[ia32] Add rcpps, rsqrtps, cvtdq2ps, cvttps2dq

test/cctest/test-disasm-ia32.cc

 // Copyright 2011 the V8 project authors. All rights reserved.
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions are
 // met:
 //
 //     * Redistributions of source code must retain the above copyright
 //       notice, this list of conditions and the following disclaimer.
 //     * Redistributions in binary form must reproduce the above
 //       copyright notice, this list of conditions and the following
 //       disclaimer in the documentation and/or other materials provided
@@ skipping 410 matching lines @@
     __ subps(xmm1, xmm0);
     __ subps(xmm1, Operand(ebx, ecx, times_4, 10000));
     __ mulps(xmm1, xmm0);
     __ mulps(xmm1, Operand(ebx, ecx, times_4, 10000));
     __ divps(xmm1, xmm0);
     __ divps(xmm1, Operand(ebx, ecx, times_4, 10000));
     __ minps(xmm1, xmm0);
     __ minps(xmm1, Operand(ebx, ecx, times_4, 10000));
     __ maxps(xmm1, xmm0);
     __ maxps(xmm1, Operand(ebx, ecx, times_4, 10000));
+    __ rcpps(xmm1, xmm0);
+    __ rcpps(xmm1, Operand(ebx, ecx, times_4, 10000));
+    __ rsqrtps(xmm1, xmm0);
+    __ rsqrtps(xmm1, Operand(ebx, ecx, times_4, 10000));
 
     __ cmpeqps(xmm5, xmm1);
     __ cmpeqps(xmm5, Operand(ebx, ecx, times_4, 10000));
     __ cmpltps(xmm5, xmm1);
     __ cmpltps(xmm5, Operand(ebx, ecx, times_4, 10000));
     __ cmpleps(xmm5, xmm1);
     __ cmpleps(xmm5, Operand(ebx, ecx, times_4, 10000));
     __ cmpneqps(xmm5, xmm1);
     __ cmpneqps(xmm5, Operand(ebx, ecx, times_4, 10000));
 
     __ ucomiss(xmm0, xmm1);
     __ ucomiss(xmm0, Operand(ebx, ecx, times_4, 10000));
   }
   {
     __ cvttss2si(edx, Operand(ebx, ecx, times_4, 10000));
     __ cvtsi2sd(xmm1, Operand(ebx, ecx, times_4, 10000));
     __ cvtss2sd(xmm1, Operand(ebx, ecx, times_4, 10000));
     __ cvtss2sd(xmm1, xmm0);
+    __ cvtdq2ps(xmm1, xmm0);
+    __ cvtdq2ps(xmm1, Operand(ebx, ecx, times_4, 10000));
+    __ cvttps2dq(xmm1, xmm0);
+    __ cvttps2dq(xmm1, Operand(ebx, ecx, times_4, 10000));
     __ movsd(xmm1, Operand(ebx, ecx, times_4, 10000));
     __ movsd(Operand(ebx, ecx, times_4, 10000), xmm1);
     // 128 bit move instructions.
     __ movdqa(xmm0, Operand(ebx, ecx, times_4, 10000));
     __ movdqa(Operand(ebx, ecx, times_4, 10000), xmm0);
     __ movdqu(xmm0, Operand(ebx, ecx, times_4, 10000));
     __ movdqu(Operand(ebx, ecx, times_4, 10000), xmm0);
 
     __ addsd(xmm1, xmm0);
     __ addsd(xmm1, Operand(ebx, ecx, times_4, 10000));
@@ skipping 109 matching lines @@
       __ vmulps(xmm0, xmm1, xmm2);
       __ vmulps(xmm0, xmm1, Operand(ebx, ecx, times_4, 10000));
       __ vsubps(xmm0, xmm1, xmm2);
       __ vsubps(xmm0, xmm1, Operand(ebx, ecx, times_4, 10000));
       __ vminps(xmm0, xmm1, xmm2);
       __ vminps(xmm0, xmm1, Operand(ebx, ecx, times_4, 10000));
       __ vdivps(xmm0, xmm1, xmm2);
       __ vdivps(xmm0, xmm1, Operand(ebx, ecx, times_4, 10000));
       __ vmaxps(xmm0, xmm1, xmm2);
       __ vmaxps(xmm0, xmm1, Operand(ebx, ecx, times_4, 10000));
+      __ vrcpps(xmm1, xmm0);
+      __ vrcpps(xmm1, Operand(ebx, ecx, times_4, 10000));
+      __ vrsqrtps(xmm1, xmm0);
+      __ vrsqrtps(xmm1, Operand(ebx, ecx, times_4, 10000));
 
       __ vcmpeqps(xmm5, xmm4, xmm1);
       __ vcmpeqps(xmm5, xmm4, Operand(ebx, ecx, times_4, 10000));
       __ vcmpltps(xmm5, xmm4, xmm1);
       __ vcmpltps(xmm5, xmm4, Operand(ebx, ecx, times_4, 10000));
       __ vcmpleps(xmm5, xmm4, xmm1);
       __ vcmpleps(xmm5, xmm4, Operand(ebx, ecx, times_4, 10000));
       __ vcmpneqps(xmm5, xmm4, xmm1);
       __ vcmpneqps(xmm5, xmm4, Operand(ebx, ecx, times_4, 10000));
 
@@ skipping 13 matching lines @@
       __ vdivpd(xmm0, xmm1, Operand(ebx, ecx, times_4, 10000));
       __ vmaxpd(xmm0, xmm1, xmm2);
       __ vmaxpd(xmm0, xmm1, Operand(ebx, ecx, times_4, 10000));
 
       __ vpsllw(xmm0, xmm7, 21);
       __ vpslld(xmm0, xmm7, 21);
       __ vpsrlw(xmm0, xmm7, 21);
       __ vpsrld(xmm0, xmm7, 21);
       __ vpsraw(xmm0, xmm7, 21);
       __ vpsrad(xmm0, xmm7, 21);
+
+      __ vcvtdq2ps(xmm1, xmm0);
+      __ vcvtdq2ps(xmm1, Operand(ebx, ecx, times_4, 10000));
+      __ vcvttps2dq(xmm1, xmm0);
+      __ vcvttps2dq(xmm1, Operand(ebx, ecx, times_4, 10000));
+
 #define EMIT_SSE2_AVXINSTR(instruction, notUsed1, notUsed2, notUsed3) \
   __ v##instruction(xmm7, xmm5, xmm1);                                \
   __ v##instruction(xmm7, xmm5, Operand(edx, 4));
 
       SSE2_INSTRUCTION_LIST(EMIT_SSE2_AVXINSTR)
 #undef EMIT_SSE2_AVXINSTR
 
 #define EMIT_SSE4_AVXINSTR(instruction, notUsed1, notUsed2, notUsed3, \
                            notUsed4)                                  \
   __ v##instruction(xmm7, xmm5, xmm1);                                \
@@ skipping 167 matching lines @@
 #ifdef OBJECT_PRINT
   OFStream os(stdout);
   code->Print(os);
   byte* begin = code->instruction_start();
   byte* end = begin + code->instruction_size();
   disasm::Disassembler::Disassemble(stdout, begin, end);
 #endif
 }
 
 #undef __