Model the divider in Lujiazui processors as a separate automaton to significantly reduce the overall model size. This should also result in improved accuracy, as pipe 0 should be able to accept new instructions while the divider is occupied.
It is unclear why integer divisions are modeled as if pipes 0-3 are all occupied. I've opted to keep a single-cycle reservation of all four pipes together, so GCC should continue trying to pack instructions around a division accordingly. Currently top three symbols in insn-automata.o are: 106102 r lujiazui_core_check 106102 r lujiazui_core_transitions 196123 r lujiazui_core_min_issue_delay This patch shrinks all lujiazui tables to: 3 r lujiazui_decoder_min_issue_delay 20 r lujiazui_decoder_transitions 32 r lujiazui_agu_min_issue_delay 126 r lujiazui_agu_transitions 304 r lujiazui_div_base 352 r lujiazui_div_check 352 r lujiazui_div_transitions 1152 r lujiazui_core_min_issue_delay 1592 r lujiazui_agu_translate 1592 r lujiazui_core_translate 1592 r lujiazui_decoder_translate 1592 r lujiazui_div_translate 3952 r lujiazui_div_min_issue_delay 9216 r lujiazui_core_transitions This continues the work on reducing i386 insn-automata.o size started with similar fixes for division and multiplication instructions in znver.md [1][2]. I plan to submit corresponding fixes for b[td]ver[123].md as well. [1] https://inbox.sourceware.org/gcc-patches/23c795d6-403c-5927-e610-f0f1215f5...@ispras.ru/T/#m36e069d43d07d768d4842a779e26b4a0915cc543 [2] https://inbox.sourceware.org/gcc-patches/20221101162637.14238-1-amona...@ispras.ru/ gcc/ChangeLog: PR target/87832 * config/i386/lujiazui.md (lujiazui_div): New automaton. (lua_div): New unit. (lua_idiv_qi): Correct unit in the reservation. (lua_idiv_qi_load): Ditto. (lua_idiv_hi): Ditto. (lua_idiv_hi_load): Ditto. (lua_idiv_si): Ditto. (lua_idiv_si_load): Ditto. (lua_idiv_di): Ditto. (lua_idiv_di_load): Ditto. (lua_fdiv_SF): Ditto. (lua_fdiv_SF_load): Ditto. (lua_fdiv_DF): Ditto. (lua_fdiv_DF_load): Ditto. (lua_fdiv_XF): Ditto. (lua_fdiv_XF_load): Ditto. (lua_ssediv_SF): Ditto. (lua_ssediv_load_SF): Ditto. (lua_ssediv_V4SF): Ditto. (lua_ssediv_load_V4SF): Ditto. (lua_ssediv_V8SF): Ditto. (lua_ssediv_load_V8SF): Ditto. (lua_ssediv_SD): Ditto. (lua_ssediv_load_SD): Ditto. (lua_ssediv_V2DF): Ditto. (lua_ssediv_load_V2DF): Ditto. (lua_ssediv_V4DF): Ditto. (lua_ssediv_load_V4DF): Ditto. (lua_sseicvt_si): Ditto. --- gcc/config/i386/lujiazui.md | 58 +++++++++++++++++++------------------ 1 file changed, 30 insertions(+), 28 deletions(-) diff --git a/gcc/config/i386/lujiazui.md b/gcc/config/i386/lujiazui.md index 9046c09f2..58a230c70 100644 --- a/gcc/config/i386/lujiazui.md +++ b/gcc/config/i386/lujiazui.md @@ -19,8 +19,8 @@ ;; Scheduling for ZHAOXIN lujiazui processor. -;; Modeling automatons for decoders, execution pipes and AGU pipes. -(define_automaton "lujiazui_decoder,lujiazui_core,lujiazui_agu") +;; Modeling automatons for decoders, execution pipes, AGU pipes, and divider. +(define_automaton "lujiazui_decoder,lujiazui_core,lujiazui_agu,lujiazui_div") ;; The rules for the decoder are simple: ;; - an instruction with 1 uop can be decoded by any of the three @@ -55,6 +55,8 @@ (define_reservation "lua_decoder01" "lua_decoder0|lua_decoder1") (define_cpu_unit "lua_p0,lua_p1,lua_p2,lua_p3" "lujiazui_core") (define_cpu_unit "lua_p4,lua_p5" "lujiazui_agu") +(define_cpu_unit "lua_div" "lujiazui_div") + (define_reservation "lua_p03" "lua_p0|lua_p3") (define_reservation "lua_p12" "lua_p1|lua_p2") (define_reservation "lua_p1p2" "lua_p1+lua_p2") @@ -229,56 +231,56 @@ (define_insn_reservation "lua_idiv_qi" 21 (and (eq_attr "memory" "none") (and (eq_attr "mode" "QI") (eq_attr "type" "idiv")))) - "lua_decoder0,lua_p0p1p2p3*21") + "lua_decoder0,lua_p0p1p2p3,lua_div*21") (define_insn_reservation "lua_idiv_qi_load" 25 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "QI") (eq_attr "type" "idiv")))) - "lua_decoder0,lua_p45,lua_p0p1p2p3*21") + "lua_decoder0,lua_p45,lua_p0p1p2p3,lua_div*21") (define_insn_reservation "lua_idiv_hi" 22 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "HI") (eq_attr "type" "idiv")))) - "lua_decoder0,lua_p0p1p2p3*22") + "lua_decoder0,lua_p0p1p2p3,lua_div*22") (define_insn_reservation "lua_idiv_hi_load" 26 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "HI") (eq_attr "type" "idiv")))) - "lua_decoder0,lua_p45,lua_p0p1p2p3*22") + "lua_decoder0,lua_p45,lua_p0p1p2p3,lua_div*22") (define_insn_reservation "lua_idiv_si" 20 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "SI") (eq_attr "type" "idiv")))) - "lua_decoder0,lua_p0p1p2p3*20") + "lua_decoder0,lua_p0p1p2p3,lua_div*20") (define_insn_reservation "lua_idiv_si_load" 24 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "SI") (eq_attr "type" "idiv")))) - "lua_decoder0,lua_p45,lua_p0p1p2p3*20") + "lua_decoder0,lua_p45,lua_p0p1p2p3,lua_div*20") (define_insn_reservation "lua_idiv_di" 150 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "DI") (eq_attr "type" "idiv")))) - "lua_decoder0,lua_p0p1p2p3*150") + "lua_decoder0,lua_p0p1p2p3,lua_div*150") (define_insn_reservation "lua_idiv_di_load" 154 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "DI") (eq_attr "type" "idiv")))) - "lua_decoder0,lua_p45,lua_p0p1p2p3*150") + "lua_decoder0,lua_p45,lua_p0p1p2p3,lua_div*150") ;; x87 floating point operations. @@ -406,42 +408,42 @@ (define_insn_reservation "lua_fdiv_SF" 15 (and (eq_attr "memory" "none") (and (eq_attr "mode" "SF") (eq_attr "type" "fdiv,fpspc")))) - "lua_decodern,lua_p0*15") + "lua_decodern,lua_p0,lua_div*15") (define_insn_reservation "lua_fdiv_SF_load" 19 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "SF") (eq_attr "type" "fdiv,fpspc")))) - "lua_decoder01,lua_p45,lua_p0*15") + "lua_decoder01,lua_p45,lua_p0,lua_div*15") (define_insn_reservation "lua_fdiv_DF" 18 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "DF") (eq_attr "type" "fdiv,fpspc")))) - "lua_decodern,lua_p0*18") + "lua_decodern,lua_p0,lua_div*18") (define_insn_reservation "lua_fdiv_DF_load" 22 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "DF") (eq_attr "type" "fdiv,fpspc")))) - "lua_decoder01,lua_p45,lua_p0*18") + "lua_decoder01,lua_p45,lua_p0,lua_div*18") (define_insn_reservation "lua_fdiv_XF" 22 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "XF") (eq_attr "type" "fdiv,fpspc")))) - "lua_decoder0,lua_p0*22") + "lua_decoder0,lua_p0,lua_div*22") (define_insn_reservation "lua_fdiv_XF_load" 26 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "XF") (eq_attr "type" "fdiv,fpspc")))) - "lua_decoder0,lua_p45,lua_p0*22") + "lua_decoder0,lua_p45,lua_p0,lua_div*22") ;; MMX instructions. @@ -593,84 +595,84 @@ (define_insn_reservation "lua_ssediv_SF" 13 (and (eq_attr "memory" "none") (and (eq_attr "mode" "SF") (eq_attr "type" "ssediv")))) - "lua_decodern,lua_p0*13") + "lua_decodern,lua_p0,lua_div*13") (define_insn_reservation "lua_ssediv_load_SF" 17 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "SF") (eq_attr "type" "ssediv")))) - "lua_decoder01,lua_p45,lua_p0*13") + "lua_decoder01,lua_p45,lua_p0,lua_div*13") (define_insn_reservation "lua_ssediv_V4SF" 23 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "V4SF") (eq_attr "type" "ssediv")))) - "lua_decodern,lua_p0*23") + "lua_decodern,lua_p0,lua_div*23") (define_insn_reservation "lua_ssediv_load_V4SF" 27 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "V4SF") (eq_attr "type" "ssediv")))) - "lua_decoder01,lua_p45,lua_p0*23") + "lua_decoder01,lua_p45,lua_p0,lua_div*23") (define_insn_reservation "lua_ssediv_V8SF" 47 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "V8SF") (eq_attr "type" "ssediv")))) - "lua_decoder0,lua_p0*47") + "lua_decoder0,lua_p0,lua_div*47") (define_insn_reservation "lua_ssediv_load_V8SF" 51 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "V8SF") (eq_attr "type" "ssediv")))) - "lua_decoder0,lua_p45,lua_p0*47") + "lua_decoder0,lua_p45,lua_p0,lua_div*47") (define_insn_reservation "lua_ssediv_SD" 17 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "DF") (eq_attr "type" "ssediv")))) - "lua_decodern,lua_p0*17") + "lua_decodern,lua_p0,lua_div*17") (define_insn_reservation "lua_ssediv_load_SD" 21 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "DF") (eq_attr "type" "ssediv")))) - "lua_decoder01,lua_p45,lua_p0*17") + "lua_decoder01,lua_p45,lua_p0,lua_div*17") (define_insn_reservation "lua_ssediv_V2DF" 30 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "V2DF") (eq_attr "type" "ssediv")))) - "lua_decodern,lua_p0*30") + "lua_decodern,lua_p0,lua_div*30") (define_insn_reservation "lua_ssediv_load_V2DF" 34 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "V2DF") (eq_attr "type" "ssediv")))) - "lua_decoder01,lua_p45,lua_p0*30") + "lua_decoder01,lua_p45,lua_p0,lua_div*30") (define_insn_reservation "lua_ssediv_V4DF" 56 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "none") (and (eq_attr "mode" "V4DF") (eq_attr "type" "ssediv")))) - "lua_decoder0,lua_p0*56") + "lua_decoder0,lua_p0,lua_div*56") (define_insn_reservation "lua_ssediv_load_V4DF" 60 (and (eq_attr "cpu" "lujiazui") (and (eq_attr "memory" "load") (and (eq_attr "mode" "V4DF") (eq_attr "type" "ssediv")))) - "lua_decoder0,lua_p4p5,lua_p0*56") + "lua_decoder0,lua_p4p5,lua_p0,lua_div*56") (define_insn_reservation "lua_sseicvt_si" 2 -- 2.37.2
