WebAssembly SIMD conversion instructions - WebAssembly

Convert between types

convert_low_i32x4_s: Converts the lanes of a v128 f64x2 value interpretation into a signed i32x4 value interpretation.
convert_low_i32x4_u: Converts the lanes of a v128 f64x2 value interpretation into an unsigned i32x4 value interpretation.
convert_i32x4_s: Converts the lanes of a v128 f32x4 value interpretation into a signed i32x4 value interpretation.
convert_i32x4_u: Converts the lanes of a v128 f32x4 value interpretation into an unsigned i32x4 value interpretation.
demote_f64x2_zero: Converts the lanes of a v128 f64x2 value interpretation into an f32x4 value interpretation. The two higher lanes of the result are initialized to zero.
promote_low_f32x4: Converts the first two lanes of a v128 f32x4 value interpretation into an f64x2 value interpretation.

Convert from narrower to wider type

extend_high_i8x16_s: Converts lanes 8–15 of a signed v128 i8x16 value interpretation into an i16x8 value interpretation.
extend_high_i8x16_u: Converts lanes 8–15 of an unsigned v128 i8x16 value interpretation into an i16x8 value interpretation.
extend_high_i16x8_s: Converts lanes 4–7 of a signed v128 i16x8 value interpretation into an i32x4 value interpretation.
extend_high_i16x8_u: Converts lanes 4–7 of an unsigned v128 i16x8 value interpretation into an i32x4 value interpretation.
extend_high_i32x4_s: Converts lanes 2–3 of a signed v128 i32x4 value interpretation into an i64x2 value interpretation.
extend_high_i32x4_u: Converts lanes 2–3 of an unsigned v128 i32x4 value interpretation into an i64x2 value interpretation.
extend_low_i8x16_s: Converts lanes 0–7 of a signed v128 i8x16 value interpretation into an i16x8 value interpretation.
extend_low_i8x16_u: Converts lanes 0–7 of an unsigned v128 i8x16 value interpretation into an i16x8 value interpretation.
extend_low_i16x8_s: Converts lanes 0–3 of a signed v128 i16x8 value interpretation into an i32x4 value interpretation.
extend_low_i16x8_u: Converts lanes 0–3 of an unsigned v128 i16x8 value interpretation into an i32x4 value interpretation.
extend_low_i32x4_s: Converts lanes 0–1 of a signed v128 i32x4 value interpretation into an i64x2 value interpretation.
extend_low_i32x4_u: Converts lanes 0–1 of an unsigned v128 i32x4 value interpretation into an i64x2 value interpretation.

Convert from wider to narrower type

narrow_i16x8_s: Converts two signed v128 i16x8 value interpretations into an i8x16 value interpretation.
narrow_i16x8_u: Converts two signed v128 i16x8 value interpretations into an i8x16 value interpretation using unsigned saturation.
narrow_i32x4_s: Converts two signed v128 i32x4 value interpretations into an i16x8 value interpretation.
narrow_i32x4_u: Converts two signed v128 i32x4 value interpretations into an i16x8 value interpretation using unsigned saturation.

Value replacement

replace_lane: Replaces the specified lane of a v128 value interpretation with a new value and returns the resulting v128 value interpretation.
shuffle: Returns a new v128 value interpretation with its lane values selected from two input v128 values, determined by provided index values.
splat: Copies the same value to all lanes of a v128 value interpretation.
swizzle: Returns a new v128 value interpretation with its lane values selected from an input v128 value, determined by indices provided in a second input v128.

SIMD-specific trunc instructions

trunc_sat_f32x4_s: Performs a saturating conversion of the lanes of a v128 f32x4 value interpretation into a signed i32x4 value interpretation, clamping the output to the range allowed by the value type.
trunc_sat_f32x4_u: Performs a saturating conversion of the lanes of a v128 f32x4 value interpretation into an unsigned i32x4 value interpretation, clamping the output to the range allowed by the value type.
trunc_sat_f64x2_s_zero: Performs a saturating conversion of the lanes of a v128 f64x2 value interpretation into a signed i32x4 value interpretation, clamping the output to the range allowed by the value type. The two higher lanes of the result are initialized to zero.
trunc_sat_f64x2_u_zero: Performs a saturating conversion of the lanes of a v128 f64x2 value interpretation into an unsigned i32x4 value interpretation, clamping the output to the range allowed by the value type. The two higher lanes of the result are initialized to zero.

Note: See also the non-SIMD-specific trunc instruction.

Help improve MDN

Learn how to contribute

This page was last modified on Jan 1, 1970 by MDN contributors.

View this page on GitHub • Report a problem with this content