daiduo/TurboPFor-Integer-Compression
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

TurboRLE: decode

Browse Source
This commit is contained in:
x
2019-10-16 19:47:31 +02:00
parent 35fd0926ac
commit 018b99b9ed
1 changed files with 15 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
trled.c Executable file → Normal file
View File

@ -41,6 +41,12 @@
#endif
#include "sse_neon.h"
#ifdef __ARM_NEON
#define PREFETCH(_ip_,_rw_)
#else
#define PREFETCH(_ip_,_rw_) __builtin_prefetch(_ip_,_rw_)
#endif
#include "trle.h"
#include "trle_.h"
//------------------------------------- RLE 8 with Escape char ------------------------------------------------------------------
@ -74,9 +80,9 @@ unsigned _srled8(const unsigned char *__restrict in, unsigned char *__restrict o
v = _mm_loadu_si128((__m128i*)ip); _mm_storeu_si128((__m128i *)op, v); mask = _mm_movemask_epi8(_mm_cmpeq_epi8(v, ev)); if(mask) goto a; ip += 16; op += 16;
#endif
#endif
__builtin_prefetch(ip+512, 0);
PREFETCH(ip+512, 0);
continue;
a: r = ctz32(mask); ip += r+1; __builtin_prefetch(ip+512, 0);
a: r = ctz32(mask); ip += r+1; PREFETCH(ip+512, 0);
op += r;
#else
if(likely((c = *ip++) != e)) { *op++ = c; continue; }
@ -198,7 +204,7 @@ unsigned _trled(const unsigned char *__restrict in, unsigned char *__restrict ou
vlzget(ip, i, m, c-1);
c = *ip++;
i += TMIN;
rmemset(op,c,i); __builtin_prefetch(ip+512, 0);
rmemset(op,c,i); PREFETCH(ip+512, 0);
}
while(op < out+outlen) {
@ -294,22 +300,22 @@ unsigned TEMPLATE2(_srled, USIZE)(const unsigned char *__restrict in, unsigned c
uint32_t mask;
__m256i v = _mm256_loadu_si256((__m256i*)ip); _mm256_storeu_si256((__m256i *)op, v); mask = _mm256_movemask_epi8(TEMPLATE2(_mm256_cmpeq_epi,USIZE)(v, ev)); if(mask) goto a; ip += 32; op += 256/USIZE;
v = _mm256_loadu_si256((__m256i*)ip); _mm256_storeu_si256((__m256i *)op, v); mask = _mm256_movemask_epi8(TEMPLATE2(_mm256_cmpeq_epi,USIZE)(v, ev)); if(mask) goto a; ip += 32; op += 256/USIZE;
__builtin_prefetch(ip+512, 0);
PREFETCH(ip+512, 0);
continue;
a: r = ctz32(mask)/(USIZE/8);
op += r;
ip += (r+1)*sizeof(uint_t); __builtin_prefetch(ip+512, 0);
ip += (r+1)*sizeof(uint_t); PREFETCH(ip+512, 0);
#elif (__SSE__ != 0 /*|| __ARM_NEON != 0*/) && USIZE != 64
uint32_t mask;
__m128i v = _mm_loadu_si128((__m128i*)ip); _mm_storeu_si128((__m128i *)op, v); mask = _mm_movemask_epi8(TEMPLATE2(_mm_cmpeq_epi,USIZE)(v, ev)); if(mask) goto a; ip += 16; op += 128/USIZE;
v = _mm_loadu_si128((__m128i*)ip); _mm_storeu_si128((__m128i *)op, v); mask = _mm_movemask_epi8(TEMPLATE2(_mm_cmpeq_epi,USIZE)(v, ev)); if(mask) goto a; ip += 16; op += 128/USIZE;
v = _mm_loadu_si128((__m128i*)ip); _mm_storeu_si128((__m128i *)op, v); mask = _mm_movemask_epi8(TEMPLATE2(_mm_cmpeq_epi,USIZE)(v, ev)); if(mask) goto a; ip += 16; op += 128/USIZE;
v = _mm_loadu_si128((__m128i*)ip); _mm_storeu_si128((__m128i *)op, v); mask = _mm_movemask_epi8(TEMPLATE2(_mm_cmpeq_epi,USIZE)(v, ev)); if(mask) goto a; ip += 16; op += 128/USIZE;
__builtin_prefetch(ip+512, 0);
PREFETCH(ip+512, 0);
continue;
a: r = ctz32(mask)/(USIZE/8);
op += r;
ip += (r+1)*sizeof(uint_t); __builtin_prefetch(ip+512, 0);
ip += (r+1)*sizeof(uint_t); PREFETCH(ip+512, 0);
#else
if(((c = ctout(ip)) == e)) goto a; ip += sizeof(uint_t); *op++ = c;
if(((c = ctout(ip)) == e)) goto a; ip += sizeof(uint_t); *op++ = c;
@ -318,9 +324,9 @@ unsigned TEMPLATE2(_srled, USIZE)(const unsigned char *__restrict in, unsigned c
if(((c = ctout(ip)) == e)) goto a; ip += sizeof(uint_t); *op++ = c;
if(((c = ctout(ip)) == e)) goto a; ip += sizeof(uint_t); *op++ = c;
if(((c = ctout(ip)) == e)) goto a; ip += sizeof(uint_t); *op++ = c;
if(((c = ctout(ip)) == e)) goto a; ip += sizeof(uint_t); *op++ = c; __builtin_prefetch(ip +512, 0);
if(((c = ctout(ip)) == e)) goto a; ip += sizeof(uint_t); *op++ = c; PREFETCH(ip +512, 0);
continue;
a: ip += sizeof(uint_t); __builtin_prefetch(ip +512, 0);
a: ip += sizeof(uint_t); PREFETCH(ip +512, 0);
#endif
vlget32(ip, r);
if(likely(r) >= 3) {