Merge pull request #27 from Alexhuszagh/read_u64

Added read_u64 optimizations to big endian.
aldanor · May 24, 2021 · ec1b7d4 · ec1b7d4
2 parents 0c714bb + bc9ed2e
commit ec1b7d4
Show file tree

Hide file tree

Showing 3 changed files with 51 additions and 42 deletions.
diff --git a/src/common.rs b/src/common.rs
@@ -90,11 +90,8 @@ impl<'a> AsciiStr<'a> {
     #[inline]
     pub fn read_u64(&self) -> u64 {
         debug_assert!(self.check_len(8));
-        let mut value = 0_u64;
-        let src = self.ptr;
-        let dst = &mut value as *mut _ as *mut u8;
-        unsafe { ptr::copy_nonoverlapping(src, dst, 8) };
-        value
+        let src = self.ptr as *const u64;
+        u64::from_le(unsafe { ptr::read_unaligned(src) })
     }
 
     #[inline]
@@ -159,26 +156,22 @@ pub trait ByteSlice: AsRef<[u8]> + AsMut<[u8]> {
     #[inline]
     fn read_u64(&self) -> u64 {
         debug_assert!(self.as_ref().len() >= 8);
-        let mut value = 0_u64;
-        let src = self.as_ref().as_ptr();
-        let dst = &mut value as *mut _ as *mut u8;
-        unsafe { ptr::copy_nonoverlapping(src, dst, 8) };
-        value
+        let src = self.as_ref().as_ptr() as *const u64;
+        u64::from_le(unsafe { ptr::read_unaligned(src) })
     }
 
     #[inline]
     fn write_u64(&mut self, value: u64) {
         debug_assert!(self.as_ref().len() >= 8);
-        let src = &value as *const _ as *const u8;
-        let dst = self.as_mut().as_mut_ptr();
-        unsafe { ptr::copy_nonoverlapping(src, dst, 8) };
+        let dst = self.as_mut().as_mut_ptr() as *mut u64;
+        unsafe { ptr::write_unaligned(dst, u64::to_le(value)) };
     }
 }
 
 impl ByteSlice for [u8] {}
 
 #[inline]
-pub fn is_8digits_le(v: u64) -> bool {
+pub fn is_8digits(v: u64) -> bool {
     let a = v.wrapping_add(0x4646_4646_4646_4646);
     let b = v.wrapping_sub(0x3030_3030_3030_3030);
     (a | b) & 0x8080_8080_8080_8080 == 0
@@ -212,3 +205,23 @@ impl AdjustedMantissa {
         }
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_read_write_u64() {
+        let bytes = b"01234567";
+        let string = AsciiStr::new(bytes);
+        let int = string.read_u64();
+        assert_eq!(int, 0x3736353433323130);
+
+        let int = bytes.read_u64();
+        assert_eq!(int, 0x3736353433323130);
+
+        let mut slc = [0u8; 8];
+        slc.write_u64(0x3736353433323130);
+        assert_eq!(&slc, bytes);
+    }
+}
diff --git a/src/decimal.rs b/src/decimal.rs
@@ -1,6 +1,6 @@
 use core::fmt::{self, Debug};
 
-use crate::common::{is_8digits_le, parse_digits, ByteSlice};
+use crate::common::{is_8digits, parse_digits, ByteSlice};
 
 #[derive(Clone)]
 pub struct Decimal {
@@ -204,16 +204,14 @@ pub fn parse_decimal(mut s: &[u8]) -> Decimal {
         if d.num_digits == 0 {
             s = s.skip_chars(b'0');
         }
-        if cfg!(target_endian = "little") {
-            while s.len() >= 8 && d.num_digits + 8 < Decimal::MAX_DIGITS {
-                let v = s.read_u64();
-                if !is_8digits_le(v) {
-                    break;
-                }
-                d.digits[d.num_digits..].write_u64(v - 0x3030_3030_3030_3030);
-                d.num_digits += 8;
-                s = s.advance(8);
+        while s.len() >= 8 && d.num_digits + 8 < Decimal::MAX_DIGITS {
+            let v = s.read_u64();
+            if !is_8digits(v) {
+                break;
             }
+            d.digits[d.num_digits..].write_u64(v - 0x3030_3030_3030_3030);
+            d.num_digits += 8;
+            s = s.advance(8);
         }
         parse_digits(&mut s, |digit| d.try_add_digit(digit));
         d.decimal_point = s.len() as i32 - first.len() as i32;

diff --git a/src/number.rs b/src/number.rs
@@ -1,4 +1,4 @@
-use crate::common::{is_8digits_le, AsciiStr, ByteSlice};
+use crate::common::{is_8digits, AsciiStr, ByteSlice};
 use crate::float::Float;
 
 const MIN_19DIGIT_INT: u64 = 100_0000_0000_0000_0000;
@@ -70,7 +70,7 @@ impl Number {
 }
 
 #[inline]
-fn parse_8digits_le(mut v: u64) -> u64 {
+fn parse_8digits(mut v: u64) -> u64 {
     const MASK: u64 = 0x0000_00FF_0000_00FF;
     const MUL1: u64 = 0x000F_4240_0000_0064;
     const MUL2: u64 = 0x0000_2710_0000_0001;
@@ -98,22 +98,20 @@ fn try_parse_19digits(s: &mut AsciiStr<'_>, x: &mut u64) {
 }
 
 #[inline]
-fn try_parse_8digits_le(s: &mut AsciiStr<'_>, x: &mut u64) {
+fn try_parse_8digits(s: &mut AsciiStr<'_>, x: &mut u64) {
     // may cause overflows, to be handled later
-    if cfg!(target_endian = "little") {
-        if let Some(v) = s.try_read_u64() {
-            if is_8digits_le(v) {
-                *x = x
-                    .wrapping_mul(1_0000_0000)
-                    .wrapping_add(parse_8digits_le(v));
-                s.step_by(8);
-                if let Some(v) = s.try_read_u64() {
-                    if is_8digits_le(v) {
-                        *x = x
-                            .wrapping_mul(1_0000_0000)
-                            .wrapping_add(parse_8digits_le(v));
-                        s.step_by(8);
-                    }
+    if let Some(v) = s.try_read_u64() {
+        if is_8digits(v) {
+            *x = x
+                .wrapping_mul(1_0000_0000)
+                .wrapping_add(parse_8digits(v));
+            s.step_by(8);
+            if let Some(v) = s.try_read_u64() {
+                if is_8digits(v) {
+                    *x = x
+                        .wrapping_mul(1_0000_0000)
+                        .wrapping_add(parse_8digits(v));
+                    s.step_by(8);
                 }
             }
         }
@@ -180,7 +178,7 @@ pub fn parse_number(s: &[u8]) -> Option<(Number, usize)> {
     if s.check_first(b'.') {
         s.step();
         let before = s;
-        try_parse_8digits_le(&mut s, &mut mantissa);
+        try_parse_8digits(&mut s, &mut mantissa);
         try_parse_digits(&mut s, &mut mantissa);
         n_after_dot = s.offset_from(&before);
         exponent = -n_after_dot as i64;