Use subfic (val - reg) and mulli (reg * val).

In the instruction list, put /* kills xer */ for sraw, srawi, subfic;
and correct the (now unused) "addi." and "lfdu".

Change MACHOPT_F from -m3 to -m2.  This changes the code for 15 * i
from

    slwi r3,r4,4
    subfic r5,r4,0
    add r3,r3,r5

to

    mulli r3,r4,15

If the sequence "slwi subfic addi" takes 3 cycles and 12 bytes, and
mulli takes 3 cycles and 4 bytes, then mulli is better.

diff --git a/mach/powerpc/ncg/table b/mach/powerpc/ncg/table
index efdb681..82cada7 100644 (file)
--- a/mach/powerpc/ncg/table
+++ b/mach/powerpc/ncg/table
@@ -170,8 +170,10 @@ TOKENS
        SUM_RL      = { GPR reg; ADDR adr; }   4.   /* reg + lo16[adr] */
        SUM_RR      = { GPR reg1; GPR reg2; }  4.   /* reg1 + reg2 */
 
+       SUB_CR      = { INT val; GPR reg; }    4.   /* val - reg */
        SUB_RR      = { GPR reg1; GPR reg2; }  4.   /* reg1 - reg2 */
        NEG_R       = { GPR reg; }             4.   /* -reg */
+       MUL_RC      = { GPR reg; INT val; }    4.   /* reg * val */
        MUL_RR      = { GPR reg1; GPR reg2; }  4.   /* reg1 * reg2 */
        DIV_RR      = { GPR reg1; GPR reg2; }  4.   /* reg1 / reg2 signed */
        DIV_RR_U    = { GPR reg1; GPR reg2; }  4.   /* reg1 / reg2 unsigned */
@@ -280,9 +282,10 @@ SETS
        /* any integer from stack that we can easily move to GPR */
        INT_W   = SPFP + REG + CONST_STACK + SEX_B + SEX_H +
                  SUM_RIS + SUM_RC + SUM_RL + SUM_RR +
-                 SUB_RR + NEG_R + MUL_RR + DIV_RR + DIV_RR_U +
+                 SUB_CR + SUB_RR + NEG_R +
+                 MUL_RC + MUL_RR + DIV_RR + DIV_RR_U +
                  IND_ALL_B + IND_ALL_H + IND_ALL_W +
-                 FRAME_B + FRAME_H + FRAME_H_S + FRAME_W +
+                 FRAME_B + FRAME_H + FRAME_H_S + FRAME_W +
                  NOT_R + AND_RIS + AND_RC + AND_RR + ANDC_RR +
                  OR_RIS + OR_RC + OR_RR + ORC_RR +
                  XOR_RIS + XOR_RC + XOR_RR + NAND_RR + NOR_RR + EQV_RR +
@@ -307,7 +310,7 @@ INSTRUCTIONS
   cost(4, 1) /* space, time */
 
   add             GPR:wo, GPR:ro, GPR:ro.
-  addX "add."     GPR:wo, GPR:ro, GPR:ro.
+  addX "add."     GPR:wo:cc, GPR:ro, GPR:ro.
   addi            GPR:wo, GPR:ro, CONST+LABEL_LO:ro.
     li            GPR:wo, CONST:ro.
   addis           GPR:wo, GPR:ro, CONST+LABEL_HI+LABEL_HA:ro.
@@ -365,7 +368,7 @@ INSTRUCTIONS
   lbz             GPR:wo, SET_RC_B:ro cost(4, 3).
   lbzx            GPR:wo, GPR:ro, GPR:ro cost(4, 3).
   lfd             FPR+DLOCAL:wo, SET_RC_D:ro cost(4, 5).
-  lfdu            FPR:wo, IND_RC_D:ro cost(4, 5).
+  lfdu            FPR:wo, IND_RC_D:rw cost(4, 5).
   lfdx            FPR:wo, GPR:ro, GPR:ro cost(4, 5).
   lfs             FSREG+LOCAL:wo, SET_RC_W:ro cost(4, 4).
   lfsu            FSREG:wo, IND_RC_W:rw cost(4, 4).
@@ -380,6 +383,7 @@ INSTRUCTIONS
   mfcr            GPR:wo cost(4,2).
   mfspr           GPR:wo, SPR:ro cost(4, 3).
   mtspr           SPR:wo, GPR:ro cost(4, 2).
+  mulli           GPR:wo, GPR:ro, CONST:ro cost(4, 3).
   mullw           GPR:wo, GPR:ro, GPR:ro cost(4, 4).
   nand            GPR:wo, GPR:ro, GPR:ro.
   neg             GPR:wo, GPR:ro.
@@ -401,8 +405,8 @@ INSTRUCTIONS
   rlwnm           GPR:wo, GPR:ro, GPR:ro, CONST:ro, CONST:ro.
     rotlw         GPR+LOCAL:wo, GPR:ro, GPR:ro.
   slw             GPR+LOCAL:wo, GPR:ro, GPR:ro.
-  sraw            GPR+LOCAL:wo, GPR:ro, GPR:ro cost(4, 2).
-  srawi           GPR+LOCAL:wo, GPR:ro, CONST:ro cost(4, 2).
+  sraw            GPR+LOCAL:wo, GPR:ro, GPR:ro /* kills xer */ cost(4, 2).
+  srawi           GPR+LOCAL:wo, GPR:ro, CONST:ro /* kills xer */ cost(4, 2).
   srw             GPR+LOCAL:wo, GPR:ro, GPR:ro.
   stb             GPR:ro, SET_RC_B:rw cost(4, 3).
   stbx            GPR:ro, GPR:ro, GPR:ro cost(4, 3).
@@ -418,6 +422,7 @@ INSTRUCTIONS
   stwx            GPR:ro, GPR:ro, GPR:ro cost(4, 3).
   stwu            GPR:ro, IND_RC_W:rw cost(4, 3).
   subf            GPR:wo, GPR:ro, GPR:ro.
+  subfic          GPR:wo, GPR:ro, CONST:ro /* kills xer */.
   xor             GPR:wo, GPR:ro, GPR:ro.
   xori            GPR:wo, GPR:ro, CONST:ro.
   xoris           GPR:wo, GPR:ro, CONST:ro.
@@ -490,6 +495,10 @@ MOVES
 
 /* Other arithmetic */
 
+       from SUB_CR to GPR
+               /* val - reg -> subtract reg from val */
+               gen subfic %2, %1.reg, {C, %1.val}
+
        from SUB_RR to GPR
                /* reg1 - reg2 -> subtract reg2 from reg1 */
                gen subf %2, %1.reg2, %1.reg1
@@ -497,6 +506,9 @@ MOVES
        from NEG_R to GPR
                gen neg %2, %1.reg
 
+       from MUL_RC to GPR
+               gen mulli %2, %1.reg, {C, %1.val}
+
        from MUL_RR to GPR
                gen mullw %2, %1.reg1, %1.reg2
 
@@ -1471,6 +1483,8 @@ PATTERNS
                        yields {SUB_RR, %2, %1}
                with CONST2_WHEN_NEG REG
                        yields {SUM_RC, %2, 0-%1.val}
+               with REG CONST2
+                       yields {SUB_CR, %2.val, %1}
                with CONST_HI_ZR REG
                        yields {SUM_RIS, %2, his(0-%1.val)}
                with CONST_STACK-CONST2_WHEN_NEG-CONST_HI_ZR REG
@@ -1482,6 +1496,10 @@ PATTERNS
                        yields {NEG_R, %1}
 
        pat mli $1==4                      /* Multiply word (second * top) */
+               with CONST2 REG
+                       yields {MUL_RC, %2, %1.val}
+               with REG CONST2
+                       yields {MUL_RC, %1, %2.val}
                with REG REG
                        yields {MUL_RR, %2, %1}
 

diff --git a/plat/linuxppc/descr b/plat/linuxppc/descr
index 1bbb9fb..7f6f8fc 100644 (file)
--- a/plat/linuxppc/descr
+++ b/plat/linuxppc/descr
@@ -19,7 +19,7 @@ var PLATFORM=linuxppc
 var PLATFORMDIR={EM}/share/ack/{PLATFORM}
 var CPP_F=-D__unix
 var ALIGN=-a0:4 -a1:4 -a2:4 -a3:4 -b0:0x10000054
-var MACHOPT_F=-m3
+var MACHOPT_F=-m2
 var EGO_PLAT_FLAGS=-M{EM}/share/ack/ego/{ARCH}.descr
 
 # Override the setting in fe so that files compiled for linuxppc can see

diff --git a/plat/osxppc/descr b/plat/osxppc/descr
index 5f416c4..072a79d 100644 (file)
--- a/plat/osxppc/descr
+++ b/plat/osxppc/descr
@@ -19,7 +19,7 @@ var PLATFORM=osxppc
 var PLATFORMDIR={EM}/share/ack/{PLATFORM}
 var CPP_F=-D__unix
 var ALIGN=-a0:4 -a1:4 -a2:4096 -a3:4 -b0:0x129c
-var MACHOPT_F=-m3
+var MACHOPT_F=-m2
 var EGO_PLAT_FLAGS=-M{EM}/share/ack/ego/{ARCH}.descr
 
 # Override the setting in fe so that files compiled for osxppc can see