rs6000: Support variable insert and Expand vec_insert in expander [PR79251]

vec_insert accepts 3 arguments, arg0 is input vector, arg1 is the value
to be insert, arg2 is the place to insert arg1 to arg0.  Current expander
generates stxv+stwx+lxv if arg2 is variable instead of constant, which
causes serious store hit load performance issue on Power.  This patch tries
 1) Build VIEW_CONVERT_EXPR for vec_insert (i, v, n) like v[n&3] = i to
unify the gimple code, then expander could use vec_set_optab to expand.
 2) Expand the IFN VEC_SET to fast instructions: lvsr+insert+lvsl.
In this way, "vec_insert (i, v, n)" and "v[n&3] = i" won't be expanded too
early in gimple stage if arg2 is variable, avoid generating store hit load
instructions.

For Power9 V4SI:
	addi 9,1,-16
	rldic 6,6,2,60
	stxv 34,-16(1)
	stwx 5,9,6
	lxv 34,-16(1)
=>
	rlwinm 6,6,2,28,29
	mtvsrwz 0,5
	lvsr 1,0,6
	lvsl 0,0,6
	xxperm 34,34,33
	xxinsertw 34,0,12
	xxperm 34,34,32

Though instructions increase from 5 to 7, the performance is improved
60% in typical cases.
Tested with V2DI, V2DF V4SI, V4SF, V8HI, V16QI on Power9-LE.

2021-01-22  Xionghu Luo  <luoxhu@linux.ibm.com>

gcc/ChangeLog:

	PR target/79251
	PR target/98065

	* config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
	Ajdust variable index vec_insert from address dereference to
	ARRAY_REF(VIEW_CONVERT_EXPR) tree expression.
	* config/rs6000/rs6000-protos.h (rs6000_expand_vector_set_var):
	New declaration.
	* config/rs6000/rs6000.c (rs6000_expand_vector_set_var): New function.

2021-01-22  Xionghu Luo  <luoxhu@linux.ibm.com>

gcc/testsuite/ChangeLog:

	* gcc.target/powerpc/pr79251.p9.c: New test.
	* gcc.target/powerpc/pr79251-run.c: New test.
	* gcc.target/powerpc/pr79251.h: New header.
This commit is contained in:
Xionghu Luo 2021-01-21 21:01:24 -06:00
parent b46027c654
commit b292255975
6 changed files with 130 additions and 14 deletions

View File

@ -1512,9 +1512,7 @@ altivec_resolve_overloaded_builtin (location_t loc, tree fndecl,
tree arg1;
tree arg2;
tree arg1_type;
tree arg1_inner_type;
tree decl, stmt;
tree innerptrtype;
machine_mode mode;
/* No second or third arguments. */
@ -1566,8 +1564,13 @@ altivec_resolve_overloaded_builtin (location_t loc, tree fndecl,
return build_call_expr (call, 3, arg1, arg0, arg2);
}
/* Build *(((arg1_inner_type*)&(vector type){arg1})+arg2) = arg0. */
arg1_inner_type = TREE_TYPE (arg1_type);
/* Build *(((arg1_inner_type*)&(vector type){arg1})+arg2) = arg0 with
VIEW_CONVERT_EXPR. i.e.:
D.3192 = v1;
_1 = n & 3;
VIEW_CONVERT_EXPR<int[4]>(D.3192)[_1] = i;
v1 = D.3192;
D.3194 = v1; */
if (TYPE_VECTOR_SUBPARTS (arg1_type) == 1)
arg2 = build_int_cst (TREE_TYPE (arg2), 0);
else
@ -1582,6 +1585,7 @@ altivec_resolve_overloaded_builtin (location_t loc, tree fndecl,
TREE_USED (decl) = 1;
TREE_TYPE (decl) = arg1_type;
TREE_READONLY (decl) = TYPE_READONLY (arg1_type);
TREE_ADDRESSABLE (decl) = 1;
if (c_dialect_cxx ())
{
stmt = build4 (TARGET_EXPR, arg1_type, decl, arg1,
@ -1592,19 +1596,12 @@ altivec_resolve_overloaded_builtin (location_t loc, tree fndecl,
{
DECL_INITIAL (decl) = arg1;
stmt = build1 (DECL_EXPR, arg1_type, decl);
TREE_ADDRESSABLE (decl) = 1;
SET_EXPR_LOCATION (stmt, loc);
stmt = build1 (COMPOUND_LITERAL_EXPR, arg1_type, stmt);
}
innerptrtype = build_pointer_type (arg1_inner_type);
stmt = build_unary_op (loc, ADDR_EXPR, stmt, 0);
stmt = convert (innerptrtype, stmt);
stmt = build_binary_op (loc, PLUS_EXPR, stmt, arg2, 1);
stmt = build_indirect_ref (loc, stmt, RO_NULL);
stmt = build2 (MODIFY_EXPR, TREE_TYPE (stmt), stmt,
convert (TREE_TYPE (stmt), arg0));
stmt = build_array_ref (loc, stmt, arg2);
stmt = fold_build2 (MODIFY_EXPR, TREE_TYPE (arg0), stmt,
convert (TREE_TYPE (stmt), arg0));
stmt = build2 (COMPOUND_EXPR, arg1_type, stmt, decl);
return stmt;
}

View File

@ -58,6 +58,7 @@ extern bool rs6000_split_128bit_ok_p (rtx []);
extern void rs6000_expand_float128_convert (rtx, rtx, bool);
extern void rs6000_expand_vector_init (rtx, rtx);
extern void rs6000_expand_vector_set (rtx, rtx, rtx);
extern void rs6000_expand_vector_set_var (rtx, rtx, rtx);
extern void rs6000_expand_vector_extract (rtx, rtx, rtx);
extern void rs6000_split_vec_extract_var (rtx, rtx, rtx, rtx, rtx);
extern rtx rs6000_adjust_vec_address (rtx, rtx, rtx, rtx, machine_mode);

View File

@ -6993,6 +6993,12 @@ rs6000_expand_vector_set (rtx target, rtx val, rtx elt_rtx)
if (VECTOR_MEM_VSX_P (mode))
{
if (!CONST_INT_P (elt_rtx))
{
rs6000_expand_vector_set_var (target, val, elt_rtx);
return;
}
rtx insn = NULL_RTX;
if (mode == V2DFmode)
@ -7083,6 +7089,53 @@ rs6000_expand_vector_set (rtx target, rtx val, rtx elt_rtx)
emit_insn (gen_rtx_SET (target, x));
}
/* Insert VAL into IDX of TARGET, VAL size is same of the vector element, IDX
is variable and also counts by vector element size. */
void
rs6000_expand_vector_set_var (rtx target, rtx val, rtx idx)
{
machine_mode mode = GET_MODE (target);
gcc_assert (VECTOR_MEM_VSX_P (mode) && !CONST_INT_P (idx));
gcc_assert (GET_MODE (idx) == E_SImode);
machine_mode inner_mode = GET_MODE (val);
rtx tmp = gen_reg_rtx (GET_MODE (idx));
int width = GET_MODE_SIZE (inner_mode);
gcc_assert (width >= 1 && width <= 8);
int shift = exact_log2 (width);
/* Generate the IDX for permute shift, width is the vector element size.
idx = idx * width. */
emit_insn (gen_ashlsi3 (tmp, idx, GEN_INT (shift)));
tmp = convert_modes (DImode, SImode, tmp, 1);
/* lvsr v1,0,idx. */
rtx pcvr = gen_reg_rtx (V16QImode);
emit_insn (gen_altivec_lvsr_reg (pcvr, tmp));
/* lvsl v2,0,idx. */
rtx pcvl = gen_reg_rtx (V16QImode);
emit_insn (gen_altivec_lvsl_reg (pcvl, tmp));
rtx sub_target = simplify_gen_subreg (V16QImode, target, mode, 0);
rtx permr
= gen_altivec_vperm_v8hiv16qi (sub_target, sub_target, sub_target, pcvr);
emit_insn (permr);
rs6000_expand_vector_set (target, val, const0_rtx);
rtx perml
= gen_altivec_vperm_v8hiv16qi (sub_target, sub_target, sub_target, pcvl);
emit_insn (perml);
}
/* Extract field ELT from VEC into TARGET. */
void

View File

@ -0,0 +1,28 @@
/* { dg-options "-O2 -maltivec" } */
#include <stddef.h>
#include <altivec.h>
#include "pr79251.h"
TEST_VEC_INSERT_ALL (test)
#define run_test(TYPE, num) \
{ \
vector TYPE v; \
vector TYPE u = {0x0}; \
for (long k = 0; k < 16 / sizeof (TYPE); k++) \
v[k] = 0xaa; \
for (long k = 0; k < 16 / sizeof (TYPE); k++) \
{ \
u = test##num (v, 254, k); \
if (u[k] != (TYPE) 254) \
__builtin_abort (); \
} \
}
int
main (void)
{
TEST_VEC_INSERT_ALL (run_test)
return 0;
}

View File

@ -0,0 +1,19 @@
#define test(TYPE, num) \
__attribute__ ((noinline, noclone)) \
vector TYPE test##num (vector TYPE v, TYPE i, signed int n) \
{ \
return vec_insert (i, v, n); \
}
#define TEST_VEC_INSERT_ALL(T) \
T (char, 0) \
T (unsigned char, 1) \
T (short, 2) \
T (unsigned short, 3) \
T (int, 4) \
T (unsigned int, 5) \
T (long long, 6) \
T (unsigned long long, 7) \
T (float, 8) \
T (double, 9)

View File

@ -0,0 +1,18 @@
/* { dg-do compile } */
/* { dg-require-effective-target powerpc_p9vector_ok } */
/* { dg-options "-O2 -mdejagnu-cpu=power9 -maltivec" } */
#include <stddef.h>
#include <altivec.h>
#include "pr79251.h"
TEST_VEC_INSERT_ALL (test)
/* { dg-final { scan-assembler-not {\mstxw\M} } } */
/* { dg-final { scan-assembler-times {\mlvsl\M} 10 } } */
/* { dg-final { scan-assembler-times {\mlvsr\M} 10 } } */
/* { dg-final { scan-assembler-times {\mxxperm\M} 20 } } */
/* { dg-final { scan-assembler-times {\mxxinsertw\M} 3 } } */
/* { dg-final { scan-assembler-times {\mvinserth\M} 2 } } */
/* { dg-final { scan-assembler-times {\mvinsertb\M} 2 } } */
/* { dg-final { scan-assembler-times {\mxxpermdi\M} 3 } } */