@@ -5469,6 +5469,11 @@ ix86_get_ssemov (rtx *operands, unsigned size,
bool evex_reg_p = (size == 64
|| EXT_REX_SSE_REG_P (operands[0])
|| EXT_REX_SSE_REG_P (operands[1]));
+
+ bool egpr_p = (TARGET_APX_EGPR
+ && (x86_extended_rex2reg_mentioned_p (operands[0])
+ || x86_extended_rex2reg_mentioned_p (operands[1])));
+
machine_mode scalar_mode;
const char *opcode = NULL;
@@ -5547,6 +5552,12 @@ ix86_get_ssemov (rtx *operands, unsigned size,
? "vmovdqu16"
: "vmovdqu64")
: "vmovdqa64");
+ else if (egpr_p)
+ opcode = (misaligned_p
+ ? (TARGET_AVX512BW
+ ? "vmovdqu16"
+ : "%vmovups")
+ : "%vmovaps");
else
opcode = (misaligned_p
? (TARGET_AVX512BW
@@ -5563,6 +5574,8 @@ ix86_get_ssemov (rtx *operands, unsigned size,
case E_TFmode:
if (evex_reg_p)
opcode = misaligned_p ? "vmovdqu64" : "vmovdqa64";
+ else if (egpr_p)
+ opcode = misaligned_p ? "%vmovups" : "%vmovaps";
else
opcode = misaligned_p ? "%vmovdqu" : "%vmovdqa";
break;
@@ -5581,6 +5594,12 @@ ix86_get_ssemov (rtx *operands, unsigned size,
? "vmovdqu8"
: "vmovdqu64")
: "vmovdqa64");
+ else if (egpr_p)
+ opcode = (misaligned_p
+ ? (TARGET_AVX512BW
+ ? "vmovdqu8"
+ : "%vmovups")
+ : "%vmovaps");
else
opcode = (misaligned_p
? (TARGET_AVX512BW
@@ -5589,12 +5608,18 @@ ix86_get_ssemov (rtx *operands, unsigned size,
: "%vmovdqa");
break;
case E_HImode:
- if (evex_reg_p)
+ if (evex_reg_p || egpr_p)
opcode = (misaligned_p
? (TARGET_AVX512BW
? "vmovdqu16"
: "vmovdqu64")
: "vmovdqa64");
+ else if (egpr_p)
+ opcode = (misaligned_p
+ ? (TARGET_AVX512BW
+ ? "vmovdqu16"
+ : "%vmovups")
+ : "%vmovaps");
else
opcode = (misaligned_p
? (TARGET_AVX512BW
@@ -5605,6 +5630,8 @@ ix86_get_ssemov (rtx *operands, unsigned size,
case E_SImode:
if (evex_reg_p)
opcode = misaligned_p ? "vmovdqu32" : "vmovdqa32";
+ else if (egpr_p)
+ opcode = misaligned_p ? "%vmovups" : "%vmovaps";
else
opcode = misaligned_p ? "%vmovdqu" : "%vmovdqa";
break;
@@ -5613,6 +5640,8 @@ ix86_get_ssemov (rtx *operands, unsigned size,
case E_OImode:
if (evex_reg_p)
opcode = misaligned_p ? "vmovdqu64" : "vmovdqa64";
+ else if (egpr_p)
+ opcode = misaligned_p ? "%vmovups" : "%vmovaps";
else
opcode = misaligned_p ? "%vmovdqu" : "%vmovdqa";
break;
@@ -18918,6 +18918,12 @@ (define_insn "*<extract_type>_vinsert<shuffletype><extract_suf>_0"
{
if (which_alternative == 0)
return "vinsert<shuffletype><extract_suf>\t{$0, %2, %1, %0|%0, %1, %2, 0}";
+ bool egpr_used = (TARGET_APX_EGPR
+ && x86_extended_rex2reg_mentioned_p (operands[2]));
+ const char *align_templ = egpr_used ? "vmovdqa\t{%2, %x0|%x0, %2}"
+ : "vmovaps\t{%2, %x0|%x0, %2}";
+ const char *unalign_templ = egpr_used ? "vmovdqu\t{%2, %x0|%x0, %2}"
+ : "vmovups\t{%2, %x0|%x0, %2}";
switch (<MODE>mode)
{
case E_V8DFmode:
@@ -18933,17 +18939,17 @@ (define_insn "*<extract_type>_vinsert<shuffletype><extract_suf>_0"
case E_V8DImode:
if (misaligned_operand (operands[2], <ssequartermode>mode))
return which_alternative == 2 ? "vmovdqu64\t{%2, %x0|%x0, %2}"
- : "vmovdqu\t{%2, %x0|%x0, %2}";
+ : unalign_templ;
else
return which_alternative == 2 ? "vmovdqa64\t{%2, %x0|%x0, %2}"
- : "vmovdqa\t{%2, %x0|%x0, %2}";
+ : align_templ;
case E_V16SImode:
if (misaligned_operand (operands[2], <ssequartermode>mode))
return which_alternative == 2 ? "vmovdqu32\t{%2, %x0|%x0, %2}"
- : "vmovdqu\t{%2, %x0|%x0, %2}";
+ : unalign_templ;
else
return which_alternative == 2 ? "vmovdqa32\t{%2, %x0|%x0, %2}"
- : "vmovdqa\t{%2, %x0|%x0, %2}";
+ : align_templ;
default:
gcc_unreachable ();
}
@@ -27652,11 +27658,13 @@ (define_insn "avx_vec_concat<mode>"
[(set (match_operand:V_256_512 0 "register_operand" "=x,v,x,Yv")
(vec_concat:V_256_512
(match_operand:<ssehalfvecmode> 1 "nonimmediate_operand" "x,v,xm,vm")
- (match_operand:<ssehalfvecmode> 2 "nonimm_or_0_operand" "xm,vm,C,C")))]
+ (match_operand:<ssehalfvecmode> 2 "nonimm_or_0_operand" "xBt,vm,C,C")))]
"TARGET_AVX
&& (operands[2] == CONST0_RTX (<ssehalfvecmode>mode)
|| !MEM_P (operands[1]))"
{
+ bool egpr_used = (TARGET_APX_EGPR
+ && x86_extended_rex2reg_mentioned_p (operands[1]));
switch (which_alternative)
{
case 0:
@@ -27704,7 +27712,8 @@ (define_insn "avx_vec_concat<mode>"
if (misaligned_operand (operands[1], <ssehalfvecmode>mode))
{
if (which_alternative == 2)
- return "vmovdqu\t{%1, %t0|%t0, %1}";
+ return egpr_used ? "vmovups\t{%1, %t0|%t0, %1}"
+ : "vmovdqu\t{%1, %t0|%t0, %1}";
else if (GET_MODE_SIZE (<ssescalarmode>mode) == 8)
return "vmovdqu64\t{%1, %t0|%t0, %1}";
else
@@ -27713,7 +27722,8 @@ (define_insn "avx_vec_concat<mode>"
else
{
if (which_alternative == 2)
- return "vmovdqa\t{%1, %t0|%t0, %1}";
+ return egpr_used ? "vmovaps\t{%1, %t0|%t0, %1}"
+ : "vmovdqa\t{%1, %t0|%t0, %1}";
else if (GET_MODE_SIZE (<ssescalarmode>mode) == 8)
return "vmovdqa64\t{%1, %t0|%t0, %1}";
else
@@ -27723,7 +27733,8 @@ (define_insn "avx_vec_concat<mode>"
if (misaligned_operand (operands[1], <ssehalfvecmode>mode))
{
if (which_alternative == 2)
- return "vmovdqu\t{%1, %x0|%x0, %1}";
+ return egpr_used ? "vmovups\t{%1, %x0|%x0, %1}"
+ : "vmovdqu\t{%1, %x0|%x0, %1}";
else if (GET_MODE_SIZE (<ssescalarmode>mode) == 8)
return "vmovdqu64\t{%1, %x0|%x0, %1}";
else
@@ -27732,7 +27743,8 @@ (define_insn "avx_vec_concat<mode>"
else
{
if (which_alternative == 2)
- return "vmovdqa\t{%1, %x0|%x0, %1}";
+ return egpr_used ? "vmovaps\t{%1, %x0|%x0, %1}"
+ : "vmovdqa\t{%1, %x0|%x0, %1}";
else if (GET_MODE_SIZE (<ssescalarmode>mode) == 8)
return "vmovdqa64\t{%1, %x0|%x0, %1}";
else
@@ -27745,7 +27757,9 @@ (define_insn "avx_vec_concat<mode>"
gcc_unreachable ();
}
}
- [(set_attr "type" "sselog,sselog,ssemov,ssemov")
+ [(set_attr "isa" "noavx512f,avx512f,*,*")
+ (set_attr "gpr32" "0,1,1,1")
+ (set_attr "type" "sselog,sselog,ssemov,ssemov")
(set_attr "prefix_extra" "1,1,*,*")
(set_attr "length_immediate" "1,1,*,*")
(set_attr "prefix" "maybe_evex")