1 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2 ; Copyright(c) 2011-2015 Intel Corporation All rights reserved.
4 ; Redistribution and use in source and binary forms, with or without
5 ; modification, are permitted provided that the following conditions
7 ; * Redistributions of source code must retain the above copyright
8 ; notice, this list of conditions and the following disclaimer.
9 ; * Redistributions in binary form must reproduce the above copyright
10 ; notice, this list of conditions and the following disclaimer in
11 ; the documentation and/or other materials provided with the
13 ; * Neither the name of Intel Corporation nor the names of its
14 ; contributors may be used to endorse or promote products derived
15 ; from this software without specific prior written permission.
17 ; THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
18 ; "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
19 ; LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
20 ; A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
21 ; OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
22 ; SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
23 ; LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
24 ; DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
25 ; THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
26 ; (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
27 ; OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
31 ;;; gf_vect_mad_avx512(len, vec, vec_i, mul_array, src, dest);
34 %include "reg_sizes.asm"
36 %ifdef HAVE_AS_KNOWS_AVX512
38 %ifidn __OUTPUT_FORMAT__, elf64
52 %ifidn __OUTPUT_FORMAT__, win64
57 %define arg4 r12 ; must be saved and loaded
62 %define stack_size 16*3 + 3*8
63 %define arg(x) [rsp + stack_size + PS + PS*x]
64 %define func(x) proc_frame x
68 vmovdqa [rsp+16*0],xmm6
69 vmovdqa [rsp+16*1],xmm7
70 vmovdqa [rsp+16*2],xmm8
71 save_reg r12, 3*16 + 0*8
72 save_reg r15, 3*16 + 1*8
79 vmovdqa xmm6, [rsp+16*0]
80 vmovdqa xmm7, [rsp+16*1]
81 vmovdqa xmm8, [rsp+16*2]
82 mov r12, [rsp + 3*16 + 0*8]
83 mov r15, [rsp + 3*16 + 1*8]
88 ;;; gf_vect_mad_avx512(len, vec, vec_i, mul_array, src, dest)
92 %define mul_array arg3
97 %ifndef EC_ALIGNED_ADDR
98 ;;; Use Un-aligned load/store
100 %define XSTR vmovdqu8
102 ;;; Use Non-temporal load/stor
107 %define XLDR vmovntdqa
108 %define XSTR vmovntdq
126 %define xgft_loy ymm7
130 global gf_vect_mad_avx512:function
131 func(gf_vect_mad_avx512)
137 vpbroadcastb xmask0f, tmp ;Construct mask 0x0f0f0f...
138 sal vec_i, 5 ;Multiply by 32
139 vmovdqu8 xgft_loy, [vec_i+mul_array] ;Load array Cx{00}..{0f}, Cx{00}..{f0}
140 vshufi64x2 xgft_hi, xgft_lo, xgft_lo, 0x55
141 vshufi64x2 xgft_lo, xgft_lo, xgft_lo, 0x00
146 XLDR xd, [dest+pos] ;Get next dest vector
147 XLDR x0, [src+pos] ;Get next source vector
149 vpandq xtmpa, x0, xmask0f ;Mask low src nibble in bits 4-0
150 vpsraw x0, x0, 4 ;Shift to put high nibble into bits 4-0
151 vpandq x0, x0, xmask0f ;Mask high src nibble in bits 4-0
153 vpshufb xtmph {k1}{z}, xgft_hi, x0 ;Lookup mul table of high nibble
154 vpshufb xtmpl {k1}{z}, xgft_lo, xtmpa ;Lookup mul table of low nibble
155 vpxorq xtmph, xtmph, xtmpl ;GF add high and low partials
156 vpxorq xd, xd, xtmph ;xd += partial
159 add pos, 64 ;Loop on 64 bytes at a time
169 lea tmp, [len + 64 - 1]
173 mov pos, len ;Overlapped offset length-64
174 jmp .loop64 ;Do one more overlap pass
189 %ifidn __OUTPUT_FORMAT__, win64
190 global no_gf_vect_mad_avx512
191 no_gf_vect_mad_avx512:
193 %endif ; ifdef HAVE_AS_KNOWS_AVX512