From 24efacc29cf4d2a52d09e8d424a55ac3e88cbd7d Mon Sep 17 00:00:00 2001 From: x Date: Fri, 6 Jan 2017 18:06:52 +0100 Subject: [PATCH] IcBench: Integer Compression codecs --- plugins.cc | 93 ++++++++++++++++++++++++++++++++---------------------- 1 file changed, 55 insertions(+), 38 deletions(-) diff --git a/plugins.cc b/plugins.cc index afa90ce..0fe9cd2 100644 --- a/plugins.cc +++ b/plugins.cc @@ -93,6 +93,8 @@ enum { #define C_TURBOPFORV CODEC0V TB_PFOR128, TB_PFOR256, + TB_PFORN128, + TB_PFORN256, TB_PACK128H, TB_PACK128V, TB_PACK256H, @@ -353,7 +355,9 @@ unsigned char sbuf[BLK_SIZE*2+64]; //------------------------------------------------- registry ------------------------------------------------------------------------------------------------- struct plugs plugs[] = { { TB_PFOR128, "TurboPFor", C_TURBOPFOR, "", "TurboPFor", "", "https://github.com/powturbo/TurboPFor", "", 0,BLK_V128 }, + { TB_PFORN128, "TurboPForN", C_TURBOPFOR, "", "TurboPFor", "", "https://github.com/powturbo/TurboPFor", ""/*, 0,BLK_V128*/ }, { TB_PFOR256, "TurboPFor256", C_TURBOPFOR, "", "TurboPFor", "", "https://github.com/powturbo/TurboPFor", "", 0,BLK_V256 }, + { TB_PFORN256, "TurboPForN256", C_TURBOPFOR, "", "TurboPFor", "", "https://github.com/powturbo/TurboPFor", ""/*, 0,BLK_V256*/ }, { TB_PFORDA, "TurboPForDA", C_TURBOPFOR, "", "TurboPFor", "", "https://github.com/powturbo/TurboPFor", "", 0,BLK_V128 }, { TB_FOR, "TurboFor", C_TURBOPFOR, "", "TurboPFor", "", "https://github.com/powturbo/TurboPFor", "", 0,BLK_V128 }, { TB_FORDA, "TurboForDA", C_TURBOPFOR, "", "TurboPFor", "", "https://github.com/powturbo/TurboPFor", "", 0,BLK_V128 }, @@ -411,17 +415,17 @@ struct plugs plugs[] = { { LI_HORPACK, "LI_HorPack", C_LITTLEPACK, "", "LittlePack", "", "", ""/*, 0,BLK_V128*/}, { LI_BMIPACK, "LI_BMIPack256", C_LITTLEPACK, "", "LittlePack", "", "", ""/*, 0,BLK_V128*/}, - { P_QMX, "qmx", C_QMX, "", "QMX", "https://bitbucket.org/andrewtrotman/bench.git", "", ""/*, 0,BLK_V128*/}, - { P_QMX2, "qmx2", C_QMX, "", "QMX", "https://bitbucket.org/andrewtrotman/bench.git", "", ""/*, 0,BLK_V128*/}, - { P_QMX3, "qmx3", C_QMX, "", "QMX", "https://bitbucket.org/andrewtrotman/bench.git", "", ""/*, 0,BLK_V128*/}, - { P_QMX4, "qmx4", C_QMX, "", "QMX", "https://bitbucket.org/andrewtrotman/bench.git", "", ""/*, 0,BLK_V128*/}, + { P_QMX, "qmx", C_QMX, "", "QMX", "https://bitbucket.org/andrewtrotman/bench.git", "", ""/*, 0,BLK_V128*/}, + { P_QMX2, "qmx2", C_QMX, "", "QMX", "https://bitbucket.org/andrewtrotman/bench.git", "", ""/*, 0,BLK_V128*/}, + { P_QMX3, "qmx3", C_QMX, "", "QMX", "https://bitbucket.org/andrewtrotman/bench.git", "", ""/*, 0,BLK_V128*/}, + { P_QMX4, "qmx4", C_QMX, "", "QMX", "https://bitbucket.org/andrewtrotman/bench.git", "", ""/*, 0,BLK_V128*/}, { P_LZT, "LzTurbo", C_LZTURBO, "", "LzTurbo", "", "https://sites.google.com/site/powturbo", "20,21,22,32", 0,BLK_SIZE }, { P_VSHUF, "VSimpleANS", C_LZTURBO, "", "LzTurbo", "", "https://sites.google.com/site/powturbo", "20,21,22,32", 0,BLK_SIZE }, - { LZ4_, "lz4", C_LZ4, "", "Lz4", "BSD license", "https://github.com/Cyan4973/lz4", "" }, - { LZ4_S1, "lz4_bitshufle", C_LZ4, "", "Lz4", "BSD license", "https://github.com/Cyan4973/lz4", "", 0,BLK_SIZE }, - { LZ4_S4, "lz4_tp4", C_LZ4, "", "Lz4", "BSD license", "https://github.com/Cyan4973/lz4", "", 0,BLK_SIZE }, - { LZ4_S8, "lz4_tp8", C_LZ4, "", "Lz4", "BSD license", "https://github.com/Cyan4973/lz4", "", 0,BLK_SIZE }, + { LZ4_, "lz4", C_LZ4, "", "Lz4", "BSD license", "https://github.com/Cyan4973/lz4", "" }, + { LZ4_S1, "lz4_bitshufle", C_LZ4, "", "lz4+bitshuffle", "BSD license", "https://github.com/Cyan4973/lz4", "", 0,BLK_SIZE }, + { LZ4_S4, "lz4_tp4", C_LZ4, "", "lz4+Nibble transpose", "BSD license", "https://github.com/Cyan4973/lz4", "", 0,BLK_SIZE }, + { LZ4_S8, "lz4_tp8", C_LZ4, "", "lz+byte transpose", "BSD license", "https://github.com/Cyan4973/lz4", "", 0,BLK_SIZE }, { P_ZLIB, "zlib", C_ZLIB, "1.2.8","zlib", "zlib license", "http://zlib.net\thttps://github.com/madler/zlib", "", 0,BLK_SIZE }, // { P_ZSTD, "zstd", C_ZSTD, "1.0.0","ZSTD", "BSD license+Patents","https://github.com/facebook/zstd", "1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22" }, //----------------------------------------------------------------------------------- @@ -553,14 +557,17 @@ unsigned char *codcomps(unsigned char *_in, unsigned _n, unsigned char *out, int if(inc) { return efano1enc32(in, n, out, x+1); } else { return efanoenc32( in, n, out, x ); } #if C_TURBOPFORV - case TB_ELIASFANOV: x = *in++; VBPUT32(out, x); --n; + case TB_ELIASFANOV: x = *in++; VBPUT32(out, x); --n; if(inc) { return n == 128?efano1enc128v32(in, n, out, x+1):efano1enc32(in, n, out, x+1); } else { return n == 128?efanoenc128v32( in, n, out, x ):efanoenc32( in, n, out, x ); } - case TB_PFOR128: bitdelta32( in+1, --n, pa, in[0], inc); VBPUT32(out, in[0]); return n == 128?p4enc128v32(pa, n, out):p4enc32(pa, n, out); + case TB_PFOR128: if(inc) { --n; VBPUT32(out, in[0]); return n < 128?p4d1enc32(in+1, n, out, in[0]):p4d1enc128v32(in+1, n, out, in[0]); } + else { --n; VBPUT32(out, in[0]); return n < 128?p4denc32( in+1, n, out, in[0]):p4denc128v32( in+1, n, out, in[0]); } + case TB_PFORN128: --n; VBPUT32(out, in[0]); return inc?p4nd1enc128v32(in+1, n, out, in[0]):p4ndenc128v32(in+1, n, out, in[0]); case TB_PACK128V: if(inc) { b = bitd132(in+1, --n, in[0]); VBPUT32(out, in[0]); *out++=b; return n < 128?bitd1pack32(in+1, n, out, in[0], b):bitd1pack128v32(in+1, n, out, in[0], b); } else { b = bitd32( in+1, --n, in[0]); VBPUT32(out, in[0]); *out++=b; return n < 128?bitdpack32( in+1, n, out, in[0], b):bitdpack128v32( in+1, n, out, in[0], b); } #ifdef __AVX2__ case TB_PFOR256: bitdelta32( in+1, --n, pa, in[0], inc); VBPUT32(out, in[0]); return n == 256?p4enc256v32(pa, n, out):p4enc32(pa, n, out); + case TB_PFORN256: --n; VBPUT32(out, in[0]); return inc?p4nd1enc256v32(in+1, n, out, in[0]):p4ndenc256v32(in+1, n, out, in[0]); case TB_PACK256V: b = bitdelta32( in+1, --n, pa, in[0], inc); VBPUT32(out, in[0]); *out++=b; return n == 256?bitpack256v32(pa, n, out,b):bitpack32(pa, n, out,b); //case TB_PACK256V: if(inc) { b = bitd132(in+1, --n, in[0]); VBPUT32(out, in[0]); *out++=b; return n < 256?bitd1pack32(in+1, n, out, in[0], b):bitd1pack256v32(in+1, n, out, in[0], b); } // else { b = bitd32( in+1, --n, in[0]); VBPUT32(out, in[0]); *out++=b; return n < 256?bitdpack32( in+1, n, out, in[0], b):bitdpack256v32( in+1, n, out, in[0], b); } @@ -697,6 +704,9 @@ unsigned char *coddecomps(unsigned char *in, unsigned _n, unsigned char *_out, i case TB_PFOR128: VBGET32(in, x); *out++ = x; --n; //__builtin_prefetch(in+256); if(inc) { return n==128?p4d1dec128v32( in, n, out, x ):p4d1dec32(in, n, out, x); } else { return n==128?p4ddec128v32( in, n, out, x ):p4ddec32( in, n, out, x); } + case TB_PFORN128: VBGET32(in, x); *out++ = x; --n; //__builtin_prefetch(in+256); + if(inc) { return p4nd1dec128v32( in, n, out, x ); } + else { return p4nddec128v32( in, n, out, x ); } case TB_PACK128V: VBGET32(in, x);*out = x; b = *in++; if(n <= 128) { return inc?bitd1unpack32( in, n-1, out+1, x, b):bitdunpack32( in,n-1, out+1, x, b); } else { return inc?bitd1unpack128v32(in,n, out+1, x, b):bitdunpack128v32(in,n, out+1, x, b); } @@ -704,12 +714,15 @@ unsigned char *coddecomps(unsigned char *in, unsigned _n, unsigned char *_out, i case TB_PFOR256: VBGET32(in, x); *out++ = x; --n; //__builtin_prefetch(in+256); if(inc) { return n==256?p4d1dec256v32( in, n, out, x ):p4d1dec32(in, n, out, x); } else { return n==256?p4ddec256v32( in, n, out, x ):p4ddec32( in, n, out, x); } + case TB_PFORN256: VBGET32(in, x); *out++ = x; --n; + if(inc) { return p4nd1dec256v32( in, n, out, x ); } + else { return p4nddec256v32( in, n, out, x ); } /*case TB_PACK256V: VBGET32(in, x);*out = x; b = *in++; if(n <= 256) return inc?bitd1unpack32( in, n-1, out+1, x, b):bitdunpack32( in, n-1, out+1, x, b); else { in = bitunpack256v32( in, out+1, b);bitundx32(out, n, -inc, inc); } break;*/ case TB_PACK256V: VBGET32(in, x);*out = x; b = *in++; if(n <= 256) { return inc?bitd1unpack32( in, n-1, out+1, x, b):bitdunpack32( in, n-1, out+1, x, b); } - else { return inc?bitd1unpack256v32(in, out+1, x, b):bitdunpack256v32( in, out+1, x, b); } + else { return inc?bitd1unpack256v32(in, n-1, out+1, x, b):bitdunpack256v32( in, n-1, out+1, x, b); } #endif #endif @@ -816,28 +829,30 @@ unsigned char *codcomp(unsigned char *_in, unsigned _n, unsigned char *out, int int i,xb; switch(codec&0x3f) { //---------- copy ---------------------------------------------------- - case P_COPY: return u32enc( in, n, (unsigned *)out); - case P_MCPY: return u32enc( in, n, (unsigned *)out); //memcpy(out, _in, _n); return out+_n; + case P_COPY: return u32enc( in, n, (unsigned *)out); + case P_MCPY: return u32enc( in, n, (unsigned *)out); //memcpy(out, _in, _n); return out+_n; // --------- variable byte ------------------------------------------- - case TB_VBYTE: return vbenc32( in, n, out); - case TB_VSIMPLE: return vsenc32( in, n, out); + case TB_VBYTE: return vbenc32( in, n, out); + case TB_VSIMPLE: return vsenc32( in, n, out); #if C_SIMPLE8B - case AM_SIMPLE8B: return vs8benc( in, n, out); + case AM_SIMPLE8B: return vs8benc( in, n, out); #endif - case TB_ELIASFANO:return out; + case TB_ELIASFANO: return out; case TB_FOR : case TB_FORDA: case TB_PACK128H: - case TB_PACK: if(b < 0) { BITSIZE32(in, n, b); *out++ = b; } return bitpack32(in, n, out, b); - case TB_PFORDA: return p4encx32( in, n, out); + case TB_PACK: if(b < 0) { BITSIZE32(in, n, b); *out++ = b; } return bitpack32(in, n, out, b); + case TB_PFORDA: return p4encx32( in, n, out); #if C_TURBOPFORV case TB_ELIASFANOV:return out; - case TB_PFOR128: return n == 128?p4enc128v32(in, n, out):p4enc32(in, n, out); - case TB_PACK128V: if(b < 0) { BITSIZE32(in, n, b); *out++ = b; } return n != 128?bitpack32(in, n, out, b):bitpack128v32(in, n, out, b); + case TB_PFOR128: return n == 128?p4enc128v32(in, n, out):p4enc32(in, n, out); + case TB_PFORN128: return p4nenc128v32(in, n, out); + case TB_PACK128V: if(b < 0) { BITSIZE32(in, n, b); *out++ = b; } return n != 128?bitpack32(in, n, out, b):bitpack128v32(in, n, out, b); #ifdef __AVX2__ - case TB_PFOR256: return n == 256?p4enc256v32(in, n, out):p4enc32(in, n, out); - case TB_PACK256V: if(b < 0) { BITSIZE32(in, n, b); *out++ = b; } return n != 256?bitpack32(in, n, out, b):bitpack256v32(in, n, out, b); + case TB_PFOR256: return n == 256?p4enc256v32(in, n, out):p4enc32(in, n, out); + case TB_PFORN256: return p4nenc256v32(in, n, out); + case TB_PACK256V: if(b < 0) { BITSIZE32(in, n, b); *out++ = b; } return n != 256?bitpack32(in, n, out, b):bitpack256v32(in, n, out, b); #endif #endif @@ -992,10 +1007,10 @@ unsigned char *codcomp(unsigned char *_in, unsigned _n, unsigned char *out, int unsigned char *coddecomp(unsigned char *in, unsigned _n, unsigned char *_out, int outlen, int codec, int lev, char *prm, int b) { unsigned *out = (unsigned *)_out, n = (outlen+3) / 4; switch(codec&0x3f) { - case P_COPY: return u32dec( (unsigned *)in, n, out); - case P_MCPY: return u32dec( (unsigned *)in, n, out); //memcpy(_out, in, _n); return in+_n; + case P_COPY: return u32dec( (unsigned *)in, n, out); + case P_MCPY: return u32dec( (unsigned *)in, n, out); //memcpy(_out, in, _n); return in+_n; - // case P_VU: return vudec32( in, n, out); + // case P_VU: return vudec32( in, n, out); case TB_VBYTE: return vbdec32( in, n, out); case TB_PFORDA : return p4decx32( in, n, out); @@ -1008,14 +1023,16 @@ unsigned char *coddecomp(unsigned char *in, unsigned _n, unsigned char *_out, in case TB_FORDA: if(b < 0) b = *in++; return _bitunpackx32( in, n, out, b); case TB_PACK: if(b < 0) b = *in++; return bitunpack32( in, n, out, b); #if C_TURBOPFORV - case TB_PFOR128 : __builtin_prefetch(in+256);return n == 128?p4dec128v32(in, n, out):p4dec32(in, n, out); + case TB_PFOR128 : __builtin_prefetch(in+256);return n == 128?p4dec128v32(in, n, out):p4dec32(in, n, out); + case TB_PFORN128 : return p4ndec128v32(in, n, out); #ifdef __AVX2__ - case TB_PFOR256 : __builtin_prefetch(in+256);return n == 256?p4dec256v32(in, n, out):p4dec32(in, n, out); - case TB_PACK256V: if(b < 0) b = *in++; return n != 256?bitunpack32(in, n, out, b):bitunpack256v32(in, n, out, b); + case TB_PFOR256 : __builtin_prefetch(in+256);return n == 256?p4dec256v32(in, n, out):p4dec32(in, n, out); + case TB_PFORN256 : return p4ndec256v32(in, n, out); + case TB_PACK256V: if(b < 0) b = *in++; return n != 256?bitunpack32(in, n, out, b):bitunpack256v32(in, n, out, b); #endif case TB_ELIASFANOV: return in; - //case TB_PACK128H: if(b < 0) b = *in++; return n != 128?bitunpack32(in, n, out, b):bitunpack128h32(in, out, b); - case TB_PACK128V: if(b < 0) b = *in++; return n != 128?bitunpack32(in, n, out, b):bitunpack128v32(in, n, out, b); + //case TB_PACK128H: if(b < 0) b = *in++; return n != 128?bitunpack32(in, n, out, b):bitunpack128h32(in, out, b); + case TB_PACK128V: if(b < 0) b = *in++; return n != 128?bitunpack32(in, n, out, b):bitunpack128v32(in, n, out, b); #endif #if C_FASTPFOR @@ -1077,24 +1094,24 @@ unsigned char *coddecomp(unsigned char *in, unsigned _n, unsigned char *_out, in #endif #if C_POLYCOM - case PC_VBYTE: return vbpolydec(in, n, out); - case PC_SIMPLE16: return vs16dec( (unsigned *)in, n, out); - case PC_RICE: return rcdec32( (unsigned *)in, n, out); - case PC_OPTPFD : return optpfddec32( in, n, out); //if(n < 128) return vbytedec(in, n, out); else { unsigned all_array[2048]; return (unsigned char *)detailed_p4_decode(out, (unsigned *)in, all_array); } + case PC_VBYTE: return vbpolydec(in, n, out); + case PC_SIMPLE16: return vs16dec( (unsigned *)in, n, out); + case PC_RICE: return rcdec32( (unsigned *)in, n, out); + case PC_OPTPFD : return optpfddec32( in, n, out); //if(n < 128) return vbytedec(in, n, out); else { unsigned all_array[2048]; return (unsigned char *)detailed_p4_decode(out, (unsigned *)in, all_array); } #endif #if C_SIMDCOMP //case SC_PACK: if(b < 0) b = *in++; return fastunpack32((uint32_t *)in, n, out, b); case SC_FOR: - case SC_SIMDPACK128: if(b < 0) b = *in++; return (unsigned char *)simdunpack_length( (__m128i *)in, n, out, b); + case SC_SIMDPACK128: if(b < 0) b = *in++; return (unsigned char *)simdunpack_length( (__m128i *)in, n, out, b); //case SC_FORDA: if(b < 0) b = *in++; return simdfor_selectx(in, n, out, 0, b); #ifdef __AVX2__ - case SC_SIMDPACK256: if(b < 0) b = *in++; return (unsigned char *)avxunpackn( (unsigned *)in, n, out, b); + case SC_SIMDPACK256: if(b < 0) b = *in++; return (unsigned char *)avxunpackn( (unsigned *)in, n, out, b); #endif #endif #if C_STREAMVBYTE - case P_STREAMVBYTE: return in + streamvbyte_decode(in, out, n); + case P_STREAMVBYTE: return in + streamvbyte_decode(in, out, n); #endif #if C_QMX //case P_QMX: return qmx_dec(in+4, ctou32(in), out, n);