diff options
Diffstat (limited to 'sysdeps/x86_64/fpu/multiarch')
-rw-r--r-- | sysdeps/x86_64/fpu/multiarch/svml_d_sincos8_core_avx512.S | 19 | ||||
-rw-r--r-- | sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core_avx512.S | 63 |
2 files changed, 8 insertions, 74 deletions
diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_sincos8_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_d_sincos8_core_avx512.S index c9207558c5..3667faa0fa 100644 --- a/sysdeps/x86_64/fpu/multiarch/svml_d_sincos8_core_avx512.S +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_sincos8_core_avx512.S @@ -599,24 +599,9 @@ libmvec_hidden_def(_ZGVeN8vl8l8_sincos_skx) cfi_def_cfa_register (%rbp) andq $-64, %rsp subq $256, %rsp - /* Encoding for vmovups %zmm1, 128(%rsp). */ - .byte 0x62 - .byte 0xf1 - .byte 0x7c - .byte 0x48 - .byte 0x11 - .byte 0x4c - .byte 0x24 - .byte 0x02 + vmovups %zmm1, 128(%rsp) lea (%rsp), %rdi - /* Encoding for vmovups %zmm2, 192(%rdi). */ - .byte 0x62 - .byte 0xf1 - .byte 0x7c - .byte 0x48 - .byte 0x11 - .byte 0x57 - .byte 0x03 + vmovups %zmm2, 192(%rdi) lea 64(%rsp), %rsi call HIDDEN_JUMPTARGET(\callee) movq 128(%rsp), %rdx diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core_avx512.S index f73ab7de7c..8fa4255d6d 100644 --- a/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core_avx512.S +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_sincosf16_core_avx512.S @@ -510,40 +510,11 @@ libmvec_hidden_def(_ZGVeN16vl4l4_sincosf_skx) cfi_def_cfa_register (%rbp) andq $-64, %rsp subq $384, %rsp - /* Encoding for vmovups %zmm1, 128(%rsp). */ - .byte 0x62 - .byte 0xf1 - .byte 0x7c - .byte 0x48 - .byte 0x11 - .byte 0x4c - .byte 0x24 - .byte 0x02 + vmovups %zmm1, 128(%rsp) lea (%rsp), %rdi - /* Encoding for vmovups %zmm2, 192(%rdi). */ - .byte 0x62 - .byte 0xf1 - .byte 0x7c - .byte 0x48 - .byte 0x11 - .byte 0x57 - .byte 0x03 - /* Encoding for vmovups %zmm3, 256(%rdi). */ - .byte 0x62 - .byte 0xf1 - .byte 0x7c - .byte 0x48 - .byte 0x11 - .byte 0x5f - .byte 0x04 - /* Encoding for vmovups %zmm4, 320(%rdi). */ - .byte 0x62 - .byte 0xf1 - .byte 0x7c - .byte 0x48 - .byte 0x11 - .byte 0x67 - .byte 0x05 + vmovups %zmm2, 192(%rdi) + vmovups %zmm3, 256(%rdi) + vmovups %zmm4, 320(%rdi) lea 64(%rsp), %rsi call HIDDEN_JUMPTARGET(\callee) movq 128(%rsp), %rdx @@ -661,30 +632,8 @@ libmvec_hidden_def(_ZGVeN16vl4l4_sincosf_skx) leal -112(%rbp), %esi leal -176(%rbp), %edi subl $296, %esp - /* Encoding for vmovdqa64 %zmm1, -240(%ebp). */ - .byte 0x67 - .byte 0x62 - .byte 0xf1 - .byte 0xfd - .byte 0x48 - .byte 0x7f - .byte 0x8d - .byte 0x10 - .byte 0xff - .byte 0xff - .byte 0xff - /* Encoding for vmovdqa64 %zmm2, -304(%ebp). */ - .byte 0x67 - .byte 0x62 - .byte 0xf1 - .byte 0xfd - .byte 0x48 - .byte 0x7f - .byte 0x95 - .byte 0xd0 - .byte 0xfe - .byte 0xff - .byte 0xff + vmovdqa64 %zmm1, -240(%ebp) + vmovdqa64 %zmm2, -304(%ebp) call HIDDEN_JUMPTARGET(\callee) movl -240(%ebp), %eax vmovss -176(%ebp), %xmm0 |