https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92097
Bug ID: 92097
Summary: instrutions only vaild under TARGET_AVX
Product: gcc
Version: 10.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: crazylht at gmail dot com
CC: hjl.tools at gmail dot com, uros at gcc dot gnu.org
Target Milestone: ---
Target: i386, x86-64
According to intel sdm, "sqrtss" is under TARGET_SSE, "vsqrtss" is under
TARGET_AVX, so define_insn *rsqrtsf2_sse should be under TARGET_AVX but it is
TARGET_SSE now, is it on purpose?
-----
(define_insn "*rsqrtsf2_sse"
[(set (match_operand:SF 0 "register_operand" "=x,x,x")
(unspec:SF [(match_operand:SF 1 "nonimmediate_operand" "0,x,m")]
UNSPEC_RSQRT))]
"TARGET_SSE && TARGET_SSE_MATH"
"@
%vrsqrtss\t{%d1, %0|%0, %d1}
%vrsqrtss\t{%d1, %0|%0, %d1}
%vrsqrtss\t{%1, %d0|%d0, %1}"
[(set_attr "type" "sse")
(set_attr "atom_sse_attr" "rcp")
(set_attr "btver2_sse_attr" "rcp")
(set_attr "prefix" "maybe_vex")
(set_attr "mode" "SF")
(set (attr "preferred_for_speed")
(cond [(eq_attr "alternative" "1")
(symbol_ref "TARGET_AVX || !TARGET_SSE_PARTIAL_REG_DEPENDENCY")
(eq_attr "alternative" "2")
(symbol_ref "!TARGET_SSE_PARTIAL_REG_DEPENDENCY")
]
(symbol_ref "true")))])
-----
also for *sqrt<mode>2_sse,*rcpsf2_sse
Similar for sse4_1_round<mode>2
