Fix RISC-V Packed SIMD intrinsics compilation
This commit is contained in:
Asuna
Amanieu d'Antras
parent
7bb6e1978a
commit
2f608b514b
1 changed files with 136 additions and 19 deletions
|
|
@ -2,9 +2,22 @@
|
|||
//!
|
||||
//! RV64 only part is placed in riscv64 folder.
|
||||
use crate::arch::asm;
|
||||
#[cfg(test)]
|
||||
use stdarch_test::assert_instr;
|
||||
|
||||
// FIXME: Currently the P extension is still unratified, so there is no support
|
||||
// for it in the upstream LLVM for now, and thus no LLVM built-in functions or
|
||||
// serialization of instructions are provided.
|
||||
//
|
||||
// We add `assert_instr(unknown)` to each function so that we can at least make
|
||||
// sure they compile. Since there is no serialization yet, we can only write
|
||||
// "unknown" here, so that if LLVM upstream provides support for the P extension
|
||||
// at some point in the future, we can know in time and then update our
|
||||
// implementation.
|
||||
|
||||
/// Adds packed 16-bit signed numbers, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn add16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -16,6 +29,7 @@ pub fn add16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Halves the sum of packed 16-bit signed numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn radd16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -27,6 +41,7 @@ pub fn radd16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Halves the sum of packed 16-bit unsigned numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn uradd16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -38,6 +53,7 @@ pub fn uradd16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Adds packed 16-bit signed numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kadd16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -49,6 +65,7 @@ pub fn kadd16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Adds packed 16-bit unsigned numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ukadd16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -60,6 +77,7 @@ pub fn ukadd16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Subtracts packed 16-bit signed numbers, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sub16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -71,6 +89,7 @@ pub fn sub16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Halves the subtraction result of packed 16-bit signed numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn rsub16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -82,6 +101,7 @@ pub fn rsub16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Halves the subtraction result of packed 16-bit unsigned numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ursub16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -93,6 +113,7 @@ pub fn ursub16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Subtracts packed 16-bit signed numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ksub16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -104,6 +125,7 @@ pub fn ksub16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Subtracts packed 16-bit unsigned numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn uksub16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -115,6 +137,7 @@ pub fn uksub16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross adds and subtracts packed 16-bit signed numbers, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn cras16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -126,6 +149,7 @@ pub fn cras16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross halves of adds and subtracts packed 16-bit signed numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn rcras16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -137,6 +161,7 @@ pub fn rcras16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross halves of adds and subtracts packed 16-bit unsigned numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn urcras16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -148,6 +173,7 @@ pub fn urcras16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross adds and subtracts packed 16-bit signed numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kcras16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -159,6 +185,7 @@ pub fn kcras16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross adds and subtracts packed 16-bit unsigned numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ukcras16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -170,6 +197,7 @@ pub fn ukcras16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross subtracts and adds packed 16-bit signed numbers, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn crsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -181,6 +209,7 @@ pub fn crsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross halves of subtracts and adds packed 16-bit signed numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn rcrsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -192,6 +221,7 @@ pub fn rcrsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross halves of subtracts and adds packed 16-bit unsigned numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn urcrsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -203,6 +233,7 @@ pub fn urcrsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross subtracts and adds packed 16-bit signed numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kcrsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -214,6 +245,7 @@ pub fn kcrsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Cross subtracts and adds packed 16-bit unsigned numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ukcrsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -225,6 +257,7 @@ pub fn ukcrsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight adds and subtracts packed 16-bit signed numbers, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn stas16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -236,6 +269,7 @@ pub fn stas16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight halves of adds and subtracts packed 16-bit signed numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn rstas16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -247,6 +281,7 @@ pub fn rstas16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight halves of adds and subtracts packed 16-bit unsigned numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn urstas16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -258,6 +293,7 @@ pub fn urstas16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight adds and subtracts packed 16-bit signed numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kstas16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -269,6 +305,7 @@ pub fn kstas16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight adds and subtracts packed 16-bit unsigned numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ukstas16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -280,6 +317,7 @@ pub fn ukstas16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight subtracts and adds packed 16-bit signed numbers, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn stsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -291,6 +329,7 @@ pub fn stsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight halves of subtracts and adds packed 16-bit signed numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn rstsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -302,6 +341,7 @@ pub fn rstsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight halves of subtracts and adds packed 16-bit unsigned numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn urstsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -313,6 +353,7 @@ pub fn urstsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight subtracts and adds packed 16-bit signed numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kstsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -324,6 +365,7 @@ pub fn kstsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Straight subtracts and adds packed 16-bit unsigned numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ukstsa16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -335,6 +377,7 @@ pub fn ukstsa16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Adds packed 8-bit signed numbers, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn add8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -346,6 +389,7 @@ pub fn add8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Halves the sum of packed 8-bit signed numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn radd8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -357,6 +401,7 @@ pub fn radd8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Halves the sum of packed 8-bit unsigned numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn uradd8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -368,6 +413,7 @@ pub fn uradd8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Adds packed 8-bit signed numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kadd8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -379,6 +425,7 @@ pub fn kadd8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Adds packed 8-bit unsigned numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ukadd8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -390,6 +437,7 @@ pub fn ukadd8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Subtracts packed 8-bit signed numbers, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sub8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -401,6 +449,7 @@ pub fn sub8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Halves the subtraction result of packed 8-bit signed numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn rsub8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -412,6 +461,7 @@ pub fn rsub8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Halves the subtraction result of packed 8-bit unsigned numbers, dropping least bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ursub8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -423,6 +473,7 @@ pub fn ursub8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Subtracts packed 8-bit signed numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ksub8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -434,6 +485,7 @@ pub fn ksub8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Subtracts packed 8-bit unsigned numbers, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn uksub8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -445,6 +497,7 @@ pub fn uksub8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Arithmetic right shift packed 16-bit elements without rounding up
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sra16(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -456,6 +509,7 @@ pub fn sra16(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Arithmetic right shift packed 16-bit elements with rounding up
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sra16u(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -467,6 +521,7 @@ pub fn sra16u(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical right shift packed 16-bit elements without rounding up
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn srl16(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -478,6 +533,7 @@ pub fn srl16(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical right shift packed 16-bit elements with rounding up
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn srl16u(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -489,6 +545,7 @@ pub fn srl16u(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical left shift packed 16-bit elements, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sll16(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -500,6 +557,7 @@ pub fn sll16(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical left shift packed 16-bit elements, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ksll16(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -511,6 +569,7 @@ pub fn ksll16(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical saturating left then arithmetic right shift packed 16-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kslra16(a: usize, b: i32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -522,6 +581,7 @@ pub fn kslra16(a: usize, b: i32) -> usize {
|
|||
|
||||
/// Logical saturating left then arithmetic right shift packed 16-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kslra16u(a: usize, b: i32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -533,6 +593,7 @@ pub fn kslra16u(a: usize, b: i32) -> usize {
|
|||
|
||||
/// Arithmetic right shift packed 8-bit elements without rounding up
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sra8(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -544,6 +605,7 @@ pub fn sra8(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Arithmetic right shift packed 8-bit elements with rounding up
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sra8u(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -555,6 +617,7 @@ pub fn sra8u(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical right shift packed 8-bit elements without rounding up
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn srl8(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -566,6 +629,7 @@ pub fn srl8(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical right shift packed 8-bit elements with rounding up
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn srl8u(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -577,6 +641,7 @@ pub fn srl8u(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical left shift packed 8-bit elements, discarding overflow bits
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sll8(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -588,6 +653,7 @@ pub fn sll8(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical left shift packed 8-bit elements, saturating at the numeric bounds
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ksll8(a: usize, b: u32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -599,6 +665,7 @@ pub fn ksll8(a: usize, b: u32) -> usize {
|
|||
|
||||
/// Logical saturating left then arithmetic right shift packed 8-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kslra8(a: usize, b: i32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -610,6 +677,7 @@ pub fn kslra8(a: usize, b: i32) -> usize {
|
|||
|
||||
/// Logical saturating left then arithmetic right shift packed 8-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kslra8u(a: usize, b: i32) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -621,6 +689,7 @@ pub fn kslra8u(a: usize, b: i32) -> usize {
|
|||
|
||||
/// Compare equality for packed 16-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn cmpeq16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -632,6 +701,7 @@ pub fn cmpeq16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compare whether 16-bit packed signed integers are less than the others
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn scmplt16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -643,6 +713,7 @@ pub fn scmplt16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compare whether 16-bit packed signed integers are less than or equal to the others
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn scmple16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -654,6 +725,7 @@ pub fn scmple16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compare whether 16-bit packed unsigned integers are less than the others
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ucmplt16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -665,6 +737,7 @@ pub fn ucmplt16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compare whether 16-bit packed unsigned integers are less than or equal to the others
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ucmple16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -676,6 +749,7 @@ pub fn ucmple16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compare equality for packed 8-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn cmpeq8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -687,6 +761,7 @@ pub fn cmpeq8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compare whether 8-bit packed signed integers are less than the others
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn scmplt8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -698,6 +773,7 @@ pub fn scmplt8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compare whether 8-bit packed signed integers are less than or equal to the others
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn scmple8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -709,6 +785,7 @@ pub fn scmple8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compare whether 8-bit packed unsigned integers are less than the others
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ucmplt8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -720,6 +797,7 @@ pub fn ucmplt8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compare whether 8-bit packed unsigned integers are less than or equal to the others
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ucmple8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -731,6 +809,7 @@ pub fn ucmple8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Get minimum values from 16-bit packed signed integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn smin16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -742,6 +821,7 @@ pub fn smin16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Get minimum values from 16-bit packed unsigned integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn umin16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -753,6 +833,7 @@ pub fn umin16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Get maximum values from 16-bit packed signed integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn smax16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -764,6 +845,7 @@ pub fn smax16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Get maximum values from 16-bit packed unsigned integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn umax16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -777,39 +859,43 @@ pub fn umax16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compute the absolute value of packed 16-bit signed integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kabs16(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAD1", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAD1)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Count the number of redundant sign bits of the packed 16-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn clrs16(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAE8", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAE8)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Count the number of leading zero bits of the packed 16-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn clz16(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAE9", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAE9)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Swap the 16-bit halfwords within each 32-bit word of a register
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn swap16(a: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -822,6 +908,7 @@ pub fn swap16(a: usize) -> usize {
|
|||
|
||||
/// Get minimum values from 8-bit packed signed integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn smin8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -833,6 +920,7 @@ pub fn smin8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Get minimum values from 8-bit packed unsigned integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn umin8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -844,6 +932,7 @@ pub fn umin8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Get maximum values from 8-bit packed signed integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn smax8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -855,6 +944,7 @@ pub fn smax8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Get maximum values from 8-bit packed unsigned integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn umax8(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -868,154 +958,168 @@ pub fn umax8(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Compute the absolute value of packed 8-bit signed integers
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kabs8(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAD0", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAD0)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Count the number of redundant sign bits of the packed 8-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn clrs8(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAE0", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAE0)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Count the number of leading zero bits of the packed 8-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn clz8(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAE1", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAE1)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Swap the 8-bit bytes within each 16-bit halfword of a register.
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn swap8(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAD8", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAD8)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack first and zeroth into two 16-bit signed halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sunpkd810(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAC8", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAC8)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack second and zeroth into two 16-bit signed halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sunpkd820(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAC9", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAC9)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack third and zeroth into two 16-bit signed halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sunpkd830(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xACA", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xACA)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack third and first into two 16-bit signed halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sunpkd831(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xACB", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xACB)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack third and second into two 16-bit signed halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn sunpkd832(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAD3", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAD3)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack first and zeroth into two 16-bit unsigned halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn zunpkd810(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xACC", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xACC)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack second and zeroth into two 16-bit unsigned halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn zunpkd820(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xACD", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xACD)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack third and zeroth into two 16-bit unsigned halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn zunpkd830(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xACE", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xACE)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack third and first into two 16-bit unsigned halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn zunpkd831(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xACF", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xACF)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Unpack third and second into two 16-bit unsigned halfwords in each 32-bit chunk
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn zunpkd832(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAD7", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAD7)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
|
@ -1024,6 +1128,7 @@ pub fn zunpkd832(a: usize) -> usize {
|
|||
|
||||
/// Pack two 16-bit data from bottom and top half from 32-bit chunks
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn pkbt16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -1035,6 +1140,7 @@ pub fn pkbt16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Pack two 16-bit data from top and bottom half from 32-bit chunks
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn pktb16(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -1046,28 +1152,31 @@ pub fn pktb16(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Count the number of redundant sign bits of the packed 32-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn clrs32(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAF8", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAF8)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Count the number of leading zero bits of the packed 32-bit elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn clz32(a: usize) -> usize {
|
||||
let value: usize;
|
||||
unsafe {
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, 0xAF9", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
asm!(".insn i 0x77, 0x0, {}, {}, %lo(0xAF9)", lateout(reg) value, in(reg) a, options(pure, nomem, nostack))
|
||||
}
|
||||
value
|
||||
}
|
||||
|
||||
/// Calculate the sum of absolute difference of unsigned 8-bit data elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn pbsad(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -1079,6 +1188,7 @@ pub fn pbsad(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Calculate and accumulate the sum of absolute difference of unsigned 8-bit data elements
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn pbsada(t: usize, a: usize, b: usize) -> usize {
|
||||
let mut value: usize;
|
||||
|
|
@ -1090,6 +1200,7 @@ pub fn pbsada(t: usize, a: usize, b: usize) -> usize {
|
|||
|
||||
/// Multiply signed 8-bit elements and add 16-bit elements on results for packed 32-bit chunks
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn smaqa(t: usize, a: usize, b: usize) -> usize {
|
||||
let mut value: usize;
|
||||
|
|
@ -1101,6 +1212,7 @@ pub fn smaqa(t: usize, a: usize, b: usize) -> usize {
|
|||
|
||||
/// Multiply unsigned 8-bit elements and add 16-bit elements on results for packed 32-bit chunks
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn umaqa(t: usize, a: usize, b: usize) -> usize {
|
||||
let mut value: usize;
|
||||
|
|
@ -1112,6 +1224,7 @@ pub fn umaqa(t: usize, a: usize, b: usize) -> usize {
|
|||
|
||||
/// Multiply signed to unsigned 8-bit and add 16-bit elements on results for packed 32-bit chunks
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn smaqasu(t: usize, a: usize, b: usize) -> usize {
|
||||
let mut value: usize;
|
||||
|
|
@ -1123,6 +1236,7 @@ pub fn smaqasu(t: usize, a: usize, b: usize) -> usize {
|
|||
|
||||
/// Adds signed lower 16-bit content of two registers with Q15 saturation
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn kaddh(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -1134,6 +1248,7 @@ pub fn kaddh(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Subtracts signed lower 16-bit content of two registers with Q15 saturation
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ksubh(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -1145,6 +1260,7 @@ pub fn ksubh(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Adds signed lower 16-bit content of two registers with U16 saturation
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn ukaddh(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
@ -1156,6 +1272,7 @@ pub fn ukaddh(a: usize, b: usize) -> usize {
|
|||
|
||||
/// Subtracts signed lower 16-bit content of two registers with U16 saturation
|
||||
#[inline]
|
||||
#[cfg_attr(test, assert_instr(unknown))]
|
||||
#[unstable(feature = "riscv_ext_intrinsics", issue = "114544")]
|
||||
pub fn uksubh(a: usize, b: usize) -> usize {
|
||||
let value: usize;
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue