search_line_fast in the lexer is documented as being a critical path.
Several architectures already provide an optimized implementation of
this routine using vector operations. This patch provides a version for
ARM using the Neon vector extensions.
Although it has not been tested on big-endian (I don't have anything
big-endian to do native testing on), I believe the code doesn't need
further porting: the use of vld1q_u8() to read the data from memory will
always put the lowest addressed byte into lane 0 of the vector.
Bootstrapped on arm-linux-gnueabi with the necessary options to enable
the use of Neon.
OK for trunk?
* lex.c (search_line_fast): Provide Neon-optimized version for ARM.--- lex.c (revision 185620)
+++ lex.c (local)
@@ -629,6 +629,69 @@ search_line_fast (const uchar *s, const
}
}
+#elif defined (__ARM_NEON__)
+#include "arm_neon.h"
+
+static const uchar *
+search_line_fast (const uchar *s, const uchar *end ATTRIBUTE_UNUSED)
+{
+ const uint8x16_t repl_nl = vdupq_n_u8 ('\n');
+ const uint8x16_t repl_cr = vdupq_n_u8 ('\r');
+ const uint8x16_t repl_bs = vdupq_n_u8 ('\\');
+ const uint8x16_t repl_qm = vdupq_n_u8 ('?');
+ const uint8x16_t xmask = (uint8x16_t) vdupq_n_u64 (0x8040201008040201ULL);
+
+ unsigned int misalign, found, mask;
+ const uint8_t *p;
+ uint8x16_t data;
+
+ /* Align the source pointer. */
+ misalign = (uintptr_t)s & 15;
+ p = (const uint8_t *)((uintptr_t)s & -16);
+ data = vld1q_u8 (p);
+
+ /* Create a mask for the bytes that are valid within the first
+ 16-byte block. The Idea here is that the AND with the mask
+ within the loop is "free", since we need some AND or TEST
+ insn in order to set the flags for the branch anyway. */
+ mask = (-1u << misalign) & 0x;
+
+ /* Main loop, processing 16 bytes at a time. */
+ goto start;
+
+ do
+{
+ uint8x8_t l;
+ uint16x4_t m;
+ uint32x2_t n;
+ uint8x16_t t, u, v, w;
+
+ p += 16;
+ data = vld1q_u8 (p);
+ mask = 0x;
+
+start:
+ t = vceqq_u8 (data, repl_nl);
+ u = vceqq_u8 (data, repl_cr);
+ v = vorrq_u8 (t, vceqq_u8 (data, repl_bs));
+ w = vorrq_u8 (u, vceqq_u8 (data, repl_qm));
+ t = vandq_u8 (vorrq_u8 (v, w), xmask);
+ l = vpadd_u8 (vget_low_u8 (t), vget_high_u8 (t));
+ m = vpaddl_u8 (l);
+ n = vpaddl_u16 (m);
+
+ found = vget_lane_u32 ((uint32x2_t) vorr_u64 ((uint64x1_t) n,
+ vshr_n_u64 ((uint64x1_t) n, 24)), 0);
+ found &= mask;
+}
+ while (!found);
+
+ /* FOUND contains 1 in bits for which we matched a relevant
+ character. Conversion to the byte index is trivial. */
+ found = __builtin_ctz (found);
+ return (const uchar *)p + found;
+}
+
#else
/* We only have one accellerated alternative. Use a direct call so that
