OpenE2K/gcc
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[i386]Migrate reduction optabs to reduc_<op>_scal

Browse Source 
* config/i386/sse.md (reduc_splus_v8df): Rename to...
	(reduc_plus_scal_v8df): ...here; reduce to temp and extract scalar.

	(reduc_splus_v4df): Rename to...
	(reduc_plus_scal_v4df): ...here; reduce to temp and extract scalar.

	(reduc_splus_v2df): Rename to...
	(reduc_plus_scal_v2df): ...here; reduce to temp and extract scalar.

	(reduc_splus_v16sf): Rename to...
	(reduc_plus_scal_v16sf): ...here; reduce to temp and extract scalar.

	(reduc_splus_v8sf): Rename to...
	(reduc_plus_scal_v8sf): ...here; reduce to temp and extract scalar.

	(reduc_splus_v4sf): Rename to...
	(reduc_plus_scal_v4sf): ...here; reduce to temp and extract scalar.

	(reduc_<code>_<mode>, all 3 variants): Rename each to...
	(reduc_<code>_scal_<mode>): ...here; reduce to temp and extract scalar.

	(reduc_umin_v8hf): Rename to...
	(reduc_umin_scal_v8hf): ...here; reduce to temp and extract scalar.

From-SVN: r230423
This commit is contained in:
Alan Lawrence 2015-11-16 13:53:23 +00:00 committed by Alan Lawrence
parent fff7721799
commit def4ad19ea
2 changed files with 77 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
gcc/ChangeLog
View File

@ -1,3 +1,29 @@
2015-11-16 Alan Lawrence <alan.lawrence@arm.com>
* config/i386/sse.md (reduc_splus_v8df): Rename to...
(reduc_plus_scal_v8df): ...here; reduce to temp and extract scalar.
(reduc_splus_v4df): Rename to...
(reduc_plus_scal_v4df): ...here; reduce to temp and extract scalar.
(reduc_splus_v2df): Rename to...
(reduc_plus_scal_v2df): ...here; reduce to temp and extract scalar.
(reduc_splus_v16sf): Rename to...
(reduc_plus_scal_v16sf): ...here; reduce to temp and extract scalar.
(reduc_splus_v8sf): Rename to...
(reduc_plus_scal_v8sf): ...here; reduce to temp and extract scalar.
(reduc_splus_v4sf): Rename to...
(reduc_plus_scal_v4sf): ...here; reduce to temp and extract scalar.
(reduc_<code>_<mode>, all 3 variants): Rename each to...
(reduc_<code>_scal_<mode>): ...here; reduce to temp and extract scalar.
(reduc_umin_v8hf): Rename to...
(reduc_umin_scal_v8hf): ...here; reduce to temp and extract scalar.
2015-11-16 Kirill Yukhin <kirill.yukhin@intel.com> 2015-11-16 Kirill Yukhin <kirill.yukhin@intel.com>
* omp-low.c (pass_omp_simd_clone::gate): If target allows - call * omp-low.c (pass_omp_simd_clone::gate): If target allows - call

82
gcc/config/i386/sse.md
View File

@ -2441,73 +2441,85 @@
(set_attr "prefix_rep" "1,*") (set_attr "prefix_rep" "1,*")
(set_attr "mode" "V4SF")]) (set_attr "mode" "V4SF")])
(define_expand "reduc_splus_v8df" (define_expand "reduc_plus_scal_v8df"
[(match_operand:V8DF 0 "register_operand") [(match_operand:DF 0 "register_operand")
(match_operand:V8DF 1 "register_operand")] (match_operand:V8DF 1 "register_operand")]
"TARGET_AVX512F" "TARGET_AVX512F"
{ {
ix86_expand_reduc (gen_addv8df3, operands[0], operands[1]); rtx tmp = gen_reg_rtx (V8DFmode);
ix86_expand_reduc (gen_addv8df3, tmp, operands[1]);
emit_insn (gen_vec_extractv8df (operands[0], tmp, const0_rtx));
DONE; DONE;
}) })
(define_expand "reduc_splus_v4df" (define_expand "reduc_plus_scal_v4df"
[(match_operand:V4DF 0 "register_operand") [(match_operand:DF 0 "register_operand")
(match_operand:V4DF 1 "register_operand")] (match_operand:V4DF 1 "register_operand")]
"TARGET_AVX" "TARGET_AVX"
{ {
rtx tmp = gen_reg_rtx (V4DFmode); rtx tmp = gen_reg_rtx (V4DFmode);
rtx tmp2 = gen_reg_rtx (V4DFmode); rtx tmp2 = gen_reg_rtx (V4DFmode);
rtx vec_res = gen_reg_rtx (V4DFmode);
emit_insn (gen_avx_haddv4df3 (tmp, operands[1], operands[1])); emit_insn (gen_avx_haddv4df3 (tmp, operands[1], operands[1]));
emit_insn (gen_avx_vperm2f128v4df3 (tmp2, tmp, tmp, GEN_INT (1))); emit_insn (gen_avx_vperm2f128v4df3 (tmp2, tmp, tmp, GEN_INT (1)));
emit_insn (gen_addv4df3 (operands[0], tmp, tmp2)); emit_insn (gen_addv4df3 (vec_res, tmp, tmp2));
emit_insn (gen_vec_extractv4df (operands[0], vec_res, const0_rtx));
DONE; DONE;
}) })
(define_expand "reduc_splus_v2df" (define_expand "reduc_plus_scal_v2df"
[(match_operand:V2DF 0 "register_operand") [(match_operand:DF 0 "register_operand")
(match_operand:V2DF 1 "register_operand")] (match_operand:V2DF 1 "register_operand")]
"TARGET_SSE3" "TARGET_SSE3"
{ {
emit_insn (gen_sse3_haddv2df3 (operands[0], operands[1], operands[1])); rtx tmp = gen_reg_rtx (V2DFmode);
emit_insn (gen_sse3_haddv2df3 (tmp, operands[1], operands[1]));
emit_insn (gen_vec_extractv2df (operands[0], tmp, const0_rtx));
DONE; DONE;
}) })
(define_expand "reduc_splus_v16sf" (define_expand "reduc_plus_scal_v16sf"
[(match_operand:V16SF 0 "register_operand") [(match_operand:SF 0 "register_operand")
(match_operand:V16SF 1 "register_operand")] (match_operand:V16SF 1 "register_operand")]
"TARGET_AVX512F" "TARGET_AVX512F"
{ {
ix86_expand_reduc (gen_addv16sf3, operands[0], operands[1]); rtx tmp = gen_reg_rtx (V16SFmode);
ix86_expand_reduc (gen_addv16sf3, tmp, operands[1]);
emit_insn (gen_vec_extractv16sf (operands[0], tmp, const0_rtx));
DONE; DONE;
}) })
(define_expand "reduc_splus_v8sf" (define_expand "reduc_plus_scal_v8sf"
[(match_operand:V8SF 0 "register_operand") [(match_operand:SF 0 "register_operand")
(match_operand:V8SF 1 "register_operand")] (match_operand:V8SF 1 "register_operand")]
"TARGET_AVX" "TARGET_AVX"
{ {
rtx tmp = gen_reg_rtx (V8SFmode); rtx tmp = gen_reg_rtx (V8SFmode);
rtx tmp2 = gen_reg_rtx (V8SFmode); rtx tmp2 = gen_reg_rtx (V8SFmode);
rtx vec_res = gen_reg_rtx (V8SFmode);
emit_insn (gen_avx_haddv8sf3 (tmp, operands[1], operands[1])); emit_insn (gen_avx_haddv8sf3 (tmp, operands[1], operands[1]));
emit_insn (gen_avx_haddv8sf3 (tmp2, tmp, tmp)); emit_insn (gen_avx_haddv8sf3 (tmp2, tmp, tmp));
emit_insn (gen_avx_vperm2f128v8sf3 (tmp, tmp2, tmp2, GEN_INT (1))); emit_insn (gen_avx_vperm2f128v8sf3 (tmp, tmp2, tmp2, GEN_INT (1)));
emit_insn (gen_addv8sf3 (operands[0], tmp, tmp2)); emit_insn (gen_addv8sf3 (vec_res, tmp, tmp2));
emit_insn (gen_vec_extractv8sf (operands[0], vec_res, const0_rtx));
DONE; DONE;
}) })
(define_expand "reduc_splus_v4sf" (define_expand "reduc_plus_scal_v4sf"
[(match_operand:V4SF 0 "register_operand") [(match_operand:SF 0 "register_operand")
(match_operand:V4SF 1 "register_operand")] (match_operand:V4SF 1 "register_operand")]
"TARGET_SSE" "TARGET_SSE"
{ {
rtx vec_res = gen_reg_rtx (V4SFmode);
if (TARGET_SSE3) if (TARGET_SSE3)
{ {
rtx tmp = gen_reg_rtx (V4SFmode); rtx tmp = gen_reg_rtx (V4SFmode);
emit_insn (gen_sse3_haddv4sf3 (tmp, operands[1], operands[1])); emit_insn (gen_sse3_haddv4sf3 (tmp, operands[1], operands[1]));
emit_insn (gen_sse3_haddv4sf3 (operands[0], tmp, tmp)); emit_insn (gen_sse3_haddv4sf3 (vec_res, tmp, tmp));
} }
else else
ix86_expand_reduc (gen_addv4sf3, operands[0], operands[1]); ix86_expand_reduc (gen_addv4sf3, vec_res, operands[1]);
emit_insn (gen_vec_extractv4sf (operands[0], vec_res, const0_rtx));
DONE; DONE;
}) })
@ -2521,43 +2533,51 @@
(V8DI "TARGET_AVX512F") (V16SF "TARGET_AVX512F") (V8DI "TARGET_AVX512F") (V16SF "TARGET_AVX512F")
(V8DF "TARGET_AVX512F")]) (V8DF "TARGET_AVX512F")])
(define_expand "reduc_<code>_<mode>" (define_expand "reduc_<code>_scal_<mode>"
[(smaxmin:REDUC_SMINMAX_MODE [(smaxmin:REDUC_SMINMAX_MODE
(match_operand:REDUC_SMINMAX_MODE 0 "register_operand") (match_operand:<ssescalarmode> 0 "register_operand")
(match_operand:REDUC_SMINMAX_MODE 1 "register_operand"))] (match_operand:REDUC_SMINMAX_MODE 1 "register_operand"))]
"" ""
{ {
ix86_expand_reduc (gen_<code><mode>3, operands[0], operands[1]); rtx tmp = gen_reg_rtx (<MODE>mode);
ix86_expand_reduc (gen_<code><mode>3, tmp, operands[1]);
emit_insn (gen_vec_extract<mode> (operands[0], tmp, const0_rtx));
DONE; DONE;
}) })
(define_expand "reduc_<code>_<mode>" (define_expand "reduc_<code>_scal_<mode>"
[(umaxmin:VI_AVX512BW [(umaxmin:VI_AVX512BW
(match_operand:VI_AVX512BW 0 "register_operand") (match_operand:<ssescalarmode> 0 "register_operand")
(match_operand:VI_AVX512BW 1 "register_operand"))] (match_operand:VI_AVX512BW 1 "register_operand"))]
"TARGET_AVX512F" "TARGET_AVX512F"
{ {
ix86_expand_reduc (gen_<code><mode>3, operands[0], operands[1]); rtx tmp = gen_reg_rtx (<MODE>mode);
ix86_expand_reduc (gen_<code><mode>3, tmp, operands[1]);
emit_insn (gen_vec_extract<mode> (operands[0], tmp, const0_rtx));
DONE; DONE;
}) })
(define_expand "reduc_<code>_<mode>" (define_expand "reduc_<code>_scal_<mode>"
[(umaxmin:VI_256 [(umaxmin:VI_256
(match_operand:VI_256 0 "register_operand") (match_operand:<ssescalarmode> 0 "register_operand")
(match_operand:VI_256 1 "register_operand"))] (match_operand:VI_256 1 "register_operand"))]
"TARGET_AVX2" "TARGET_AVX2"
{ {
ix86_expand_reduc (gen_<code><mode>3, operands[0], operands[1]); rtx tmp = gen_reg_rtx (<MODE>mode);
ix86_expand_reduc (gen_<code><mode>3, tmp, operands[1]);
emit_insn (gen_vec_extract<mode> (operands[0], tmp, const0_rtx));
DONE; DONE;
}) })
(define_expand "reduc_umin_v8hi" (define_expand "reduc_umin_scal_v8hi"
[(umin:V8HI [(umin:V8HI
(match_operand:V8HI 0 "register_operand") (match_operand:HI 0 "register_operand")
(match_operand:V8HI 1 "register_operand"))] (match_operand:V8HI 1 "register_operand"))]
"TARGET_SSE4_1" "TARGET_SSE4_1"
{ {
ix86_expand_reduc (gen_uminv8hi3, operands[0], operands[1]); rtx tmp = gen_reg_rtx (V8HImode);
ix86_expand_reduc (gen_uminv8hi3, tmp, operands[1]);
emit_insn (gen_vec_extractv8hi (operands[0], tmp, const0_rtx));
DONE; DONE;
}) })