simd_adler32/imp/
wasm.rs

use super::Adler32Imp;

/// Resolves update implementation if CPU supports simd128 instructions.
pub fn get_imp() -> Option<Adler32Imp> {
  get_imp_inner()
}

#[inline]
#[cfg(target_feature = "simd128")]
fn get_imp_inner() -> Option<Adler32Imp> {
  Some(imp::update)
}

#[inline]
#[cfg(not(target_feature = "simd128"))]
fn get_imp_inner() -> Option<Adler32Imp> {
  None
}

#[cfg(target_feature = "simd128")]
mod imp {
  const MOD: u32 = 65521;
  const NMAX: usize = 5552;
  const BLOCK_SIZE: usize = 32;
  const CHUNK_SIZE: usize = NMAX / BLOCK_SIZE * BLOCK_SIZE;

  #[cfg(target_arch = "wasm32")]
  use core::arch::wasm32::*;
  #[cfg(target_arch = "wasm64")]
  use core::arch::wasm64::*;

  pub fn update(a: u16, b: u16, data: &[u8]) -> (u16, u16) {
    update_imp(a, b, data)
  }

  #[inline]
  #[target_feature(enable = "simd128")]
  fn update_imp(a: u16, b: u16, data: &[u8]) -> (u16, u16) {
    let mut a = a as u32;
    let mut b = b as u32;

    let chunks = data.chunks_exact(CHUNK_SIZE);
    let remainder = chunks.remainder();
    for chunk in chunks {
      update_chunk_block(&mut a, &mut b, chunk);
    }

    update_block(&mut a, &mut b, remainder);

    (a as u16, b as u16)
  }

  fn update_chunk_block(a: &mut u32, b: &mut u32, chunk: &[u8]) {
    debug_assert_eq!(
      chunk.len(),
      CHUNK_SIZE,
      "Unexpected chunk size (expected {}, got {})",
      CHUNK_SIZE,
      chunk.len()
    );

    reduce_add_blocks(a, b, chunk);

    *a %= MOD;
    *b %= MOD;
  }

  fn update_block(a: &mut u32, b: &mut u32, chunk: &[u8]) {
    debug_assert!(
      chunk.len() <= CHUNK_SIZE,
      "Unexpected chunk size (expected <= {}, got {})",
      CHUNK_SIZE,
      chunk.len()
    );

    for byte in reduce_add_blocks(a, b, chunk) {
      *a += *byte as u32;
      *b += *a;
    }

    *a %= MOD;
    *b %= MOD;
  }

  #[inline(always)]
  fn reduce_add_blocks<'a>(a: &mut u32, b: &mut u32, chunk: &'a [u8]) -> &'a [u8] {
    if chunk.len() < BLOCK_SIZE {
      return chunk;
    }

    let blocks = chunk.chunks_exact(BLOCK_SIZE);
    let blocks_remainder = blocks.remainder();

    let weight_hi_v = get_weight_hi();
    let weight_lo_v = get_weight_lo();

    let mut p_v = u32x4(*a * blocks.len() as u32, 0, 0, 0);
    let mut a_v = u32x4(0, 0, 0, 0);
    let mut b_v = u32x4(*b, 0, 0, 0);

    for block in blocks {
      let block_ptr = block.as_ptr() as *const v128;
      let v_lo = unsafe { block_ptr.read_unaligned() };
      let v_hi = unsafe { block_ptr.add(1).read_unaligned() };

      p_v = u32x4_add(p_v, a_v);

      a_v = u32x4_add(a_v, u32x4_extadd_quarters_u8x16(v_lo));
      let mad = i32x4_dot_i8x16(v_lo, weight_lo_v);
      b_v = u32x4_add(b_v, mad);

      a_v = u32x4_add(a_v, u32x4_extadd_quarters_u8x16(v_hi));
      let mad = i32x4_dot_i8x16(v_hi, weight_hi_v);
      b_v = u32x4_add(b_v, mad);
    }

    b_v = u32x4_add(b_v, u32x4_shl(p_v, 5));

    *a += reduce_add(a_v);
    *b = reduce_add(b_v);

    blocks_remainder
  }

  #[inline(always)]
  fn i32x4_dot_i8x16(a: v128, b: v128) -> v128 {
    let a_lo = u16x8_extend_low_u8x16(a);
    let a_hi = u16x8_extend_high_u8x16(a);

    let b_lo = u16x8_extend_low_u8x16(b);
    let b_hi = u16x8_extend_high_u8x16(b);

    let lo = i32x4_dot_i16x8(a_lo, b_lo);
    let hi = i32x4_dot_i16x8(a_hi, b_hi);

    i32x4_add(lo, hi)
  }

  #[inline(always)]
  fn u32x4_extadd_quarters_u8x16(a: v128) -> v128 {
    u32x4_extadd_pairwise_u16x8(u16x8_extadd_pairwise_u8x16(a))
  }

  #[inline(always)]
  fn reduce_add(v: v128) -> u32 {
    let arr: [u32; 4] = unsafe { std::mem::transmute(v) };
    let mut sum = 0u32;
    for val in arr {
      sum = sum.wrapping_add(val);
    }
    sum
  }

  #[inline(always)]
  fn get_weight_lo() -> v128 {
    u8x16(
      32, 31, 30, 29, 28, 27, 26, 25, 24, 23, 22, 21, 20, 19, 18, 17,
    )
  }

  #[inline(always)]
  fn get_weight_hi() -> v128 {
    u8x16(16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1)
  }
}

#[cfg(test)]
mod tests {
  use rand::Rng;

  #[test]
  fn zeroes() {
    assert_sum_eq(&[]);
    assert_sum_eq(&[0]);
    assert_sum_eq(&[0, 0]);
    assert_sum_eq(&[0; 100]);
    assert_sum_eq(&[0; 1024]);
    assert_sum_eq(&[0; 512 * 1024]);
  }

  #[test]
  fn ones() {
    assert_sum_eq(&[]);
    assert_sum_eq(&[1]);
    assert_sum_eq(&[1, 1]);
    assert_sum_eq(&[1; 100]);
    assert_sum_eq(&[1; 1024]);
    assert_sum_eq(&[1; 512 * 1024]);
  }

  #[test]
  fn random() {
    let mut random = [0; 512 * 1024];
    rand::thread_rng().fill(&mut random[..]);

    assert_sum_eq(&random[..1]);
    assert_sum_eq(&random[..100]);
    assert_sum_eq(&random[..1024]);
    assert_sum_eq(&random[..512 * 1024]);
  }

  /// Example calculation from https://en.wikipedia.org/wiki/Adler-32.
  #[test]
  fn wiki() {
    assert_sum_eq(b"Wikipedia");
  }

  fn assert_sum_eq(data: &[u8]) {
    if let Some(update) = super::get_imp() {
      let (a, b) = update(1, 0, data);
      let left = u32::from(b) << 16 | u32::from(a);
      let right = adler::adler32_slice(data);

      assert_eq!(left, right, "len({})", data.len());
    }
  }
}
simd_adler32/imp/wasm.rs

simd_adler32/imp/
wasm.rs
