/*********************************************************************/ /* Copyright 2009, 2010 The University of Texas at Austin. */ /* All rights reserved. */ /* */ /* Redistribution and use in source and binary forms, with or */ /* without modification, are permitted provided that the following */ /* conditions are met: */ /* */ /* 1. Redistributions of source code must retain the above */ /* copyright notice, this list of conditions and the following */ /* disclaimer. */ /* */ /* 2. Redistributions in binary form must reproduce the above */ /* copyright notice, this list of conditions and the following */ /* disclaimer in the documentation and/or other materials */ /* provided with the distribution. */ /* */ /* THIS SOFTWARE IS PROVIDED BY THE UNIVERSITY OF TEXAS AT */ /* AUSTIN ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, */ /* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */ /* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */ /* DISCLAIMED. IN NO EVENT SHALL THE UNIVERSITY OF TEXAS AT */ /* AUSTIN OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, */ /* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES */ /* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE */ /* GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR */ /* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF */ /* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT */ /* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT */ /* OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE */ /* POSSIBILITY OF SUCH DAMAGE. */ /* */ /* The views and conclusions contained in the software and */ /* documentation are those of the authors and should not be */ /* interpreted as representing official policies, either expressed */ /* or implied, of The University of Texas at Austin. */ /*********************************************************************/ #define ASSEMBLER #include "common.h" #define N ARG1 /* rdi */ #define X ARG2 /* rsi */ #define INCX ARG3 /* rdx */ #define Y ARG4 /* rcx */ #ifndef WINDOWS_ABI #define INCY ARG5 /* r8 */ #else #define INCY %r10 #endif #define C %xmm14 #define S %xmm15 #include "l1param.h" PROLOGUE PROFCODE #ifdef WINDOWS_ABI movq 40(%rsp), INCY movsd 48(%rsp), %xmm0 movsd 56(%rsp), %xmm1 #endif SAVEREGISTERS leaq (, INCX, SIZE), INCX leaq (, INCY, SIZE), INCY pshufd $0x44, %xmm0, C pshufd $0x44, %xmm1, S cmpq $0, N jle .L999 cmpq $SIZE, INCX jne .L50 cmpq $SIZE, INCY jne .L50 testq $SIZE, X je .L10 movsd 0 * SIZE(Y), %xmm1 movsd 0 * SIZE(X), %xmm0 movaps %xmm1, %xmm2 movaps %xmm0, %xmm3 mulsd C, %xmm0 mulsd S, %xmm1 mulsd C, %xmm2 mulsd S, %xmm3 addsd %xmm1, %xmm0 subsd %xmm3, %xmm2 movsd %xmm0, 0 * SIZE(X) movsd %xmm2, 0 * SIZE(Y) addq $1 * SIZE, X addq $1 * SIZE, Y decq N jle .L999 ALIGN_2 .L10: testq $SIZE, Y jne .L20 movq N, %rax sarq $4, %rax jle .L14 movaps 0 * SIZE(Y), %xmm1 movaps 2 * SIZE(Y), %xmm3 movaps 4 * SIZE(Y), %xmm9 movaps 6 * SIZE(Y), %xmm11 movaps 0 * SIZE(X), %xmm0 movaps 2 * SIZE(X), %xmm2 movaps 4 * SIZE(X), %xmm8 movaps 6 * SIZE(X), %xmm10 decq %rax jle .L12 ALIGN_3 .L11: #ifdef PREFETCHW PREFETCHW (PREFETCHSIZE + 0) - PREOFFSET(X) #endif movaps %xmm1, %xmm4 mulpd S, %xmm1 movaps %xmm3, %xmm6 mulpd S, %xmm3 movaps %xmm0, %xmm5 mulpd C, %xmm0 movaps %xmm2, %xmm7 mulpd C, %xmm2 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 movaps 8 * SIZE(Y), %xmm1 addpd %xmm3, %xmm2 movaps 10 * SIZE(Y), %xmm3 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 #ifdef PREFETCHW PREFETCHW (PREFETCHSIZE + 0) - PREOFFSET(Y) #endif movaps %xmm0, 0 * SIZE(X) movaps 8 * SIZE(X), %xmm0 movaps %xmm2, 2 * SIZE(X) movaps 10 * SIZE(X), %xmm2 movaps %xmm4, 0 * SIZE(Y) movaps %xmm6, 2 * SIZE(Y) movaps %xmm9, %xmm4 mulpd S, %xmm9 movaps %xmm8, %xmm5 mulpd C, %xmm8 movaps %xmm11, %xmm6 mulpd S, %xmm11 movaps %xmm10, %xmm7 mulpd C, %xmm10 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm9, %xmm8 movaps 12 * SIZE(Y), %xmm9 addpd %xmm11, %xmm10 movaps 14 * SIZE(Y), %xmm11 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm8, 4 * SIZE(X) movaps 12 * SIZE(X), %xmm8 movaps %xmm10,6 * SIZE(X) movaps 14 * SIZE(X), %xmm10 movaps %xmm4, 4 * SIZE(Y) movaps %xmm6, 6 * SIZE(Y) #ifdef PREFETCHW PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(X) #endif movaps %xmm1, %xmm4 mulpd S, %xmm1 movaps %xmm3, %xmm6 mulpd S, %xmm3 movaps %xmm0, %xmm5 mulpd C, %xmm0 movaps %xmm2, %xmm7 mulpd C, %xmm2 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 movaps 16 * SIZE(Y), %xmm1 addpd %xmm3, %xmm2 movaps 18 * SIZE(Y), %xmm3 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 8 * SIZE(X) movaps 16 * SIZE(X), %xmm0 movaps %xmm2, 10 * SIZE(X) movaps 18 * SIZE(X), %xmm2 movaps %xmm4, 8 * SIZE(Y) movaps %xmm6, 10 * SIZE(Y) #ifdef PREFETCHW PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(Y) #endif movaps %xmm9, %xmm4 mulpd S, %xmm9 movaps %xmm8, %xmm5 mulpd C, %xmm8 movaps %xmm11, %xmm6 mulpd S, %xmm11 movaps %xmm10, %xmm7 mulpd C, %xmm10 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm9, %xmm8 movaps 20 * SIZE(Y), %xmm9 addpd %xmm11, %xmm10 movaps 22 * SIZE(Y), %xmm11 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm8, 12 * SIZE(X) movaps 20 * SIZE(X), %xmm8 movaps %xmm10, 14 * SIZE(X) movaps 22 * SIZE(X), %xmm10 movaps %xmm4, 12 * SIZE(Y) movaps %xmm6, 14 * SIZE(Y) addq $16 * SIZE, X addq $16 * SIZE, Y decq %rax jg .L11 ALIGN_3 .L12: movaps %xmm1, %xmm4 mulpd S, %xmm1 movaps %xmm3, %xmm6 mulpd S, %xmm3 movaps %xmm0, %xmm5 mulpd C, %xmm0 movaps %xmm2, %xmm7 mulpd C, %xmm2 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 movaps 8 * SIZE(Y), %xmm1 addpd %xmm3, %xmm2 movaps 10 * SIZE(Y), %xmm3 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 0 * SIZE(X) movaps 8 * SIZE(X), %xmm0 movaps %xmm2, 2 * SIZE(X) movaps 10 * SIZE(X), %xmm2 movaps %xmm4, 0 * SIZE(Y) movaps %xmm6, 2 * SIZE(Y) movaps %xmm9, %xmm4 mulpd S, %xmm9 movaps %xmm8, %xmm5 mulpd C, %xmm8 movaps %xmm11, %xmm6 mulpd S, %xmm11 movaps %xmm10, %xmm7 mulpd C, %xmm10 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm9, %xmm8 movaps 12 * SIZE(Y), %xmm9 addpd %xmm11, %xmm10 movaps 14 * SIZE(Y), %xmm11 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm8, 4 * SIZE(X) movaps 12 * SIZE(X), %xmm8 movaps %xmm10,6 * SIZE(X) movaps 14 * SIZE(X), %xmm10 movaps %xmm4, 4 * SIZE(Y) movaps %xmm6, 6 * SIZE(Y) movaps %xmm1, %xmm4 mulpd S, %xmm1 movaps %xmm3, %xmm6 mulpd S, %xmm3 movaps %xmm0, %xmm5 mulpd C, %xmm0 movaps %xmm2, %xmm7 mulpd C, %xmm2 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 addpd %xmm3, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 8 * SIZE(X) movaps %xmm2, 10 * SIZE(X) movaps %xmm4, 8 * SIZE(Y) movaps %xmm6, 10 * SIZE(Y) movaps %xmm9, %xmm4 mulpd S, %xmm9 movaps %xmm8, %xmm5 mulpd C, %xmm8 movaps %xmm11, %xmm6 mulpd S, %xmm11 movaps %xmm10, %xmm7 mulpd C, %xmm10 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm9, %xmm8 addpd %xmm11, %xmm10 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm8, 12 * SIZE(X) movaps %xmm10, 14 * SIZE(X) movaps %xmm4, 12 * SIZE(Y) movaps %xmm6, 14 * SIZE(Y) addq $16 * SIZE, X addq $16 * SIZE, Y ALIGN_3 .L14: testq $15, N jle .L999 testq $8, N jle .L15 movaps 0 * SIZE(Y), %xmm1 movaps 0 * SIZE(X), %xmm0 movaps 2 * SIZE(Y), %xmm3 movaps 2 * SIZE(X), %xmm2 movaps %xmm1, %xmm4 movaps %xmm0, %xmm5 movaps %xmm3, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 addpd %xmm3, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 0 * SIZE(X) movaps %xmm2, 2 * SIZE(X) movaps %xmm4, 0 * SIZE(Y) movaps %xmm6, 2 * SIZE(Y) movaps 4 * SIZE(Y), %xmm1 movaps 4 * SIZE(X), %xmm0 movaps 6 * SIZE(Y), %xmm3 movaps 6 * SIZE(X), %xmm2 movaps %xmm1, %xmm4 movaps %xmm0, %xmm5 movaps %xmm3, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 addpd %xmm3, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 4 * SIZE(X) movaps %xmm2, 6 * SIZE(X) movaps %xmm4, 4 * SIZE(Y) movaps %xmm6, 6 * SIZE(Y) addq $8 * SIZE, X addq $8 * SIZE, Y ALIGN_3 .L15: testq $4, N jle .L16 movaps 0 * SIZE(Y), %xmm1 movaps 0 * SIZE(X), %xmm0 movaps 2 * SIZE(Y), %xmm3 movaps 2 * SIZE(X), %xmm2 movaps %xmm1, %xmm4 movaps %xmm0, %xmm5 movaps %xmm3, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 addpd %xmm3, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 0 * SIZE(X) movaps %xmm2, 2 * SIZE(X) movaps %xmm4, 0 * SIZE(Y) movaps %xmm6, 2 * SIZE(Y) addq $4 * SIZE, X addq $4 * SIZE, Y ALIGN_3 .L16: testq $2, N jle .L17 movaps 0 * SIZE(Y), %xmm1 movaps 0 * SIZE(X), %xmm0 movaps %xmm1, %xmm2 movaps %xmm0, %xmm3 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 addpd %xmm1, %xmm0 subpd %xmm3, %xmm2 movaps %xmm0, 0 * SIZE(X) movaps %xmm2, 0 * SIZE(Y) addq $2 * SIZE, X addq $2 * SIZE, Y ALIGN_3 .L17: testq $1, N jle .L999 movsd 0 * SIZE(Y), %xmm1 movsd 0 * SIZE(X), %xmm0 movaps %xmm1, %xmm2 movaps %xmm0, %xmm3 mulsd C, %xmm0 mulsd S, %xmm1 mulsd C, %xmm2 mulsd S, %xmm3 addsd %xmm1, %xmm0 subsd %xmm3, %xmm2 movsd %xmm0, 0 * SIZE(X) movsd %xmm2, 0 * SIZE(Y) jmp .L999 ALIGN_3 .L20: movaps -1 * SIZE(Y), %xmm1 movq N, %rax sarq $4, %rax jle .L24 ALIGN_3 .L21: #ifdef PREFETCHW PREFETCHW (PREFETCHSIZE + 0) - PREOFFSET(X) #endif movaps 1 * SIZE(Y), %xmm3 movaps 3 * SIZE(Y), %xmm8 movaps 0 * SIZE(X), %xmm0 movaps 2 * SIZE(X), %xmm2 SHUFPD_1 %xmm3, %xmm1 SHUFPD_1 %xmm8, %xmm3 movaps %xmm1, %xmm4 movaps %xmm0, %xmm5 movaps %xmm3, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 addpd %xmm3, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 0 * SIZE(X) movaps %xmm2, 2 * SIZE(X) movlpd %xmm4, 0 * SIZE(Y) movhps %xmm4, 1 * SIZE(Y) movlpd %xmm6, 2 * SIZE(Y) movhps %xmm6, 3 * SIZE(Y) #ifdef PREFETCHW PREFETCHW (PREFETCHSIZE + 0) - PREOFFSET(Y) #endif movaps 5 * SIZE(Y), %xmm9 movaps 7 * SIZE(Y), %xmm1 movaps 4 * SIZE(X), %xmm0 movaps 6 * SIZE(X), %xmm2 SHUFPD_1 %xmm9, %xmm8 SHUFPD_1 %xmm1, %xmm9 movaps %xmm8, %xmm4 movaps %xmm0, %xmm5 movaps %xmm9, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm8 mulpd C, %xmm2 mulpd S, %xmm9 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm8, %xmm0 addpd %xmm9, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 4 * SIZE(X) movaps %xmm2, 6 * SIZE(X) movlpd %xmm4, 4 * SIZE(Y) movhps %xmm4, 5 * SIZE(Y) movlpd %xmm6, 6 * SIZE(Y) movhps %xmm6, 7 * SIZE(Y) #ifdef PREFETCHW PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(X) #endif movaps 9 * SIZE(Y), %xmm3 movaps 11 * SIZE(Y), %xmm8 movaps 8 * SIZE(X), %xmm0 movaps 10 * SIZE(X), %xmm2 SHUFPD_1 %xmm3, %xmm1 SHUFPD_1 %xmm8, %xmm3 movaps %xmm1, %xmm4 movaps %xmm0, %xmm5 movaps %xmm3, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 addpd %xmm3, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 8 * SIZE(X) movaps %xmm2, 10 * SIZE(X) movlpd %xmm4, 8 * SIZE(Y) movhps %xmm4, 9 * SIZE(Y) movlpd %xmm6, 10 * SIZE(Y) movhps %xmm6, 11 * SIZE(Y) #ifdef PREFETCHW PREFETCHW (PREFETCHSIZE + 64) - PREOFFSET(Y) #endif movaps 13 * SIZE(Y), %xmm9 movaps 15 * SIZE(Y), %xmm1 movaps 12 * SIZE(X), %xmm0 movaps 14 * SIZE(X), %xmm2 SHUFPD_1 %xmm9, %xmm8 SHUFPD_1 %xmm1, %xmm9 movaps %xmm8, %xmm4 movaps %xmm0, %xmm5 movaps %xmm9, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm8 mulpd C, %xmm2 mulpd S, %xmm9 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm8, %xmm0 addpd %xmm9, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 12 * SIZE(X) movaps %xmm2, 14 * SIZE(X) movlpd %xmm4, 12 * SIZE(Y) movhps %xmm4, 13 * SIZE(Y) movlpd %xmm6, 14 * SIZE(Y) movhps %xmm6, 15 * SIZE(Y) addq $16 * SIZE, X addq $16 * SIZE, Y decq %rax jg .L21 ALIGN_3 .L24: testq $15, N jle .L999 testq $8, N jle .L25 movaps 1 * SIZE(Y), %xmm3 movaps 3 * SIZE(Y), %xmm8 movaps 0 * SIZE(X), %xmm0 movaps 2 * SIZE(X), %xmm2 SHUFPD_1 %xmm3, %xmm1 SHUFPD_1 %xmm8, %xmm3 movaps %xmm1, %xmm4 movaps %xmm0, %xmm5 movaps %xmm3, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 addpd %xmm3, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 0 * SIZE(X) movaps %xmm2, 2 * SIZE(X) movlpd %xmm4, 0 * SIZE(Y) movhps %xmm4, 1 * SIZE(Y) movlpd %xmm6, 2 * SIZE(Y) movhps %xmm6, 3 * SIZE(Y) movaps 5 * SIZE(Y), %xmm9 movaps 7 * SIZE(Y), %xmm1 movaps 4 * SIZE(X), %xmm0 movaps 6 * SIZE(X), %xmm2 SHUFPD_1 %xmm9, %xmm8 SHUFPD_1 %xmm1, %xmm9 movaps %xmm8, %xmm4 movaps %xmm0, %xmm5 movaps %xmm9, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm8 mulpd C, %xmm2 mulpd S, %xmm9 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm8, %xmm0 addpd %xmm9, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 4 * SIZE(X) movaps %xmm2, 6 * SIZE(X) movlpd %xmm4, 4 * SIZE(Y) movhps %xmm4, 5 * SIZE(Y) movlpd %xmm6, 6 * SIZE(Y) movhps %xmm6, 7 * SIZE(Y) addq $8 * SIZE, X addq $8 * SIZE, Y ALIGN_3 .L25: testq $4, N jle .L26 movaps 1 * SIZE(Y), %xmm3 movaps 3 * SIZE(Y), %xmm8 movaps 0 * SIZE(X), %xmm0 movaps 2 * SIZE(X), %xmm2 SHUFPD_1 %xmm3, %xmm1 SHUFPD_1 %xmm8, %xmm3 movaps %xmm1, %xmm4 movaps %xmm0, %xmm5 movaps %xmm3, %xmm6 movaps %xmm2, %xmm7 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 mulpd C, %xmm4 mulpd S, %xmm5 mulpd C, %xmm6 mulpd S, %xmm7 addpd %xmm1, %xmm0 addpd %xmm3, %xmm2 subpd %xmm5, %xmm4 subpd %xmm7, %xmm6 movaps %xmm0, 0 * SIZE(X) movaps %xmm2, 2 * SIZE(X) movlpd %xmm4, 0 * SIZE(Y) movhps %xmm4, 1 * SIZE(Y) movlpd %xmm6, 2 * SIZE(Y) movhps %xmm6, 3 * SIZE(Y) movaps %xmm8, %xmm1 addq $4 * SIZE, X addq $4 * SIZE, Y ALIGN_3 .L26: testq $2, N jle .L27 movaps 1 * SIZE(Y), %xmm4 movaps 0 * SIZE(X), %xmm0 SHUFPD_1 %xmm4, %xmm1 movaps %xmm1, %xmm2 movaps %xmm0, %xmm3 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 addpd %xmm1, %xmm0 subpd %xmm3, %xmm2 movaps %xmm0, 0 * SIZE(X) movlpd %xmm2, 0 * SIZE(Y) movhps %xmm2, 1 * SIZE(Y) movaps %xmm4, %xmm1 addq $2 * SIZE, X addq $2 * SIZE, Y ALIGN_3 .L27: testq $1, N jle .L999 unpckhpd %xmm1, %xmm1 movsd 0 * SIZE(X), %xmm0 movaps %xmm1, %xmm2 movaps %xmm0, %xmm3 mulsd C, %xmm0 mulsd S, %xmm1 mulsd C, %xmm2 mulsd S, %xmm3 addsd %xmm1, %xmm0 subsd %xmm3, %xmm2 movsd %xmm0, 0 * SIZE(X) movsd %xmm2, 0 * SIZE(Y) jmp .L999 ALIGN_3 .L50: movq N, %rax cmpq $0, INCX je .L56 cmpq $0, INCY je .L56 sarq $2, %rax jle .L55 ALIGN_3 .L53: movsd (Y), %xmm1 movhps (Y, INCY), %xmm1 movsd (X), %xmm0 movhps (X, INCX), %xmm0 movaps %xmm1, %xmm2 movaps %xmm0, %xmm3 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 addpd %xmm1, %xmm0 subpd %xmm3, %xmm2 movlpd %xmm0, (X) movhps %xmm0, (X, INCX) movlpd %xmm2, (Y) movhps %xmm2, (Y, INCY) leaq (X, INCX, 2), X leaq (Y, INCY, 2), Y movsd (Y), %xmm1 movhps (Y, INCY), %xmm1 movsd (X), %xmm0 movhps (X, INCX), %xmm0 movaps %xmm1, %xmm2 movaps %xmm0, %xmm3 mulpd C, %xmm0 mulpd S, %xmm1 mulpd C, %xmm2 mulpd S, %xmm3 addpd %xmm1, %xmm0 subpd %xmm3, %xmm2 movlpd %xmm0, (X) movhps %xmm0, (X, INCX) movlpd %xmm2, (Y) movhps %xmm2, (Y, INCY) leaq (X, INCX, 2), X leaq (Y, INCY, 2), Y decq %rax jg .L53 ALIGN_3 .L55: movq N, %rax andq $3, %rax jle .L999 ALIGN_3 .L56: movsd (Y), %xmm1 movsd (X), %xmm0 movaps %xmm1, %xmm2 movaps %xmm0, %xmm3 mulsd C, %xmm0 mulsd S, %xmm1 mulsd C, %xmm2 mulsd S, %xmm3 addsd %xmm1, %xmm0 subsd %xmm3, %xmm2 movsd %xmm0, (X) movsd %xmm2, (Y) addq INCX, X addq INCY, Y decq %rax jg .L56 ALIGN_3 .L999: RESTOREREGISTERS ret EPILOGUE