diff options
Diffstat (limited to 'sysdeps/x86_64/fpu/svml_d_sincos8_core.S')
-rw-r--r-- | sysdeps/x86_64/fpu/svml_d_sincos8_core.S | 201 |
1 files changed, 200 insertions, 1 deletions
diff --git a/sysdeps/x86_64/fpu/svml_d_sincos8_core.S b/sysdeps/x86_64/fpu/svml_d_sincos8_core.S index 68d490e5bc..7f51ed5c8c 100644 --- a/sysdeps/x86_64/fpu/svml_d_sincos8_core.S +++ b/sysdeps/x86_64/fpu/svml_d_sincos8_core.S @@ -20,6 +20,205 @@ #include "svml_d_wrapper_impl.h" .text +ENTRY (_ZGVeN8vl8l8_sincos) +WRAPPER_IMPL_AVX512_fFF _ZGVdN4vl8l8_sincos +END (_ZGVeN8vl8l8_sincos) + +/* AVX512 ISA version as wrapper to AVX2 ISA version (for vector + function declared with #pragma omp declare simd notinbranch). */ +.macro WRAPPER_IMPL_AVX512_fFF_vvv callee +#ifndef __ILP32__ + pushq %rbp + cfi_adjust_cfa_offset (8) + cfi_rel_offset (%rbp, 0) + movq %rsp, %rbp + cfi_def_cfa_register (%rbp) + andq $-64, %rsp + subq $320, %rsp + /* Encoding for vmovups %zmm0, 256(%rsp). */ + .byte 0x62 + .byte 0xf1 + .byte 0x7c + .byte 0x48 + .byte 0x11 + .byte 0x44 + .byte 0x24 + .byte 0x04 + lea (%rsp), %rdi + /* Encoding for vmovups %zmm1, 128(%rdi). */ + .byte 0x62 + .byte 0xf1 + .byte 0x7c + .byte 0x48 + .byte 0x11 + .byte 0x4f + .byte 0x02 + /* Encoding for vmovups %zmm2, 192(%rdi). */ + .byte 0x62 + .byte 0xf1 + .byte 0x7c + .byte 0x48 + .byte 0x11 + .byte 0x57 + .byte 0x03 + lea 64(%rsp), %rsi + call HIDDEN_JUMPTARGET(\callee) + vmovdqu 288(%rsp), %ymm0 + lea 32(%rsp), %rdi + lea 96(%rsp), %rsi + call HIDDEN_JUMPTARGET(\callee) + movq 128(%rsp), %rdx + movq 192(%rsp), %rsi + movq 136(%rsp), %r8 + movq 200(%rsp), %r10 + movq (%rsp), %rax + movq 64(%rsp), %rcx + movq 8(%rsp), %rdi + movq 72(%rsp), %r9 + movq %rax, (%rdx) + movq %rcx, (%rsi) + movq 144(%rsp), %rax + movq 208(%rsp), %rcx + movq %rdi, (%r8) + movq %r9, (%r10) + movq 152(%rsp), %rdi + movq 216(%rsp), %r9 + movq 16(%rsp), %r11 + movq 80(%rsp), %rdx + movq 24(%rsp), %rsi + movq 88(%rsp), %r8 + movq %r11, (%rax) + movq %rdx, (%rcx) + movq 160(%rsp), %r11 + movq 224(%rsp), %rdx + movq %rsi, (%rdi) + movq %r8, (%r9) + movq 168(%rsp), %rsi + movq 232(%rsp), %r8 + movq 32(%rsp), %r10 + movq 96(%rsp), %rax + movq 40(%rsp), %rcx + movq 104(%rsp), %rdi + movq %r10, (%r11) + movq %rax, (%rdx) + movq 176(%rsp), %r10 + movq 240(%rsp), %rax + movq %rcx, (%rsi) + movq %rdi, (%r8) + movq 184(%rsp), %rcx + movq 248(%rsp), %rdi + movq 48(%rsp), %r9 + movq 112(%rsp), %r11 + movq 56(%rsp), %rdx + movq 120(%rsp), %rsi + movq %r9, (%r10) + movq %r11, (%rax) + movq %rdx, (%rcx) + movq %rsi, (%rdi) + movq %rbp, %rsp + cfi_def_cfa_register (%rsp) + popq %rbp + cfi_adjust_cfa_offset (-8) + cfi_restore (%rbp) + ret +#else + leal 8(%rsp), %r10d + .cfi_def_cfa 10, 0 + andl $-64, %esp + pushq -8(%r10d) + pushq %rbp + .cfi_escape 0x10,0x6,0x2,0x76,0 + movl %esp, %ebp + pushq %r12 + leal -112(%rbp), %esi + pushq %r10 + .cfi_escape 0xf,0x3,0x76,0x70,0x6 + .cfi_escape 0x10,0xc,0x2,0x76,0x78 + leal -176(%rbp), %edi + movq %rsi, %r12 + pushq %rbx + .cfi_escape 0x10,0x3,0x2,0x76,0x68 + movq %rdi, %rbx + subl $280, %esp + vmovdqa %ymm1, -208(%ebp) + vmovdqa %ymm2, -240(%ebp) + /* Encoding for vmovapd %zmm0, -304(%ebp). */ + .byte 0x67 + .byte 0x62 + .byte 0xf1 + .byte 0xfd + .byte 0x48 + .byte 0x29 + .byte 0x85 + .byte 0xd0 + .byte 0xfe + .byte 0xff + .byte 0xff + call HIDDEN_JUMPTARGET(\callee) + leal 32(%r12), %esi + vmovupd -272(%ebp), %ymm0 + leal 32(%rbx), %edi + call HIDDEN_JUMPTARGET(\callee) + movl -208(%ebp), %eax + vmovsd -176(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -204(%ebp), %eax + vmovsd -168(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -200(%ebp), %eax + vmovsd -160(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -196(%ebp), %eax + vmovsd -152(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -192(%ebp), %eax + vmovsd -144(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -188(%ebp), %eax + vmovsd -136(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -184(%ebp), %eax + vmovsd -128(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -180(%ebp), %eax + vmovsd -120(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -240(%ebp), %eax + vmovsd -112(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -236(%ebp), %eax + vmovsd -104(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -232(%ebp), %eax + vmovsd -96(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -228(%ebp), %eax + vmovsd -88(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -224(%ebp), %eax + vmovsd -80(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -220(%ebp), %eax + vmovsd -72(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -216(%ebp), %eax + vmovsd -64(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + movl -212(%ebp), %eax + vmovsd -56(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + addl $280, %esp + popq %rbx + popq %r10 + .cfi_def_cfa 10, 0 + popq %r12 + popq %rbp + leal -8(%r10), %esp + .cfi_def_cfa 7, 8 + ret +#endif +.endm + ENTRY (_ZGVeN8vvv_sincos) -WRAPPER_IMPL_AVX512_fFF _ZGVdN4vvv_sincos +WRAPPER_IMPL_AVX512_fFF_vvv _ZGVdN4vl8l8_sincos END (_ZGVeN8vvv_sincos) |