceLane()
- Returns a new float32x4 with the given lane value replaced.
Loading from and storing into typed arrays
SIMD.float32x4.load()
- Returns a new float32x4 with the lane values loaded from a typed array.
SIMD.float32x4.store()
- Stores an float32x4 into a typed array.
Arithmetic operations
SIMD.float32x4.abs()
- Returns a new float32x4 with the absolute lane values.
SIMD.float32x4.add()
- Returns a new float32x4 with the lane values added (
a + b
). SIMD.float32x4.div()
- Returns a new float32x4 with the lane values divided (
a / b
). SIMD.float32x4.mul()
- Returns a new float32x4 with the lane values multiplied (
a * b
). SIMD.float32x4.neg()
- Returns a new float32x4 with the negated lane values.
SIMD.float32x4.reciprocalApproximation()
- Returns a new float32x4 with an approximation of the reciprocal lane values.
SIMD.float32x4.reciprocalSqrtApproximation()
- Returns a new float32x4 with an approximation of the reciprocal square root lane values.
SIMD.float32x4.sub()
- Returns a new float32x4 with the lane values subtracted (
a - b
). SIMD.float32x4.sqrt()
- Returns a new float32x4 with the square root of the lane values.
Shuffling and swizzling
SIMD.float32x4.shuffle()
- Returns a new float32x4 with the lane values shuffled.
SIMD.float32x4.swizzle()
- Returns a new float32x4 with the lane values swizzled.
Min/max and clamping
SIMD.float32x4.clamp()
- Returns a new float32x4 with the lane values clamped between a lower limit and an upper limit.
SIMD.float32x4.max()
- Returns a new float32x4 with the maximum of the lane values.
SIMD.float32x4.maxNum()
- Returns a new float32x4 with the maximum of the lane values, preferring numbers over
NaN
. SIMD.float32x4.min()
- Returns a new float32x4 with the minimum of the lane values.
SIMD.float32x4.minNum()
- Returns a new float32x4 with the minimum of the lane values, preferring numbers over
NaN
.
Selections
SIMD.float32x4.select()
- Returns a new float32x4 with the lane values being a mix of the lanes depending on the selector mask.
SIMD.float32x4.selectBits()
- Returns a new float32x4 with the lane values being a mix of bits depending on the selector mask.
Comparisons
SIMD.float32x4.equal()
- Returns a selection mask depending on
a == b
. SIMD.float32x4.notEqual()
- Returns a selection mask depending on
a != b
. SIMD.float32x4.lessThan()
- Returns a selection mask depending on
a < b
. SIMD.float32x4.lessThanOrEqual()
- Returns a selection mask depending on
a <= b
. SIMD.float32x4.greaterThan()
- Returns a selection mask depending on
a > b
. SIMD.float32x4.greaterThanOrEqual()
- Returns a selection mask depending on
a >= b
.
Bitwise logical operations
SIMD.float32x4.and()
- Returns a new float32x4 with the logical AND of the lane values (
a & b
). SIMD.float32x4.or()
- Returns a new float32x4 with the logical OR of the lane values (
a | b
). SIMD.float32x4.xor()
- Returns a new float32x4 with the logical XOR of the lane values (
a ^ b
). SIMD.float32x4.not()
- Returns a new float32x4 with lane with the logical NOT of the lane values (
~a
).
Data conversions (packing / unpacking)
SIMD.float32x4.fromInt32x4()
- Creates a new float32x4 data type with a float conversion from an int32x4.
SIMD.float32x4.fromInt32x4Bits()
- Creates a new float32x4 data type with a bit-wise copy from an int32x4.
SIMD.float32x4.fromFloat64x2()
- Creates a new float32x4 data type with a float conversion from a float64x2.
SIMD.float32x4.fromFloat64x2Bits()
- Creates a new float32x4 data type with a bit-wise copy from a float64x2.
SIMD.float32x4.fromInt16x8Bits()
- Creates a new float32x4 data type with a bit-wise copy from an int16x8.
SIMD.float32x4.fromInt8x16Bits()
- Creates a new float32x4 data type with a bit-wise copy from an int8x16.
Examples
Constructing an float32x4
new SIMD.float32x4(1, 2, 3, 4); // float32x4[1, 2, 3, 4] new SIMD.float32x4(1, 2); // float32x4[1, 2, NaN, NaN] new SIMD.float32x4(); // float32x4[NaN, NaN, NaN, NaN]
Specifications
SIMD is not yet part of an official standards document or draft. For standardization work and a Polyfill implementation based on typed arrays, see the ecmascript_simd GitHub repository.
Browser compatibility
Feature | Chrome | Firefox (Gecko) | Internet Explorer | Opera | Safari |
---|---|---|---|---|---|
Basic support | Not supported | Nightly build | Not supported | Not supported | Not supported |
Feature | Android | Chrome for Android | Firefox Mobile (Gecko) | IE Mobile | Opera Mobile | Safari Mobile |
---|---|---|---|---|---|---|
Basic support | Not supported | Not supported | Nightly build | Not supported | Not supported | Not supported |