diff options
Diffstat (limited to 'sysdeps/x86_64/fpu/svml_d_sincos4_core.S')
-rw-r--r-- | sysdeps/x86_64/fpu/svml_d_sincos4_core.S | 125 |
1 files changed, 124 insertions, 1 deletions
diff --git a/sysdeps/x86_64/fpu/svml_d_sincos4_core.S b/sysdeps/x86_64/fpu/svml_d_sincos4_core.S index 2c0b011..088d5ad 100644 --- a/sysdeps/x86_64/fpu/svml_d_sincos4_core.S +++ b/sysdeps/x86_64/fpu/svml_d_sincos4_core.S @@ -20,8 +20,131 @@ #include "svml_d_wrapper_impl.h" .text +ENTRY (_ZGVdN4vl8l8_sincos) +WRAPPER_IMPL_AVX_fFF _ZGVbN2vl8l8_sincos +END (_ZGVdN4vl8l8_sincos) +libmvec_hidden_def (_ZGVdN4vl8l8_sincos) + +/* AVX2 ISA version as wrapper to SSE ISA version (for vector + function declared with #pragma omp declare simd notinbranch). */ +.macro WRAPPER_IMPL_AVX2_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 $-32, %rsp + subq $160, %rsp + vmovupd %ymm0, 128(%rsp) + lea (%rsp), %rdi + vmovdqu %ymm1, 64(%rdi) + vmovdqu %ymm2, 96(%rdi) + lea 32(%rsp), %rsi + vzeroupper + call HIDDEN_JUMPTARGET(\callee) + vmovupd 144(%rsp), %xmm0 + lea 16(%rsp), %rdi + lea 48(%rsp), %rsi + call HIDDEN_JUMPTARGET(\callee) + movq 64(%rsp), %rdx + movq 96(%rsp), %rsi + movq 72(%rsp), %r8 + movq 104(%rsp), %r10 + movq (%rsp), %rax + movq 32(%rsp), %rcx + movq 8(%rsp), %rdi + movq 40(%rsp), %r9 + movq %rax, (%rdx) + movq %rcx, (%rsi) + movq 80(%rsp), %rax + movq 112(%rsp), %rcx + movq %rdi, (%r8) + movq %r9, (%r10) + movq 88(%rsp), %rdi + movq 120(%rsp), %r9 + movq 16(%rsp), %r11 + movq 48(%rsp), %rdx + movq 24(%rsp), %rsi + movq 56(%rsp), %r8 + movq %r11, (%rax) + movq %rdx, (%rcx) + movq %rsi, (%rdi) + movq %r8, (%r9) + 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 $-32, %esp + pushq -8(%r10d) + pushq %rbp + .cfi_escape 0x10,0x6,0x2,0x76,0 + movl %esp, %ebp + pushq %r12 + leal -80(%rbp), %esi + pushq %r10 + .cfi_escape 0xf,0x3,0x76,0x70,0x6 + .cfi_escape 0x10,0xc,0x2,0x76,0x78 + leal -112(%rbp), %edi + movq %rsi, %r12 + pushq %rbx + .cfi_escape 0x10,0x3,0x2,0x76,0x68 + movq %rdi, %rbx + subl $152, %esp + vmovaps %xmm1, -128(%ebp) + vmovaps %xmm2, -144(%ebp) + vmovapd %ymm0, -176(%ebp) + vzeroupper + call HIDDEN_JUMPTARGET(\callee) + leal 16(%r12), %esi + vmovapd -160(%ebp), %xmm0 + leal 16(%rbx), %edi + call HIDDEN_JUMPTARGET(\callee) + movq -128(%ebp), %rax + vmovsd -112(%ebp), %xmm0 + vmovdqa -128(%ebp), %xmm5 + vmovdqa -144(%ebp), %xmm1 + vmovsd %xmm0, (%eax) + vmovsd -104(%ebp), %xmm0 + vpextrd $1, %xmm5, %eax + vmovsd %xmm0, (%eax) + movq -120(%ebp), %rax + vmovsd -96(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + vmovsd -88(%ebp), %xmm0 + vpextrd $3, %xmm5, %eax + vmovsd %xmm0, (%eax) + movq -144(%ebp), %rax + vmovsd -80(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + vmovsd -72(%ebp), %xmm0 + vpextrd $1, %xmm1, %eax + vmovsd %xmm0, (%eax) + movq -136(%ebp), %rax + vmovsd -64(%ebp), %xmm0 + vmovsd %xmm0, (%eax) + vmovsd -56(%ebp), %xmm0 + vpextrd $3, %xmm1, %eax + vmovsd %xmm0, (%eax) + addl $152, %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 (_ZGVdN4vvv_sincos) -WRAPPER_IMPL_AVX_fFF _ZGVbN2vvv_sincos +WRAPPER_IMPL_AVX2_fFF_vvv _ZGVbN2vl8l8_sincos END (_ZGVdN4vvv_sincos) #ifndef USE_MULTIARCH |