[PD-cvs] pd/src m_simd_sse_gcc.c,1.1.2.6,1.1.2.7 m_simd_sse_gcc.h,1.1.2.10,1.1.2.11

Update of /cvsroot/pure-data/pd/src In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv14242

Modified Files: Tag: devel_0_37 m_simd_sse_gcc.c m_simd_sse_gcc.h Log Message: further update for sse / gcc

Index: m_simd_sse_gcc.c =================================================================== RCS file: /cvsroot/pure-data/pd/src/Attic/m_simd_sse_gcc.c,v retrieving revision 1.1.2.6 retrieving revision 1.1.2.7 diff -C2 -d -r1.1.2.6 -r1.1.2.7 *** m_simd_sse_gcc.c 9 Oct 2004 07:44:29 -0000 1.1.2.6 --- m_simd_sse_gcc.c 9 Oct 2004 11:56:16 -0000 1.1.2.7 *************** *** 13,17 ****

/* dst is assumed to be aligned */ ! static void zerovec_sse_gcc(t_float *dst,int n) { asm( --- 13,17 ----

/* dst is assumed to be aligned */ ! void zerovec_sse_gcc(t_float *dst,int n) { asm( *************** *** 21,25 ****

/* should we do more loop unrolling? */ ! "1: \n" "movaps %%xmm0, (%1) \n" "movaps %%xmm0, 4*T_FLOAT(%1) \n" --- 21,26 ----

/* should we do more loop unrolling? */ ! /* *dst = 0 */ ! "1: \n" "movaps %%xmm0, (%1) \n" "movaps %%xmm0, 4*T_FLOAT(%1) \n" *************** *** 27,31 **** "movaps %%xmm0, 12*T_FLOAT(%1) \n"

! "addl $64,%1 \n" "loop 1b \n" : --- 28,32 ---- "movaps %%xmm0, 12*T_FLOAT(%1) \n"

! "addl $16*T_FLOAT,%1 \n" "loop 1b \n" : *************** *** 43,56 ****

/* should we do more loop unrolling? */ "1: \n" ! "movaps %2, (%1) \n" ! "movaps %2, 4*T_FLOAT(%1) \n" ! "movaps %2, 8*T_FLOAT(%1) \n" ! "movaps %2, 12*T_FLOAT(%1) \n"

! "addl $64,%1 \n" "loop 1b \n" : ! :"c"(n),"r"(dst),"x"(v) ); } --- 44,58 ----

/* should we do more loop unrolling? */ + /* *dst = v */ "1: \n" ! "movaps %2, (%1) \n" ! "movaps %2, 4*T_FLOAT(%1) \n" ! "movaps %2, 8*T_FLOAT(%1) \n" ! "movaps %2, 12*T_FLOAT(%1) \n"

! "addl $16*T_FLOAT,%1 \n" "loop 1b \n" : ! :"c"(n),"r"(dst),"x"((t_float)v) ); } *************** *** 64,84 **** "shr $4, %0 \n"

! /* loop: *out = *in */ "1: \n" "movaps (%1), %%xmm0 \n" "movaps 4*T_FLOAT(%1), %%xmm1 \n" "movaps 8*T_FLOAT(%1), %%xmm2 \n" ! "movaps 16*T_FLOAT(%1), %%xmm3 \n" "movaps %%xmm0, (%2) \n" "movaps %%xmm1, 4*T_FLOAT(%2) \n" "movaps %%xmm2, 8*T_FLOAT(%2) \n" ! "movaps %%xmm3, 16*T_FLOAT(%2) \n"

! "addl $64,%1 \n" ! "addl $64,%2 \n" "loop 1b \n" : ! :"r"(n),"r"(src),"r"(dst) ! :"%xmm0","%xmm1","%xmm2","%xmm3","%eax"); }

--- 66,123 ---- "shr $4, %0 \n"

! /* loop: *dst = *src */ "1: \n" "movaps (%1), %%xmm0 \n" "movaps 4*T_FLOAT(%1), %%xmm1 \n" "movaps 8*T_FLOAT(%1), %%xmm2 \n" ! "movaps 12*T_FLOAT(%1), %%xmm3 \n" "movaps %%xmm0, (%2) \n" "movaps %%xmm1, 4*T_FLOAT(%2) \n" "movaps %%xmm2, 8*T_FLOAT(%2) \n" ! "movaps %%xmm3, 12*T_FLOAT(%2) \n"

! "addl $16*T_FLOAT,%1 \n" ! "addl $16*T_FLOAT,%2 \n" "loop 1b \n" : ! :"c"(n),"r"(src),"r"(dst) ! :"%xmm0","%xmm1","%xmm2","%xmm3"); ! } ! ! ! /* dst and src are assumed to be aligned */ ! void addvec_sse_gcc(t_float *dst,const t_float *src,int n) ! { ! asm( ! ".set T_FLOAT,4 \n" /* sizeof(t_float) */ ! "shr $4, %0 \n" ! ! /* loop: *dst += *src */ ! "1: \n" ! "movaps (%2,%3),%%xmm0 \n" ! "movaps (%1,%3),%%xmm1 \n" ! "addps %%xmm0,%%xmm1 \n" ! "movaps %%xmm0,(%2,%3) \n" ! ! "movaps 4*T_FLOAT(%2,%3),%%xmm0 \n" ! "movaps 4*T_FLOAT(%1,%3),%%xmm1 \n" ! "addps %%xmm0,%%xmm1 \n" ! "movaps %%xmm0,4*T_FLOAT(%2,%3) \n" ! ! "movaps 8*T_FLOAT(%2,%3),%%xmm0 \n" ! "movaps 8*T_FLOAT(%1,%3),%%xmm1 \n" ! "addps %%xmm0,%%xmm1 \n" ! "movaps %%xmm0,8*T_FLOAT(%2,%3) \n" ! ! "movaps 12*T_FLOAT(%2,%3),%%xmm0 \n" ! "movaps 12*T_FLOAT(%1,%3),%%xmm1 \n" ! "addps %%xmm0,%%xmm1 \n" ! "movaps %%xmm0,12*T_FLOAT(%2,%3) \n" ! ! "addl $16*T_FLOAT,%3 \n" ! "loop 1b \n" ! : ! : "c"(n),"r"(src),"r"(dst),"r"(0) ! : "%xmm0","%xmm1","%xmm2","%xmm3","%xmm4","%xmm5","%xmm6","%xmm7"); }

*************** *** 95,157 **** }

! /* t_int *sig_tilde_perf_sse_gcc(t_int *w) */ ! /* { */ ! /* setvec_sse_gcc((t_float *)w[2],*(const t_float *)w[1],w[3]); */ ! /* return w+4; */ ! /* } */ ! ! /* zero_perf_sse_gcc (t_int * w)*/ ! /* asm( */ ! /* ".set T_FLOAT,4 \n" /* sizeof(t_float) */ */ ! /* ".set T_INT,4 \n" /* sizeof(t_int) */ */ ! ! /* /* header */ */ ! /* ".text \n" */ ! /* ".align 4 \n" /* alignment */ */

- /* ".globl zero_perf_sse_gcc \n" */ - /* "zero_perf_sse_gcc: \n" */ - - /* /* head of function */ */ - /* "pushl %ebp \n" */ - /* "movl %esp, %ebp \n" */ - /* "pushl %esi \n" */ - /* "movl %esi, -4(%ebp) \n" */

! /* /* get arguments */ */ ! /* "movl 8(%ebp), %esi \n" /* argument pointer */ */ ! /* "movl T_INT(%esi), %edx \n" /* out */ */ ! /* "movl 2*T_INT(%esi), %ecx \n" /* n */ */ ! ! /* "xorps %xmm0, %xmm0 \n" /* load zero */ */ ! /* "shrl $4, %ecx \n" /* divide by 16 */ */ ! ! /* /* loop: *out=0 */ */ ! /* "zpsg_loop: \n" */ ! /* "movaps %xmm0, (%edx) \n" */ ! /* "movaps %xmm0, 4*T_FLOAT(%edx) \n" */ ! /* "movaps %xmm0, 8*T_FLOAT(%edx) \n" */ ! /* "movaps %xmm0, 12*T_FLOAT(%edx) \n" */ ! ! /* "addl $64, %edx \n" /* out+=16 */ */ ! /* "loop zpsg_loop \n" */ !

! /* /* return w+3; */ */ ! /* "movl -4(%ebp), %esi \n" */ ! /* "movl 8(%ebp), %eax \n" */ ! /* "addl $12, %eax \n" */ ! /* "movl %ebp, %esp \n" */ ! /* "popl %ebp \n" */ ! /* "ret \n" */

! /* ".align 4 \n" /* alignment */ */ ! /* ".type zero_perf_sse_gcc, @function \n" */

! /* ); */

! /* /* copy_perf_sse_gcc (t_int * w)*/ */ /* asm( */ /* ".set T_FLOAT,4 \n" */ --- 134,222 ---- }

! t_int *sig_tilde_perf_sse_gcc(t_int *w) ! { ! setvec_sse_gcc((t_float *)w[2],*(const t_float *)w[1],w[3]); ! return w+4; ! }

! t_int * plus_perf_sse_gcc (t_int * w) ! { ! asm( ! ".set T_FLOAT,4 \n" ! ! "shrl $4, %3 \n" /* divide by 16 */

! /* loop: *out = *in1 + *in2 */ ! "1: \n" ! "movaps (%0,%4), %%xmm0 \n" ! "movaps (%1,%4), %%xmm1 \n" ! "addps %%xmm1, %%xmm0 \n" ! "movaps %%xmm0, (%2,%4) \n" ! ! "movaps 4*T_FLOAT(%0,%4), %%xmm2 \n" ! "movaps 4*T_FLOAT(%1,%4), %%xmm3 \n" ! "addps %%xmm3, %%xmm2 \n" ! "movaps %%xmm2, 4*T_FLOAT(%2,%4) \n" ! ! "movaps 8*T_FLOAT(%0,%4), %%xmm4 \n" ! "movaps 8*T_FLOAT(%1,%4), %%xmm5 \n" ! "addps %%xmm5, %%xmm4 \n" ! "movaps %%xmm4, 8*T_FLOAT(%2,%4) \n" ! ! "movaps 12*T_FLOAT(%0,%4), %%xmm6 \n" ! "movaps 12*T_FLOAT(%1,%4), %%xmm7 \n" ! "addps %%xmm7, %%xmm6 \n" ! "movaps %%xmm6, 12*T_FLOAT(%2,%4) \n" ! ! "addl $16*T_FLOAT, %4 \n" ! "loop 1b \n" ! : ! /* in1, in2, out, n */ ! :"r"(w[1]),"r"(w[2]),"r"(w[3]),"c"(w[4]),"r"(0) ! ); ! return w+5; ! }

! t_int *scalarplus_perf_sse_gcc(t_int *w) ! { ! asm( ! ".set T_FLOAT,4 \n" ! ! "shufps $0, %1, %1 \n" ! "shrl $4, %3 \n" /* divide by 16 */

! /* loop: *out = *in + value */ ! "1: \n" ! "movaps (%0), %%xmm1 \n" ! "addps %1, %%xmm1 \n" ! "movaps %%xmm1, (%2) \n" ! ! "movaps 4*T_FLOAT(%0), %%xmm2 \n" ! "addps %1, %%xmm2 \n" ! "movaps %%xmm2, 4*T_FLOAT(%2) \n" ! ! "movaps 8*T_FLOAT(%0), %%xmm3 \n" ! "addps %1, %%xmm3 \n" ! "movaps %%xmm3, 8*T_FLOAT(%2) \n" ! ! "movaps 12*T_FLOAT(%0), %%xmm4 \n" ! "addps %1, %%xmm4 \n" ! "movaps %%xmm4, 12*T_FLOAT(%2) \n" ! ! "addl $16*T_FLOAT, %0 \n" ! "addl $16*T_FLOAT, %2 \n" ! "loop 1b \n" ! : ! /* in, value, out, n */ ! :"r"(w[1]),"x"((t_float)w[2]),"r"(w[3]),"c"(w[4]) ! :"%xmm1","%xmm2","%xmm3","%xmm4" ! ); ! return w+5; ! }

! /* /* scalarplus_perf_sse_gcc(t_int *w) */ */ /* asm( */ /* ".set T_FLOAT,4 \n" */ *************** *** 162,167 **** /* ".align 4 \n" /* alignment */ */

! /* ".globl copy_perf_sse_gcc \n" */ ! /* "copy_perf_sse_gcc: \n" */

/* /* head of function */ */ --- 227,232 ---- /* ".align 4 \n" /* alignment */ */

! /* ".globl scalarplus_perf_sse_gcc \n" */ ! /* "scalarplus_perf_sse_gcc: \n" */

/* /* head of function */ */ *************** *** 174,402 **** /* /* get arguments */ */ /* "movl 8(%ebp), %esi \n" /* argument pointer */ */ ! /* "movl 1*T_INT(%esi), %ebx \n" /* in1 */ */ ! /* "movl 2*T_INT(%esi), %edx \n" /* out */ */ ! /* "movl 3*T_INT(%esi), %ecx \n" /* n */ */ /* "shrl $4, %ecx \n" /* divide by 16 */ */

! /* /* loop: *out = *in */ */ ! /* "cpsg_loop: \n" */ ! /* "movaps (%ebx), %xmm0 \n" */ ! /* "movaps %xmm0, (%edx) \n" */ ! /* "movaps 4*T_FLOAT(%ebx), %xmm1 \n" */ ! /* "movaps %xmm1, 4*T_FLOAT(%edx) \n" */

! /* "movaps 8*T_FLOAT(%ebx), %xmm2 \n" */ ! /* "movaps %xmm2, 8*T_FLOAT(%edx) \n" */ ! /* "movaps 12*T_FLOAT(%ebx), %xmm3 \n" */ ! /* "movaps %xmm3, 12*T_FLOAT(%edx) \n" */ ! ! /* "addl $64, %ebx \n" /* in1 +=16 */ */ ! /* "addl $64, %edx \n" /* out +=16 */ */ ! /* "loop cpsg_loop \n" */

! /* /* return w+4; */ */ /* "movl -4(%ebp), %ebx \n" */ /* "movl -8(%ebp), %esi \n" */ /* "movl 8(%ebp), %eax \n" */ ! /* "addl $16, %eax \n" */ /* "movl %ebp, %esp \n" */ /* "popl %ebp \n" */ /* "ret \n" */

- /* ".align 4 \n" /* alignment */ */ ! /* ".type copy_perf_sse_gcc, @function \n" */ ! /* ); */ ! ! ! /* sig_tilde_perf_sse_gcc(t_int * w) */ ! asm( ! ".set T_FLOAT,4 \n" ! ".set T_INT,4 \n" ! ! /* header */ ! ".text \n" ! ".align 4 \n" /* alignment */ ! ! ".globl sig_tilde_perf_sse_gcc \n" ! "sig_tilde_perf_sse_gcc: \n" ! ! /* head of function */ ! "pushl %ebp \n" ! "movl %esp, %ebp \n" ! "pushl %esi \n" ! "movl %esi, -4(%ebp) \n" ! ! /* get arguments */ ! "movl 8(%ebp), %esi \n" /* argument pointer */ ! "movl 2*T_INT(%esi), %edx \n" /* out */ ! "movl 1*T_INT(%esi), %eax \n" /* f */ ! "movl 3*T_INT(%esi), %ecx \n" /* n */ ! ! /* set registers to f */ ! "movss (%eax), %xmm0 \n" ! "shufps $0, %xmm0, %xmm0 \n" ! "movl 12(%esi), %ecx \n" ! "shrl $4, %ecx \n" /* divide by 16 */ ! ! /* loop: *out = f */ ! "sigtpsg_loop: \n" ! "movaps %xmm0, (%edx) \n" ! "movaps %xmm0, 16(%edx) \n" ! "movaps %xmm0, 32(%edx) \n" ! "movaps %xmm0, 48(%edx) \n" ! "addl $64, %edx \n" /* out+=16 */ ! "loop sigtpsg_loop \n" ! ! /* return w+4; */ ! "movl -4(%ebp), %esi \n" ! "movl 8(%ebp), %eax \n" ! "addl $16, %eax \n" ! "movl %ebp, %esp \n" ! "popl %ebp \n" ! "ret \n" ! ! ".align 4 \n" /* alignment */ ! ".type sig_tilde_perf_sse_gcc, @function \n" ! ! ); ! ! ! ! /* plus_perf_sse_gcc (t_int * w)*/ ! asm( ! ".set T_FLOAT,4 \n" ! ".set T_INT,4 \n" ! ! /* header */ ! ".text \n" ! ".align 4 \n" /* alignment */ ! ! ".globl plus_perf_sse_gcc \n" ! "plus_perf_sse_gcc: \n" ! ! /* head of function */ ! "pushl %ebp \n" ! "movl %esp, %ebp \n" ! "subl $8, %esp \n" ! "movl %ebx, -4(%ebp) \n" ! "movl %esi, -8(%ebp) \n" ! ! /* get arguments */ ! "movl 8(%ebp), %esi \n" /* argument vector */ ! ! "movl 1*T_INT(%esi), %eax \n" /* in1 */ ! "movl 2*T_INT(%esi), %ebx \n" /* in2 */ ! "movl 3*T_INT(%esi), %edx \n" /* out */ ! "movl 4*T_INT(%esi), %ecx \n" /* n */ ! "shrl $4, %ecx \n" /* divide by 16 */ ! "xorl %esi, %esi \n" /* reset index */ ! ! /* loop: *out = *in1 + *in2 */ ! "ppsg_loop: \n" ! "movaps (%eax,%esi), %xmm0 \n" ! "movaps (%ebx,%esi), %xmm1 \n" ! "addps %xmm1, %xmm0 \n" ! "movaps %xmm0, (%edx,%esi) \n" ! ! "movaps 4*T_FLOAT(%eax,%esi), %xmm2 \n" ! "movaps 4*T_FLOAT(%ebx,%esi), %xmm3 \n" ! "addps %xmm3, %xmm2 \n" ! "movaps %xmm2, 4*T_FLOAT(%edx,%esi) \n" ! ! "movaps 8*T_FLOAT(%eax,%esi), %xmm4 \n" ! "movaps 8*T_FLOAT(%ebx,%esi), %xmm5 \n" ! "addps %xmm5, %xmm4 \n" ! "movaps %xmm4, 8*T_FLOAT(%edx,%esi) \n" ! ! "movaps 12*T_FLOAT(%eax,%esi), %xmm6 \n" ! "movaps 12*T_FLOAT(%ebx,%esi), %xmm7 \n" ! "addps %xmm7, %xmm6 \n" ! "movaps %xmm6, 12*T_FLOAT(%edx,%esi) \n" ! ! "addl $64, %esi \n" /* out+=16; */ ! "loop ppsg_loop \n" ! ! /* return w+5; */ ! "movl -4(%ebp), %ebx \n" ! "movl -8(%ebp), %esi \n" ! "movl 8(%ebp), %eax \n" ! "addl $20, %eax \n" ! "movl %ebp, %esp \n" ! "popl %ebp \n" ! "ret \n" ! ! ! ".align 4 \n" /* alignment */ ! ".type plus_perf_sse_gcc, @function \n" ! ! ); ! ! /* scalarplus_perf_sse_gcc(t_int *w) */ ! asm( ! ".set T_FLOAT,4 \n" ! ".set T_INT,4 \n" ! ! /* header */ ! ".text \n" ! ".align 4 \n" /* alignment */ ! ! ".globl scalarplus_perf_sse_gcc \n" ! "scalarplus_perf_sse_gcc: \n" ! ! /* head of function */ ! "pushl %ebp \n" ! "movl %esp, %ebp \n" ! "subl $8, %esp \n" ! "movl %ebx, -4(%ebp) \n" ! "movl %esi, -8(%ebp) \n" ! ! /* get arguments */ ! "movl 8(%ebp), %esi \n" /* argument pointer */ ! "movl T_INT(%esi), %ebx \n" /* in */ ! "movl 3*T_INT(%esi), %edx \n" /* out */ ! "movl 2*T_INT(%esi), %eax \n" /* value */ ! "movl 4*T_INT(%esi), %ecx \n" /* n */ ! ! "movss (%eax), %xmm0 \n" ! "shufps $0, %xmm0, %xmm0 \n" ! "shrl $4, %ecx \n" /* divide by 16 */ ! ! /* loop: *out = *in + value */ ! "sppsg_loop: \n" ! ! "movaps (%ebx), %xmm1 \n" ! "addps %xmm0, %xmm1 \n" ! "movaps %xmm1, (%edx) \n" ! ! "movaps 4*T_FLOAT(%ebx), %xmm2 \n" ! "addps %xmm0, %xmm2 \n" ! "movaps %xmm2, 4*T_FLOAT(%edx) \n" ! ! "movaps 8*T_FLOAT(%ebx), %xmm3 \n" ! "addps %xmm0, %xmm3 \n" ! "movaps %xmm3, 8*T_FLOAT(%edx) \n" ! ! "movaps 12*T_FLOAT(%ebx), %xmm4 \n" ! "addps %xmm0, %xmm4 \n" ! "movaps %xmm4, 12*T_FLOAT(%edx) \n" ! ! "addl $64, %ebx \n" /* in += 16 */ ! "addl $64, %edx \n" /* out += 16 */ ! "loop sppsg_loop \n" ! ! /* return w+5; */ ! "movl -4(%ebp), %ebx \n" ! "movl -8(%ebp), %esi \n" ! "movl 8(%ebp), %eax \n" ! "addl $20, %eax \n" ! "movl %ebp, %esp \n" ! "popl %ebp \n" ! "ret \n" ! ! ".align 4 \n" /* alignment */ ! ".type scalarplus_perf_sse_gcc, @function \n"

! );

--- 239,287 ---- /* /* get arguments */ */ /* "movl 8(%ebp), %esi \n" /* argument pointer */ */ ! /* "movl T_INT(%esi), %ebx \n" /* in */ */ ! /* "movl 3*T_INT(%esi), %edx \n" /* out */ */ ! /* "movl 2*T_INT(%esi), %eax \n" /* value */ */ ! /* "movl 4*T_INT(%esi), %ecx \n" /* n */ */ ! ! /* "movss (%eax), %xmm0 \n" */ ! /* "shufps $0, %xmm0, %xmm0 \n" */ /* "shrl $4, %ecx \n" /* divide by 16 */ */

! /* /* loop: *out = *in + value */ */ ! /* "sppsg_loop: \n" */

! /* "movaps (%ebx), %xmm1 \n" */ ! /* "addps %xmm0, %xmm1 \n" */ ! /* "movaps %xmm1, (%edx) \n" */

! /* "movaps 4*T_FLOAT(%ebx), %xmm2 \n" */ ! /* "addps %xmm0, %xmm2 \n" */ ! /* "movaps %xmm2, 4*T_FLOAT(%edx) \n" */ ! ! /* "movaps 8*T_FLOAT(%ebx), %xmm3 \n" */ ! /* "addps %xmm0, %xmm3 \n" */ ! /* "movaps %xmm3, 8*T_FLOAT(%edx) \n" */ ! ! /* "movaps 12*T_FLOAT(%ebx), %xmm4 \n" */ ! /* "addps %xmm0, %xmm4 \n" */ ! /* "movaps %xmm4, 12*T_FLOAT(%edx) \n" */ ! ! /* "addl $64, %ebx \n" /* in += 16 */ */ ! /* "addl $64, %edx \n" /* out += 16 */ */ ! /* "loop sppsg_loop \n" */ ! ! /* /* return w+5; */ */ /* "movl -4(%ebp), %ebx \n" */ /* "movl -8(%ebp), %esi \n" */ /* "movl 8(%ebp), %eax \n" */ ! /* "addl $20, %eax \n" */ /* "movl %ebp, %esp \n" */ /* "popl %ebp \n" */ /* "ret \n" */

/* ".align 4 \n" /* alignment */ */ ! /* ".type scalarplus_perf_sse_gcc, @function \n" */

! /* ); */

Index: m_simd_sse_gcc.h =================================================================== RCS file: /cvsroot/pure-data/pd/src/Attic/m_simd_sse_gcc.h,v retrieving revision 1.1.2.10 retrieving revision 1.1.2.11 diff -C2 -d -r1.1.2.10 -r1.1.2.11 *** m_simd_sse_gcc.h 9 Oct 2004 07:44:29 -0000 1.1.2.10 --- m_simd_sse_gcc.h 9 Oct 2004 11:56:16 -0000 1.1.2.11 *************** *** 10,14 ****

/* SIMD functions for SSE with GCC */ ! //void zerovec_sse_gcc(t_float *dst,int n); void setvec_sse_gcc(t_float *dst,t_float v,int n); void copyvec_sse_gcc(t_float *dst,const t_float *src,int n); --- 10,14 ----

/* SIMD functions for SSE with GCC */ ! void zerovec_sse_gcc(t_float *dst,int n); void setvec_sse_gcc(t_float *dst,t_float v,int n); void copyvec_sse_gcc(t_float *dst,const t_float *src,int n); *************** *** 39,46 ****

! #define zerovec zerovec_8 //sse_gcc #define setvec setvec_sse_gcc #define copyvec copyvec_sse_gcc ! #define addvec addvec_8 /* SIMD not implemented */ #define testcopyvec testcopyvec_8 /* SIMD not implemented */ #define testaddvec testaddvec_8 /* SIMD not implemented */ --- 39,46 ----

! #define zerovec zerovec_sse_gcc #define setvec setvec_sse_gcc #define copyvec copyvec_sse_gcc ! #define addvec addvec_sse_gcc #define testcopyvec testcopyvec_8 /* SIMD not implemented */ #define testaddvec testaddvec_8 /* SIMD not implemented */