From 3079f652d7cc34456aefb412677c01e758922527 Mon Sep 17 00:00:00 2001 From: Noah Goldstein Date: Mon, 20 Jun 2022 13:02:10 -0700 Subject: x86: Replace all sse instructions with vex equivilent in avx+ files Most of these don't really matter as there was no dirty upper state but we should generally avoid stray sse when its not needed. The one case that really matters is in svml_d_tanh4_core_avx2.S: blendvps %xmm0, %xmm8, %xmm7 When there was a dirty upper state. Tested on x86_64-linux --- sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S') diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S index ef9581075d..c3b0f7940c 100644 --- a/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S @@ -295,12 +295,12 @@ L(SPECIAL_VALUES_LOOP): L(SCALAR_MATH_CALL): movl %r12d, %r14d - movsd 64(%rsp, %r14, 8), %xmm0 - movsd 128(%rsp, %r14, 8), %xmm1 + vmovsd 64(%rsp, %r14, 8), %xmm0 + vmovsd 128(%rsp, %r14, 8), %xmm1 call atan2@PLT # LOE rbx r14 r15 r12d r13d xmm0 - movsd %xmm0, 192(%rsp, %r14, 8) + vmovsd %xmm0, 192(%rsp, %r14, 8) /* Process special inputs in loop */ jmp L(SPECIAL_VALUES_LOOP) -- cgit 1.4.1