llvm.org GIT mirror llvm / 041b3f8 lib / Target / X86 / README-SSE.txt
041b3f8

Tree @041b3f8 (Download .tar.gz)

README-SSE.txt @041b3f8raw · history · blame

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
//===---------------------------------------------------------------------===//
// Random ideas for the X86 backend: SSE-specific stuff.
//===---------------------------------------------------------------------===//

- Consider eliminating the unaligned SSE load intrinsics, replacing them with
  unaligned LLVM load instructions.

//===---------------------------------------------------------------------===//

Expand libm rounding functions inline:  Significant speedups possible.
http://gcc.gnu.org/ml/gcc-patches/2006-10/msg00909.html

//===---------------------------------------------------------------------===//

When compiled with unsafemath enabled, "main" should enable SSE DAZ mode and
other fast SSE modes.

//===---------------------------------------------------------------------===//

Think about doing i64 math in SSE regs.

//===---------------------------------------------------------------------===//

This testcase should have no SSE instructions in it, and only one load from
a constant pool:

double %test3(bool %B) {
        %C = select bool %B, double 123.412, double 523.01123123
        ret double %C
}

Currently, the select is being lowered, which prevents the dag combiner from
turning 'select (load CPI1), (load CPI2)' -> 'load (select CPI1, CPI2)'

The pattern isel got this one right.

//===---------------------------------------------------------------------===//

SSE doesn't have [mem] op= reg instructions.  If we have an SSE instruction
like this:

  X += y

and the register allocator decides to spill X, it is cheaper to emit this as:

Y += [xslot]
store Y -> [xslot]

than as:

tmp = [xslot]
tmp += y
store tmp -> [xslot]

..and this uses one fewer register (so this should be done at load folding
time, not at spiller time).  *Note* however that this can only be done
if Y is dead.  Here's a testcase:

%.str_3 = external global [15 x sbyte]          ; <[15 x sbyte]*> [#uses=0]
implementation   ; Functions:
declare void %printf(int, ...)
void %main() {
build_tree.exit:
        br label %no_exit.i7
no_exit.i7:             ; preds = %no_exit.i7, %build_tree.exit
        %tmp.0.1.0.i9 = phi double [ 0.000000e+00, %build_tree.exit ], [ %tmp.34.i18, %no_exit.i7 ]      ; <double> [#uses=1]
        %tmp.0.0.0.i10 = phi double [ 0.000000e+00, %build_tree.exit ], [ %tmp.28.i16, %no_exit.i7 ]     ; <double> [#uses=1]
        %tmp.28.i16 = add double %tmp.0.0.0.i10, 0.000000e+00
        %tmp.34.i18 = add double %tmp.0.1.0.i9, 0.000000e+00
        br bool false, label %Compute_Tree.exit23, label %no_exit.i7
Compute_Tree.exit23:            ; preds = %no_exit.i7
        tail call void (int, ...)* %printf( int 0 )
        store double %tmp.34.i18, double* null
        ret void
}

We currently emit:

.BBmain_1:
        xorpd %XMM1, %XMM1
        addsd %XMM0, %XMM1
***     movsd %XMM2, QWORD PTR [%ESP + 8]
***     addsd %XMM2, %XMM1
***     movsd QWORD PTR [%ESP + 8], %XMM2
        jmp .BBmain_1   # no_exit.i7

This is a bugpoint reduced testcase, which is why the testcase doesn't make
much sense (e.g. its an infinite loop). :)

//===---------------------------------------------------------------------===//

SSE should implement 'select_cc' using 'emulated conditional moves' that use
pcmp/pand/pandn/por to do a selection instead of a conditional branch:

double %X(double %Y, double %Z, double %A, double %B) {
        %C = setlt double %A, %B
        %z = add double %Z, 0.0    ;; select operand is not a load
        %D = select bool %C, double %Y, double %z
        ret double %D
}

We currently emit:

_X:
        subl $12, %esp
        xorpd %xmm0, %xmm0
        addsd 24(%esp), %xmm0
        movsd 32(%esp), %xmm1
        movsd 16(%esp), %xmm2
        ucomisd 40(%esp), %xmm1
        jb LBB_X_2
LBB_X_1:
        movsd %xmm0, %xmm2
LBB_X_2:
        movsd %xmm2, (%esp)
        fldl (%esp)
        addl $12, %esp
        ret

//===---------------------------------------------------------------------===//

It's not clear whether we should use pxor or xorps / xorpd to clear XMM
registers. The choice may depend on subtarget information. We should do some
more experiments on different x86 machines.

//===---------------------------------------------------------------------===//

Currently the x86 codegen isn't very good at mixing SSE and FPStack
code:

unsigned int foo(double x) { return x; }

foo:
	subl $20, %esp
	movsd 24(%esp), %xmm0
	movsd %xmm0, 8(%esp)
	fldl 8(%esp)
	fisttpll (%esp)
	movl (%esp), %eax
	addl $20, %esp
	ret

This will be solved when we go to a dynamic programming based isel.

//===---------------------------------------------------------------------===//

Lower memcpy / memset to a series of SSE 128 bit move instructions when it's
feasible.

//===---------------------------------------------------------------------===//

Teach the coalescer to commute 2-addr instructions, allowing us to eliminate
the reg-reg copy in this example:

float foo(int *x, float *y, unsigned c) {
  float res = 0.0;
  unsigned i;
  for (i = 0; i < c; i++) {
    float xx = (float)x[i];
    xx = xx * y[i];
    xx += res;
    res = xx;
  }
  return res;
}

LBB_foo_3:      # no_exit
        cvtsi2ss %XMM0, DWORD PTR [%EDX + 4*%ESI]
        mulss %XMM0, DWORD PTR [%EAX + 4*%ESI]
        addss %XMM0, %XMM1
        inc %ESI
        cmp %ESI, %ECX
****    movaps %XMM1, %XMM0
        jb LBB_foo_3    # no_exit

//===---------------------------------------------------------------------===//

Codegen:
  if (copysign(1.0, x) == copysign(1.0, y))
into:
  if (x^y & mask)
when using SSE.

//===---------------------------------------------------------------------===//

Use movhps to update upper 64-bits of a v4sf value. Also movlps on lower half
of a v4sf value.

//===---------------------------------------------------------------------===//

Better codegen for vector_shuffles like this { x, 0, 0, 0 } or { x, 0, x, 0}.
Perhaps use pxor / xorp* to clear a XMM register first?

//===---------------------------------------------------------------------===//

How to decide when to use the "floating point version" of logical ops? Here are
some code fragments:

	movaps LCPI5_5, %xmm2
	divps %xmm1, %xmm2
	mulps %xmm2, %xmm3
	mulps 8656(%ecx), %xmm3
	addps 8672(%ecx), %xmm3
	andps LCPI5_6, %xmm2
	andps LCPI5_1, %xmm3
	por %xmm2, %xmm3
	movdqa %xmm3, (%edi)

	movaps LCPI5_5, %xmm1
	divps %xmm0, %xmm1
	mulps %xmm1, %xmm3
	mulps 8656(%ecx), %xmm3
	addps 8672(%ecx), %xmm3
	andps LCPI5_6, %xmm1
	andps LCPI5_1, %xmm3
	orps %xmm1, %xmm3
	movaps %xmm3, 112(%esp)
	movaps %xmm3, (%ebx)

Due to some minor source change, the later case ended up using orps and movaps
instead of por and movdqa. Does it matter?

//===---------------------------------------------------------------------===//

X86RegisterInfo::copyRegToReg() returns X86::MOVAPSrr for VR128. Is it possible
to choose between movaps, movapd, and movdqa based on types of source and
destination?

How about andps, andpd, and pand? Do we really care about the type of the packed
elements? If not, why not always use the "ps" variants which are likely to be
shorter.

//===---------------------------------------------------------------------===//

External test Nurbs exposed some problems. Look for
__ZN15Nurbs_SSE_Cubic17TessellateSurfaceE, bb cond_next140. This is what icc
emits:

        movaps    (%edx), %xmm2                                 #59.21
        movaps    (%edx), %xmm5                                 #60.21
        movaps    (%edx), %xmm4                                 #61.21
        movaps    (%edx), %xmm3                                 #62.21
        movl      40(%ecx), %ebp                                #69.49
        shufps    $0, %xmm2, %xmm5                              #60.21
        movl      100(%esp), %ebx                               #69.20
        movl      (%ebx), %edi                                  #69.20
        imull     %ebp, %edi                                    #69.49
        addl      (%eax), %edi                                  #70.33
        shufps    $85, %xmm2, %xmm4                             #61.21
        shufps    $170, %xmm2, %xmm3                            #62.21
        shufps    $255, %xmm2, %xmm2                            #63.21
        lea       (%ebp,%ebp,2), %ebx                           #69.49
        negl      %ebx                                          #69.49
        lea       -3(%edi,%ebx), %ebx                           #70.33
        shll      $4, %ebx                                      #68.37
        addl      32(%ecx), %ebx                                #68.37
        testb     $15, %bl                                      #91.13
        jne       L_B1.24       # Prob 5%                       #91.13

This is the llvm code after instruction scheduling:

cond_next140 (0xa910740, LLVM BB @0xa90beb0):
	%reg1078 = MOV32ri -3
	%reg1079 = ADD32rm %reg1078, %reg1068, 1, %NOREG, 0
	%reg1037 = MOV32rm %reg1024, 1, %NOREG, 40
	%reg1080 = IMUL32rr %reg1079, %reg1037
	%reg1081 = MOV32rm %reg1058, 1, %NOREG, 0
	%reg1038 = LEA32r %reg1081, 1, %reg1080, -3
	%reg1036 = MOV32rm %reg1024, 1, %NOREG, 32
	%reg1082 = SHL32ri %reg1038, 4
	%reg1039 = ADD32rr %reg1036, %reg1082
	%reg1083 = MOVAPSrm %reg1059, 1, %NOREG, 0
	%reg1034 = SHUFPSrr %reg1083, %reg1083, 170
	%reg1032 = SHUFPSrr %reg1083, %reg1083, 0
	%reg1035 = SHUFPSrr %reg1083, %reg1083, 255
	%reg1033 = SHUFPSrr %reg1083, %reg1083, 85
	%reg1040 = MOV32rr %reg1039
	%reg1084 = AND32ri8 %reg1039, 15
	CMP32ri8 %reg1084, 0
	JE mbb<cond_next204,0xa914d30>

Still ok. After register allocation:

cond_next140 (0xa910740, LLVM BB @0xa90beb0):
	%EAX = MOV32ri -3
	%EDX = MOV32rm <fi#3>, 1, %NOREG, 0
	ADD32rm %EAX<def&use>, %EDX, 1, %NOREG, 0
	%EDX = MOV32rm <fi#7>, 1, %NOREG, 0
	%EDX = MOV32rm %EDX, 1, %NOREG, 40
	IMUL32rr %EAX<def&use>, %EDX
	%ESI = MOV32rm <fi#5>, 1, %NOREG, 0
	%ESI = MOV32rm %ESI, 1, %NOREG, 0
	MOV32mr <fi#4>, 1, %NOREG, 0, %ESI
	%EAX = LEA32r %ESI, 1, %EAX, -3
	%ESI = MOV32rm <fi#7>, 1, %NOREG, 0
	%ESI = MOV32rm %ESI, 1, %NOREG, 32
	%EDI = MOV32rr %EAX
	SHL32ri %EDI<def&use>, 4
	ADD32rr %EDI<def&use>, %ESI
	%XMM0 = MOVAPSrm %ECX, 1, %NOREG, 0
	%XMM1 = MOVAPSrr %XMM0
	SHUFPSrr %XMM1<def&use>, %XMM1, 170
	%XMM2 = MOVAPSrr %XMM0
	SHUFPSrr %XMM2<def&use>, %XMM2, 0
	%XMM3 = MOVAPSrr %XMM0
	SHUFPSrr %XMM3<def&use>, %XMM3, 255
	SHUFPSrr %XMM0<def&use>, %XMM0, 85
	%EBX = MOV32rr %EDI
	AND32ri8 %EBX<def&use>, 15
	CMP32ri8 %EBX, 0
	JE mbb<cond_next204,0xa914d30>

This looks really bad. The problem is shufps is a destructive opcode. Since it
appears as operand two in more than one shufps ops. It resulted in a number of
copies. Note icc also suffers from the same problem. Either the instruction
selector should select pshufd or The register allocator can made the two-address
to three-address transformation.

It also exposes some other problems. See MOV32ri -3 and the spills.

//===---------------------------------------------------------------------===//

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25500

LLVM is producing bad code.

LBB_main_4:	# cond_true44
	addps %xmm1, %xmm2
	subps %xmm3, %xmm2
	movaps (%ecx), %xmm4
	movaps %xmm2, %xmm1
	addps %xmm4, %xmm1
	addl $16, %ecx
	incl %edx
	cmpl $262144, %edx
	movaps %xmm3, %xmm2
	movaps %xmm4, %xmm3
	jne LBB_main_4	# cond_true44

There are two problems. 1) No need to two loop induction variables. We can
compare against 262144 * 16. 2) Known register coalescer issue. We should
be able eliminate one of the movaps:

	addps %xmm2, %xmm1    <=== Commute!
	subps %xmm3, %xmm1
	movaps (%ecx), %xmm4
	movaps %xmm1, %xmm1   <=== Eliminate!
	addps %xmm4, %xmm1
	addl $16, %ecx
	incl %edx
	cmpl $262144, %edx
	movaps %xmm3, %xmm2
	movaps %xmm4, %xmm3
	jne LBB_main_4	# cond_true44

//===---------------------------------------------------------------------===//

Consider:

__m128 test(float a) {
  return _mm_set_ps(0.0, 0.0, 0.0, a*a);
}

This compiles into:

movss 4(%esp), %xmm1
mulss %xmm1, %xmm1
xorps %xmm0, %xmm0
movss %xmm1, %xmm0
ret

Because mulss doesn't modify the top 3 elements, the top elements of 
xmm1 are already zero'd.  We could compile this to:

movss 4(%esp), %xmm0
mulss %xmm0, %xmm0
ret

//===---------------------------------------------------------------------===//

Here's a sick and twisted idea.  Consider code like this:

__m128 test(__m128 a) {
  float b = *(float*)&A;
  ...
  return _mm_set_ps(0.0, 0.0, 0.0, b);
}

This might compile to this code:

movaps c(%esp), %xmm1
xorps %xmm0, %xmm0
movss %xmm1, %xmm0
ret

Now consider if the ... code caused xmm1 to get spilled.  This might produce
this code:

movaps c(%esp), %xmm1
movaps %xmm1, c2(%esp)
...

xorps %xmm0, %xmm0
movaps c2(%esp), %xmm1
movss %xmm1, %xmm0
ret

However, since the reload is only used by these instructions, we could 
"fold" it into the uses, producing something like this:

movaps c(%esp), %xmm1
movaps %xmm1, c2(%esp)
...

movss c2(%esp), %xmm0
ret

... saving two instructions.

The basic idea is that a reload from a spill slot, can, if only one 4-byte 
chunk is used, bring in 3 zeros the the one element instead of 4 elements.
This can be used to simplify a variety of shuffle operations, where the
elements are fixed zeros.

//===---------------------------------------------------------------------===//

For this:

#include <emmintrin.h>
void test(__m128d *r, __m128d *A, double B) {
  *r = _mm_loadl_pd(*A, &B);
}

We generates:

	subl $12, %esp
	movsd 24(%esp), %xmm0
	movsd %xmm0, (%esp)
	movl 20(%esp), %eax
	movapd (%eax), %xmm0
	movlpd (%esp), %xmm0
	movl 16(%esp), %eax
	movapd %xmm0, (%eax)
	addl $12, %esp
	ret

icc generates:

        movl      4(%esp), %edx                                 #3.6
        movl      8(%esp), %eax                                 #3.6
        movapd    (%eax), %xmm0                                 #4.22
        movlpd    12(%esp), %xmm0                               #4.8
        movapd    %xmm0, (%edx)                                 #4.3
        ret                                                     #5.1

So icc is smart enough to know that B is in memory so it doesn't load it and
store it back to stack.

//===---------------------------------------------------------------------===//

__m128d test1( __m128d A, __m128d B) {
  return _mm_shuffle_pd(A, B, 0x3);
}

compiles to

shufpd $3, %xmm1, %xmm0

Perhaps it's better to use unpckhpd instead?

unpckhpd %xmm1, %xmm0

Don't know if unpckhpd is faster. But it is shorter.

//===---------------------------------------------------------------------===//

This code generates ugly code, probably due to costs being off or something:

void %test(float* %P, <4 x float>* %P2 ) {
        %xFloat0.688 = load float* %P
        %loadVector37.712 = load <4 x float>* %P2
        %inFloat3.713 = insertelement <4 x float> %loadVector37.712, float 0.000000e+00, uint 3
        store <4 x float> %inFloat3.713, <4 x float>* %P2
        ret void
}

Generates:

_test:
        pxor %xmm0, %xmm0
        movd %xmm0, %eax        ;; EAX = 0!
        movl 8(%esp), %ecx
        movaps (%ecx), %xmm0
        pinsrw $6, %eax, %xmm0
        shrl $16, %eax          ;; EAX = 0 again!
        pinsrw $7, %eax, %xmm0
        movaps %xmm0, (%ecx)
        ret

It would be better to generate:

_test:
        movl 8(%esp), %ecx
        movaps (%ecx), %xmm0
	xor %eax, %eax
        pinsrw $6, %eax, %xmm0
        pinsrw $7, %eax, %xmm0
        movaps %xmm0, (%ecx)
        ret

or use pxor (to make a zero vector) and shuffle (to insert it).

//===---------------------------------------------------------------------===//

Some useful information in the Apple Altivec / SSE Migration Guide:

http://developer.apple.com/documentation/Performance/Conceptual/
Accelerate_sse_migration/index.html

e.g. SSE select using and, andnot, or. Various SSE compare translations.

//===---------------------------------------------------------------------===//

Add hooks to commute some CMPP operations.

//===---------------------------------------------------------------------===//

Apply the same transformation that merged four float into a single 128-bit load
to loads from constant pool.

//===---------------------------------------------------------------------===//

Floating point max / min are commutable when -enable-unsafe-fp-path is
specified. We should turn int_x86_sse_max_ss and X86ISD::FMIN etc. into other
nodes which are selected to max / min instructions that are marked commutable.

//===---------------------------------------------------------------------===//

We should compile this:
#include <xmmintrin.h>
typedef union {
  int i[4];
  float f[4];
  __m128 v;
} vector4_t;
void swizzle (const void *a, vector4_t * b, vector4_t * c) {
  b->v = _mm_loadl_pi (b->v, (__m64 *) a);
  c->v = _mm_loadl_pi (c->v, ((__m64 *) a) + 1);
}

to:

_swizzle:
        movl    4(%esp), %eax
        movl    8(%esp), %edx
        movl    12(%esp), %ecx
        movlps  (%eax), %xmm0
        movlps  %xmm0, (%edx)
        movlps  8(%eax), %xmm0
        movlps  %xmm0, (%ecx)
        ret

not:

swizzle:
        movl 8(%esp), %eax
        movaps (%eax), %xmm0
        movl 4(%esp), %ecx
        movlps (%ecx), %xmm0
        movaps %xmm0, (%eax)
        movl 12(%esp), %eax
        movaps (%eax), %xmm0
        movlps 8(%ecx), %xmm0
        movaps %xmm0, (%eax)
        ret

//===---------------------------------------------------------------------===//

This code:

#include <emmintrin.h>
__m128i test(long long i) { return _mm_cvtsi64x_si128(i); }

Should turn into a single 'movq %rdi, %xmm0' instruction.  Instead, we 
get this (on x86-64):

_test:
	movd %rdi, %xmm1
	xorps %xmm0, %xmm0
	movsd %xmm1, %xmm0
	ret

The LLVM IR is:

target triple = "x86_64-apple-darwin8"
define <2 x i64> @test(i64 %i) {
entry:
	%tmp10 = insertelement <2 x i64> undef, i64 %i, i32 0	
	%tmp11 = insertelement <2 x i64> %tmp10, i64 0, i32 1
	ret <2 x i64> %tmp11
}

//===---------------------------------------------------------------------===//

These functions should produce the same code:

#include <emmintrin.h>

typedef long long __m128i __attribute__ ((__vector_size__ (16)));

int foo(__m128i* val) {
  return __builtin_ia32_vec_ext_v4si(*val, 1);
}
int bar(__m128i* val) {
  union vs {
    __m128i *_v;
    int* _s;
  } v = {val};
  return v._s[1];
}

We currently produce (with -m64):

_foo:
        pshufd $1, (%rdi), %xmm0
        movd %xmm0, %eax
        ret
_bar:
        movl 4(%rdi), %eax
        ret

//===---------------------------------------------------------------------===//

We should materialize vector constants like "all ones" and "signbit" with 
code like:

     cmpeqps xmm1, xmm1   ; xmm1 = all-ones

and:
     cmpeqps xmm1, xmm1   ; xmm1 = all-ones
     psrlq   xmm1, 31     ; xmm1 = all 100000000000...

instead of using a load from the constant pool.  The later is important for
ABS/NEG/copysign etc.

//===---------------------------------------------------------------------===//

"converting 64-bit constant pool entry to 32-bit not necessarily beneficial"
http://llvm.org/PR1264

For this test case:

define double @foo(double %x) {
        %y = mul double %x, 5.000000e-01
        ret double %y
}

llc -march=x86-64 currently produces a 32-bit constant pool entry and this code:

        cvtss2sd .LCPI1_0(%rip), %xmm1
        mulsd %xmm1, %xmm0

instead of just using a 64-bit constant pool entry with this:

        mulsd .LCPI1_0(%rip), %xmm0

This is due to the code in ExpandConstantFP in LegalizeDAG.cpp. It notices that
x86-64 indeed has an instruction to load a 32-bit float from memory and convert
it into a 64-bit float in a register, however it doesn't notice that this isn't
beneficial because it prevents the load from being folded into the multiply.

//===---------------------------------------------------------------------===//

In this loop:

bb49:		; preds = %bb49, %bb49.preheader
	%indvar = phi i32 [ 0, %bb49.preheader ], [ %indvar.next, %bb49 ]		; <i32> [#uses=2]
	%dp.089.0.rec = shl i32 %indvar, 3		; <i32> [#uses=2]
	%dp.089.0 = getelementptr i32* %tmp89, i32 %dp.089.0.rec		; <i32*> [#uses=1]
	%tmp5051 = bitcast i32* %dp.089.0 to <2 x i64>*		; <<2 x i64>*> [#uses=1]
	store <2 x i64> zeroinitializer, <2 x i64>* %tmp5051, align 16
	%dp.089.0.sum105 = or i32 %dp.089.0.rec, 4		; <i32> [#uses=1]
	%tmp56 = getelementptr i32* %tmp89, i32 %dp.089.0.sum105		; <i32*> [#uses=1]
	%tmp5657 = bitcast i32* %tmp56 to <2 x i64>*		; <<2 x i64>*> [#uses=1]
	store <2 x i64> zeroinitializer, <2 x i64>* %tmp5657, align 16
	%indvar.next = add i32 %indvar, 1		; <i32> [#uses=2]
	%exitcond = icmp eq i32 %indvar.next, %tmp98		; <i1> [#uses=1]
	br i1 %exitcond, label %bb72, label %bb49

we get:

LBB5_6:	# bb49.preheader
	shlw	$2, %si
	decw	%si
	movzwl	%si, %eax
	incl	%eax
	xorl	%ecx, %ecx
LBB5_7:	# bb49
	xorps	%xmm0, %xmm0            # (1)
	movaps	%xmm0, (%edx)
	movaps	%xmm0, 16(%edx)
	addl	$32, %edx
	incl	%ecx
	cmpl	%eax, %ecx
	jne	LBB4_7	# bb47

The instruction at (1) can be moved out of the main body of the loop.

//===---------------------------------------------------------------------===//

These functions:

#include <xmmintrin.h>
__m128i a;
void x(unsigned short n) {
  a = _mm_slli_epi32 (a, n);
}
void y(unsigned n) {
  a = _mm_slli_epi32 (a, n);
}

compile to ( -O3 -static -fomit-frame-pointer):
_x:
        movzwl  4(%esp), %eax
        movd    %eax, %xmm0
        movaps  _a, %xmm1
        pslld   %xmm0, %xmm1
        movaps  %xmm1, _a
        ret
_y:
        movd    4(%esp), %xmm0
        movaps  _a, %xmm1
        pslld   %xmm0, %xmm1
        movaps  %xmm1, _a
        ret

"y" looks good, but "x" does silly movzwl stuff around into a GPR.  It seems
like movd would be sufficient in both cases as the value is already zero 
extended in the 32-bit stack slot IIRC.  For signed short, it should also be
save, as a really-signed value would be undefined for pslld.


//===---------------------------------------------------------------------===//