1 Target Independent Opportunities:
3 //===---------------------------------------------------------------------===//
5 With the recent changes to make the implicit def/use set explicit in
6 machineinstrs, we should change the target descriptions for 'call' instructions
7 so that the .td files don't list all the call-clobbered registers as implicit
8 defs. Instead, these should be added by the code generator (e.g. on the dag).
10 This has a number of uses:
12 1. PPC32/64 and X86 32/64 can avoid having multiple copies of call instructions
13 for their different impdef sets.
14 2. Targets with multiple calling convs (e.g. x86) which have different clobber
15 sets don't need copies of call instructions.
16 3. 'Interprocedural register allocation' can be done to reduce the clobber sets
19 //===---------------------------------------------------------------------===//
21 We should recognized various "overflow detection" idioms and translate them into
22 llvm.uadd.with.overflow and similar intrinsics. Here is a multiply idiom:
24 unsigned int mul(unsigned int a,unsigned int b) {
25 if ((unsigned long long)a*b>0xffffffff)
30 The legalization code for mul-with-overflow needs to be made more robust before
31 this can be implemented though.
33 //===---------------------------------------------------------------------===//
35 Get the C front-end to expand hypot(x,y) -> llvm.sqrt(x*x+y*y) when errno and
36 precision don't matter (ffastmath). Misc/mandel will like this. :) This isn't
37 safe in general, even on darwin. See the libm implementation of hypot for
38 examples (which special case when x/y are exactly zero to get signed zeros etc
41 //===---------------------------------------------------------------------===//
43 On targets with expensive 64-bit multiply, we could LSR this:
50 for (i = ...; ++i, tmp+=tmp)
53 This would be a win on ppc32, but not x86 or ppc64.
55 //===---------------------------------------------------------------------===//
57 Shrink: (setlt (loadi32 P), 0) -> (setlt (loadi8 Phi), 0)
59 //===---------------------------------------------------------------------===//
61 Reassociate should turn things like:
63 int factorial(int X) {
64 return X*X*X*X*X*X*X*X;
67 into llvm.powi calls, allowing the code generator to produce balanced
70 First, the intrinsic needs to be extended to support integers, and second the
71 code generator needs to be enhanced to lower these to multiplication trees.
73 //===---------------------------------------------------------------------===//
75 Interesting? testcase for add/shift/mul reassoc:
77 int bar(int x, int y) {
78 return x*x*x+y+x*x*x*x*x*y*y*y*y;
80 int foo(int z, int n) {
81 return bar(z, n) + bar(2*z, 2*n);
84 This is blocked on not handling X*X*X -> powi(X, 3) (see note above). The issue
85 is that we end up getting t = 2*X s = t*t and don't turn this into 4*X*X,
86 which is the same number of multiplies and is canonical, because the 2*X has
87 multiple uses. Here's a simple example:
89 define i32 @test15(i32 %X1) {
90 %B = mul i32 %X1, 47 ; X1*47
96 //===---------------------------------------------------------------------===//
98 Reassociate should handle the example in GCC PR16157:
100 extern int a0, a1, a2, a3, a4; extern int b0, b1, b2, b3, b4;
101 void f () { /* this can be optimized to four additions... */
102 b4 = a4 + a3 + a2 + a1 + a0;
103 b3 = a3 + a2 + a1 + a0;
108 This requires reassociating to forms of expressions that are already available,
109 something that reassoc doesn't think about yet.
112 //===---------------------------------------------------------------------===//
114 This function: (derived from GCC PR19988)
115 double foo(double x, double y) {
116 return ((x + 0.1234 * y) * (x + -0.1234 * y));
122 mulsd LCPI1_1(%rip), %xmm1
123 mulsd LCPI1_0(%rip), %xmm2
130 Reassociate should be able to turn it into:
132 double foo(double x, double y) {
133 return ((x + 0.1234 * y) * (x - 0.1234 * y));
136 Which allows the multiply by constant to be CSE'd, producing:
139 mulsd LCPI1_0(%rip), %xmm1
146 This doesn't need -ffast-math support at all. This is particularly bad because
147 the llvm-gcc frontend is canonicalizing the later into the former, but clang
148 doesn't have this problem.
150 //===---------------------------------------------------------------------===//
152 These two functions should generate the same code on big-endian systems:
154 int g(int *j,int *l) { return memcmp(j,l,4); }
155 int h(int *j, int *l) { return *j - *l; }
157 this could be done in SelectionDAGISel.cpp, along with other special cases,
160 //===---------------------------------------------------------------------===//
162 It would be nice to revert this patch:
163 http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20060213/031986.html
165 And teach the dag combiner enough to simplify the code expanded before
166 legalize. It seems plausible that this knowledge would let it simplify other
169 //===---------------------------------------------------------------------===//
171 For vector types, TargetData.cpp::getTypeInfo() returns alignment that is equal
172 to the type size. It works but can be overly conservative as the alignment of
173 specific vector types are target dependent.
175 //===---------------------------------------------------------------------===//
177 We should produce an unaligned load from code like this:
179 v4sf example(float *P) {
180 return (v4sf){P[0], P[1], P[2], P[3] };
183 //===---------------------------------------------------------------------===//
185 Add support for conditional increments, and other related patterns. Instead
190 je LBB16_2 #cond_next
201 //===---------------------------------------------------------------------===//
203 Combine: a = sin(x), b = cos(x) into a,b = sincos(x).
205 Expand these to calls of sin/cos and stores:
206 double sincos(double x, double *sin, double *cos);
207 float sincosf(float x, float *sin, float *cos);
208 long double sincosl(long double x, long double *sin, long double *cos);
210 Doing so could allow SROA of the destination pointers. See also:
211 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17687
213 This is now easily doable with MRVs. We could even make an intrinsic for this
214 if anyone cared enough about sincos.
216 //===---------------------------------------------------------------------===//
218 quantum_sigma_x in 462.libquantum contains the following loop:
220 for(i=0; i<reg->size; i++)
222 /* Flip the target bit of each basis state */
223 reg->node[i].state ^= ((MAX_UNSIGNED) 1 << target);
226 Where MAX_UNSIGNED/state is a 64-bit int. On a 32-bit platform it would be just
227 so cool to turn it into something like:
229 long long Res = ((MAX_UNSIGNED) 1 << target);
231 for(i=0; i<reg->size; i++)
232 reg->node[i].state ^= Res & 0xFFFFFFFFULL;
234 for(i=0; i<reg->size; i++)
235 reg->node[i].state ^= Res & 0xFFFFFFFF00000000ULL
238 ... which would only do one 32-bit XOR per loop iteration instead of two.
240 It would also be nice to recognize the reg->size doesn't alias reg->node[i], but
243 //===---------------------------------------------------------------------===//
245 This isn't recognized as bswap by instcombine (yes, it really is bswap):
247 unsigned long reverse(unsigned v) {
249 t = v ^ ((v << 16) | (v >> 16));
251 v = (v << 24) | (v >> 8);
255 //===---------------------------------------------------------------------===//
259 These idioms should be recognized as popcount (see PR1488):
261 unsigned countbits_slow(unsigned v) {
263 for (c = 0; v; v >>= 1)
267 unsigned countbits_fast(unsigned v){
270 v &= v - 1; // clear the least significant bit set
274 BITBOARD = unsigned long long
275 int PopCnt(register BITBOARD a) {
283 unsigned int popcount(unsigned int input) {
284 unsigned int count = 0;
285 for (unsigned int i = 0; i < 4 * 8; i++)
286 count += (input >> i) & i;
290 This sort of thing should be added to the loop idiom pass.
292 //===---------------------------------------------------------------------===//
294 These should turn into single 16-bit (unaligned?) loads on little/big endian
297 unsigned short read_16_le(const unsigned char *adr) {
298 return adr[0] | (adr[1] << 8);
300 unsigned short read_16_be(const unsigned char *adr) {
301 return (adr[0] << 8) | adr[1];
304 //===---------------------------------------------------------------------===//
306 -instcombine should handle this transform:
307 icmp pred (sdiv X / C1 ), C2
308 when X, C1, and C2 are unsigned. Similarly for udiv and signed operands.
310 Currently InstCombine avoids this transform but will do it when the signs of
311 the operands and the sign of the divide match. See the FIXME in
312 InstructionCombining.cpp in the visitSetCondInst method after the switch case
313 for Instruction::UDiv (around line 4447) for more details.
315 The SingleSource/Benchmarks/Shootout-C++/hash and hash2 tests have examples of
318 //===---------------------------------------------------------------------===//
322 SingleSource/Benchmarks/Misc/dt.c shows several interesting optimization
323 opportunities in its double_array_divs_variable function: it needs loop
324 interchange, memory promotion (which LICM already does), vectorization and
325 variable trip count loop unrolling (since it has a constant trip count). ICC
326 apparently produces this very nice code with -ffast-math:
328 ..B1.70: # Preds ..B1.70 ..B1.69
329 mulpd %xmm0, %xmm1 #108.2
330 mulpd %xmm0, %xmm1 #108.2
331 mulpd %xmm0, %xmm1 #108.2
332 mulpd %xmm0, %xmm1 #108.2
334 cmpl $131072, %edx #108.2
335 jb ..B1.70 # Prob 99% #108.2
337 It would be better to count down to zero, but this is a lot better than what we
340 //===---------------------------------------------------------------------===//
344 typedef unsigned U32;
345 typedef unsigned long long U64;
346 int test (U32 *inst, U64 *regs) {
349 int r1 = (temp >> 20) & 0xf;
350 int b2 = (temp >> 16) & 0xf;
351 effective_addr2 = temp & 0xfff;
352 if (b2) effective_addr2 += regs[b2];
353 b2 = (temp >> 12) & 0xf;
354 if (b2) effective_addr2 += regs[b2];
355 effective_addr2 &= regs[4];
356 if ((effective_addr2 & 3) == 0)
361 Note that only the low 2 bits of effective_addr2 are used. On 32-bit systems,
362 we don't eliminate the computation of the top half of effective_addr2 because
363 we don't have whole-function selection dags. On x86, this means we use one
364 extra register for the function when effective_addr2 is declared as U64 than
365 when it is declared U32.
367 PHI Slicing could be extended to do this.
369 //===---------------------------------------------------------------------===//
371 LSR should know what GPR types a target has from TargetData. This code:
373 volatile short X, Y; // globals
377 for (i = 0; i < N; i++) { X = i; Y = i*4; }
380 produces two near identical IV's (after promotion) on PPC/ARM:
390 add r2, r2, #1 <- [0,+,1]
391 sub r0, r0, #1 <- [0,-,1]
395 LSR should reuse the "+" IV for the exit test.
397 //===---------------------------------------------------------------------===//
399 Tail call elim should be more aggressive, checking to see if the call is
400 followed by an uncond branch to an exit block.
402 ; This testcase is due to tail-duplication not wanting to copy the return
403 ; instruction into the terminating blocks because there was other code
404 ; optimized out of the function after the taildup happened.
405 ; RUN: llvm-as < %s | opt -tailcallelim | llvm-dis | not grep call
407 define i32 @t4(i32 %a) {
409 %tmp.1 = and i32 %a, 1 ; <i32> [#uses=1]
410 %tmp.2 = icmp ne i32 %tmp.1, 0 ; <i1> [#uses=1]
411 br i1 %tmp.2, label %then.0, label %else.0
413 then.0: ; preds = %entry
414 %tmp.5 = add i32 %a, -1 ; <i32> [#uses=1]
415 %tmp.3 = call i32 @t4( i32 %tmp.5 ) ; <i32> [#uses=1]
418 else.0: ; preds = %entry
419 %tmp.7 = icmp ne i32 %a, 0 ; <i1> [#uses=1]
420 br i1 %tmp.7, label %then.1, label %return
422 then.1: ; preds = %else.0
423 %tmp.11 = add i32 %a, -2 ; <i32> [#uses=1]
424 %tmp.9 = call i32 @t4( i32 %tmp.11 ) ; <i32> [#uses=1]
427 return: ; preds = %then.1, %else.0, %then.0
428 %result.0 = phi i32 [ 0, %else.0 ], [ %tmp.3, %then.0 ],
433 //===---------------------------------------------------------------------===//
435 Tail recursion elimination should handle:
440 return 2 * pow2m1 (n - 1) + 1;
443 Also, multiplies can be turned into SHL's, so they should be handled as if
444 they were associative. "return foo() << 1" can be tail recursion eliminated.
446 //===---------------------------------------------------------------------===//
448 Argument promotion should promote arguments for recursive functions, like
451 ; RUN: llvm-as < %s | opt -argpromotion | llvm-dis | grep x.val
453 define internal i32 @foo(i32* %x) {
455 %tmp = load i32* %x ; <i32> [#uses=0]
456 %tmp.foo = call i32 @foo( i32* %x ) ; <i32> [#uses=1]
460 define i32 @bar(i32* %x) {
462 %tmp3 = call i32 @foo( i32* %x ) ; <i32> [#uses=1]
466 //===---------------------------------------------------------------------===//
468 We should investigate an instruction sinking pass. Consider this silly
484 je LBB1_2 # cond_true
492 The PIC base computation (call+popl) is only used on one path through the
493 code, but is currently always computed in the entry block. It would be
494 better to sink the picbase computation down into the block for the
495 assertion, as it is the only one that uses it. This happens for a lot of
496 code with early outs.
498 Another example is loads of arguments, which are usually emitted into the
499 entry block on targets like x86. If not used in all paths through a
500 function, they should be sunk into the ones that do.
502 In this case, whole-function-isel would also handle this.
504 //===---------------------------------------------------------------------===//
506 Investigate lowering of sparse switch statements into perfect hash tables:
507 http://burtleburtle.net/bob/hash/perfect.html
509 //===---------------------------------------------------------------------===//
511 We should turn things like "load+fabs+store" and "load+fneg+store" into the
512 corresponding integer operations. On a yonah, this loop:
517 for (b = 0; b < 10000000; b++)
518 for (i = 0; i < 256; i++)
522 is twice as slow as this loop:
527 for (b = 0; b < 10000000; b++)
528 for (i = 0; i < 256; i++)
529 a[i] ^= (1ULL << 63);
532 and I suspect other processors are similar. On X86 in particular this is a
533 big win because doing this with integers allows the use of read/modify/write
536 //===---------------------------------------------------------------------===//
538 DAG Combiner should try to combine small loads into larger loads when
539 profitable. For example, we compile this C++ example:
541 struct THotKey { short Key; bool Control; bool Shift; bool Alt; };
542 extern THotKey m_HotKey;
543 THotKey GetHotKey () { return m_HotKey; }
545 into (-m64 -O3 -fno-exceptions -static -fomit-frame-pointer):
547 __Z9GetHotKeyv: ## @_Z9GetHotKeyv
548 movq _m_HotKey@GOTPCREL(%rip), %rax
561 //===---------------------------------------------------------------------===//
563 We should add an FRINT node to the DAG to model targets that have legal
564 implementations of ceil/floor/rint.
566 //===---------------------------------------------------------------------===//
571 long long input[8] = {1,0,1,0,1,0,1,0};
575 Clang compiles this into:
577 call void @llvm.memset.p0i8.i64(i8* %tmp, i8 0, i64 64, i32 16, i1 false)
578 %0 = getelementptr [8 x i64]* %input, i64 0, i64 0
579 store i64 1, i64* %0, align 16
580 %1 = getelementptr [8 x i64]* %input, i64 0, i64 2
581 store i64 1, i64* %1, align 16
582 %2 = getelementptr [8 x i64]* %input, i64 0, i64 4
583 store i64 1, i64* %2, align 16
584 %3 = getelementptr [8 x i64]* %input, i64 0, i64 6
585 store i64 1, i64* %3, align 16
587 Which gets codegen'd into:
590 movaps %xmm0, -16(%rbp)
591 movaps %xmm0, -32(%rbp)
592 movaps %xmm0, -48(%rbp)
593 movaps %xmm0, -64(%rbp)
599 It would be better to have 4 movq's of 0 instead of the movaps's.
601 //===---------------------------------------------------------------------===//
603 http://llvm.org/PR717:
605 The following code should compile into "ret int undef". Instead, LLVM
606 produces "ret int 0":
615 //===---------------------------------------------------------------------===//
617 The loop unroller should partially unroll loops (instead of peeling them)
618 when code growth isn't too bad and when an unroll count allows simplification
619 of some code within the loop. One trivial example is:
625 for ( nLoop = 0; nLoop < 1000; nLoop++ ) {
634 Unrolling by 2 would eliminate the '&1' in both copies, leading to a net
635 reduction in code size. The resultant code would then also be suitable for
636 exit value computation.
638 //===---------------------------------------------------------------------===//
640 We miss a bunch of rotate opportunities on various targets, including ppc, x86,
641 etc. On X86, we miss a bunch of 'rotate by variable' cases because the rotate
642 matching code in dag combine doesn't look through truncates aggressively
643 enough. Here are some testcases reduces from GCC PR17886:
645 unsigned long long f5(unsigned long long x, unsigned long long y) {
646 return (x << 8) | ((y >> 48) & 0xffull);
648 unsigned long long f6(unsigned long long x, unsigned long long y, int z) {
651 return (x << 8) | ((y >> 48) & 0xffull);
653 return (x << 16) | ((y >> 40) & 0xffffull);
655 return (x << 24) | ((y >> 32) & 0xffffffull);
657 return (x << 32) | ((y >> 24) & 0xffffffffull);
659 return (x << 40) | ((y >> 16) & 0xffffffffffull);
663 //===---------------------------------------------------------------------===//
665 This (and similar related idioms):
667 unsigned int foo(unsigned char i) {
668 return i | (i<<8) | (i<<16) | (i<<24);
673 define i32 @foo(i8 zeroext %i) nounwind readnone ssp noredzone {
675 %conv = zext i8 %i to i32
676 %shl = shl i32 %conv, 8
677 %shl5 = shl i32 %conv, 16
678 %shl9 = shl i32 %conv, 24
679 %or = or i32 %shl9, %conv
680 %or6 = or i32 %or, %shl5
681 %or10 = or i32 %or6, %shl
685 it would be better as:
687 unsigned int bar(unsigned char i) {
688 unsigned int j=i | (i << 8);
694 define i32 @bar(i8 zeroext %i) nounwind readnone ssp noredzone {
696 %conv = zext i8 %i to i32
697 %shl = shl i32 %conv, 8
698 %or = or i32 %shl, %conv
699 %shl5 = shl i32 %or, 16
700 %or6 = or i32 %shl5, %or
704 or even i*0x01010101, depending on the speed of the multiplier. The best way to
705 handle this is to canonicalize it to a multiply in IR and have codegen handle
706 lowering multiplies to shifts on cpus where shifts are faster.
708 //===---------------------------------------------------------------------===//
710 We do a number of simplifications in simplify libcalls to strength reduce
711 standard library functions, but we don't currently merge them together. For
712 example, it is useful to merge memcpy(a,b,strlen(b)) -> strcpy. This can only
713 be done safely if "b" isn't modified between the strlen and memcpy of course.
715 //===---------------------------------------------------------------------===//
717 We compile this program: (from GCC PR11680)
718 http://gcc.gnu.org/bugzilla/attachment.cgi?id=4487
720 Into code that runs the same speed in fast/slow modes, but both modes run 2x
721 slower than when compile with GCC (either 4.0 or 4.2):
723 $ llvm-g++ perf.cpp -O3 -fno-exceptions
725 1.821u 0.003s 0:01.82 100.0% 0+0k 0+0io 0pf+0w
727 $ g++ perf.cpp -O3 -fno-exceptions
729 0.821u 0.001s 0:00.82 100.0% 0+0k 0+0io 0pf+0w
731 It looks like we are making the same inlining decisions, so this may be raw
732 codegen badness or something else (haven't investigated).
734 //===---------------------------------------------------------------------===//
736 Divisibility by constant can be simplified (according to GCC PR12849) from
737 being a mulhi to being a mul lo (cheaper). Testcase:
739 void bar(unsigned n) {
744 This is equivalent to the following, where 2863311531 is the multiplicative
745 inverse of 3, and 1431655766 is ((2^32)-1)/3+1:
746 void bar(unsigned n) {
747 if (n * 2863311531U < 1431655766U)
751 The same transformation can work with an even modulo with the addition of a
752 rotate: rotate the result of the multiply to the right by the number of bits
753 which need to be zero for the condition to be true, and shrink the compare RHS
754 by the same amount. Unless the target supports rotates, though, that
755 transformation probably isn't worthwhile.
757 The transformation can also easily be made to work with non-zero equality
758 comparisons: just transform, for example, "n % 3 == 1" to "(n-1) % 3 == 0".
760 //===---------------------------------------------------------------------===//
762 Better mod/ref analysis for scanf would allow us to eliminate the vtable and a
763 bunch of other stuff from this example (see PR1604):
773 std::scanf("%d", &t.val);
774 std::printf("%d\n", t.val);
777 //===---------------------------------------------------------------------===//
779 These functions perform the same computation, but produce different assembly.
781 define i8 @select(i8 %x) readnone nounwind {
782 %A = icmp ult i8 %x, 250
783 %B = select i1 %A, i8 0, i8 1
787 define i8 @addshr(i8 %x) readnone nounwind {
788 %A = zext i8 %x to i9
789 %B = add i9 %A, 6 ;; 256 - 250 == 6
791 %D = trunc i9 %C to i8
795 //===---------------------------------------------------------------------===//
799 f (unsigned long a, unsigned long b, unsigned long c)
801 return ((a & (c - 1)) != 0) || ((b & (c - 1)) != 0);
804 f (unsigned long a, unsigned long b, unsigned long c)
806 return ((a & (c - 1)) != 0) | ((b & (c - 1)) != 0);
808 Both should combine to ((a|b) & (c-1)) != 0. Currently not optimized with
809 "clang -emit-llvm-bc | opt -std-compile-opts".
811 //===---------------------------------------------------------------------===//
814 #define PMD_MASK (~((1UL << 23) - 1))
815 void clear_pmd_range(unsigned long start, unsigned long end)
817 if (!(start & ~PMD_MASK) && !(end & ~PMD_MASK))
820 The expression should optimize to something like
821 "!((start|end)&~PMD_MASK). Currently not optimized with "clang
822 -emit-llvm-bc | opt -std-compile-opts".
824 //===---------------------------------------------------------------------===//
826 unsigned int f(unsigned int i, unsigned int n) {++i; if (i == n) ++i; return
828 unsigned int f2(unsigned int i, unsigned int n) {++i; i += i == n; return i;}
829 These should combine to the same thing. Currently, the first function
830 produces better code on X86.
832 //===---------------------------------------------------------------------===//
835 #define abs(x) x>0?x:-x
838 return (abs(x)) >= 0;
840 This should optimize to x == INT_MIN. (With -fwrapv.) Currently not
841 optimized with "clang -emit-llvm-bc | opt -std-compile-opts".
843 //===---------------------------------------------------------------------===//
847 rotate_cst (unsigned int a)
849 a = (a << 10) | (a >> 22);
854 minus_cst (unsigned int a)
863 mask_gt (unsigned int a)
865 /* This is equivalent to a > 15. */
870 rshift_gt (unsigned int a)
872 /* This is equivalent to a > 23. */
876 All should simplify to a single comparison. All of these are
877 currently not optimized with "clang -emit-llvm-bc | opt
880 //===---------------------------------------------------------------------===//
883 int c(int* x) {return (char*)x+2 == (char*)x;}
884 Should combine to 0. Currently not optimized with "clang
885 -emit-llvm-bc | opt -std-compile-opts" (although llc can optimize it).
887 //===---------------------------------------------------------------------===//
889 int a(unsigned b) {return ((b << 31) | (b << 30)) >> 31;}
890 Should be combined to "((b >> 1) | b) & 1". Currently not optimized
891 with "clang -emit-llvm-bc | opt -std-compile-opts".
893 //===---------------------------------------------------------------------===//
895 unsigned a(unsigned x, unsigned y) { return x | (y & 1) | (y & 2);}
896 Should combine to "x | (y & 3)". Currently not optimized with "clang
897 -emit-llvm-bc | opt -std-compile-opts".
899 //===---------------------------------------------------------------------===//
901 int a(int a, int b, int c) {return (~a & c) | ((c|a) & b);}
902 Should fold to "(~a & c) | (a & b)". Currently not optimized with
903 "clang -emit-llvm-bc | opt -std-compile-opts".
905 //===---------------------------------------------------------------------===//
907 int a(int a,int b) {return (~(a|b))|a;}
908 Should fold to "a|~b". Currently not optimized with "clang
909 -emit-llvm-bc | opt -std-compile-opts".
911 //===---------------------------------------------------------------------===//
913 int a(int a, int b) {return (a&&b) || (a&&!b);}
914 Should fold to "a". Currently not optimized with "clang -emit-llvm-bc
915 | opt -std-compile-opts".
917 //===---------------------------------------------------------------------===//
919 int a(int a, int b, int c) {return (a&&b) || (!a&&c);}
920 Should fold to "a ? b : c", or at least something sane. Currently not
921 optimized with "clang -emit-llvm-bc | opt -std-compile-opts".
923 //===---------------------------------------------------------------------===//
925 int a(int a, int b, int c) {return (a&&b) || (a&&c) || (a&&b&&c);}
926 Should fold to a && (b || c). Currently not optimized with "clang
927 -emit-llvm-bc | opt -std-compile-opts".
929 //===---------------------------------------------------------------------===//
931 int a(int x) {return x | ((x & 8) ^ 8);}
932 Should combine to x | 8. Currently not optimized with "clang
933 -emit-llvm-bc | opt -std-compile-opts".
935 //===---------------------------------------------------------------------===//
937 int a(int x) {return x ^ ((x & 8) ^ 8);}
938 Should also combine to x | 8. Currently not optimized with "clang
939 -emit-llvm-bc | opt -std-compile-opts".
941 //===---------------------------------------------------------------------===//
943 int a(int x) {return ((x | -9) ^ 8) & x;}
944 Should combine to x & -9. Currently not optimized with "clang
945 -emit-llvm-bc | opt -std-compile-opts".
947 //===---------------------------------------------------------------------===//
949 unsigned a(unsigned a) {return a * 0x11111111 >> 28 & 1;}
950 Should combine to "a * 0x88888888 >> 31". Currently not optimized
951 with "clang -emit-llvm-bc | opt -std-compile-opts".
953 //===---------------------------------------------------------------------===//
955 unsigned a(char* x) {if ((*x & 32) == 0) return b();}
956 There's an unnecessary zext in the generated code with "clang
957 -emit-llvm-bc | opt -std-compile-opts".
959 //===---------------------------------------------------------------------===//
961 unsigned a(unsigned long long x) {return 40 * (x >> 1);}
962 Should combine to "20 * (((unsigned)x) & -2)". Currently not
963 optimized with "clang -emit-llvm-bc | opt -std-compile-opts".
965 //===---------------------------------------------------------------------===//
967 This was noticed in the entryblock for grokdeclarator in 403.gcc:
969 %tmp = icmp eq i32 %decl_context, 4
970 %decl_context_addr.0 = select i1 %tmp, i32 3, i32 %decl_context
971 %tmp1 = icmp eq i32 %decl_context_addr.0, 1
972 %decl_context_addr.1 = select i1 %tmp1, i32 0, i32 %decl_context_addr.0
974 tmp1 should be simplified to something like:
975 (!tmp || decl_context == 1)
977 This allows recursive simplifications, tmp1 is used all over the place in
978 the function, e.g. by:
980 %tmp23 = icmp eq i32 %decl_context_addr.1, 0 ; <i1> [#uses=1]
981 %tmp24 = xor i1 %tmp1, true ; <i1> [#uses=1]
982 %or.cond8 = and i1 %tmp23, %tmp24 ; <i1> [#uses=1]
986 //===---------------------------------------------------------------------===//
990 Store sinking: This code:
992 void f (int n, int *cond, int *res) {
995 for (i = 0; i < n; i++)
997 *res ^= 234; /* (*) */
1000 On this function GVN hoists the fully redundant value of *res, but nothing
1001 moves the store out. This gives us this code:
1003 bb: ; preds = %bb2, %entry
1004 %.rle = phi i32 [ 0, %entry ], [ %.rle6, %bb2 ]
1005 %i.05 = phi i32 [ 0, %entry ], [ %indvar.next, %bb2 ]
1006 %1 = load i32* %cond, align 4
1007 %2 = icmp eq i32 %1, 0
1008 br i1 %2, label %bb2, label %bb1
1011 %3 = xor i32 %.rle, 234
1012 store i32 %3, i32* %res, align 4
1015 bb2: ; preds = %bb, %bb1
1016 %.rle6 = phi i32 [ %3, %bb1 ], [ %.rle, %bb ]
1017 %indvar.next = add i32 %i.05, 1
1018 %exitcond = icmp eq i32 %indvar.next, %n
1019 br i1 %exitcond, label %return, label %bb
1021 DSE should sink partially dead stores to get the store out of the loop.
1023 Here's another partial dead case:
1024 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12395
1026 //===---------------------------------------------------------------------===//
1028 Scalar PRE hoists the mul in the common block up to the else:
1030 int test (int a, int b, int c, int g) {
1040 It would be better to do the mul once to reduce codesize above the if.
1041 This is GCC PR38204.
1044 //===---------------------------------------------------------------------===//
1045 This simple function from 179.art:
1048 struct { double y; int reset; } *Y;
1053 for (i=0;i<numf2s;i++)
1054 if (Y[i].y > Y[winner].y)
1058 Compiles into (with clang TBAA):
1060 for.body: ; preds = %for.inc, %bb.nph
1061 %indvar = phi i64 [ 0, %bb.nph ], [ %indvar.next, %for.inc ]
1062 %i.01718 = phi i32 [ 0, %bb.nph ], [ %i.01719, %for.inc ]
1063 %tmp4 = getelementptr inbounds %struct.anon* %tmp3, i64 %indvar, i32 0
1064 %tmp5 = load double* %tmp4, align 8, !tbaa !4
1065 %idxprom7 = sext i32 %i.01718 to i64
1066 %tmp10 = getelementptr inbounds %struct.anon* %tmp3, i64 %idxprom7, i32 0
1067 %tmp11 = load double* %tmp10, align 8, !tbaa !4
1068 %cmp12 = fcmp ogt double %tmp5, %tmp11
1069 br i1 %cmp12, label %if.then, label %for.inc
1071 if.then: ; preds = %for.body
1072 %i.017 = trunc i64 %indvar to i32
1075 for.inc: ; preds = %for.body, %if.then
1076 %i.01719 = phi i32 [ %i.01718, %for.body ], [ %i.017, %if.then ]
1077 %indvar.next = add i64 %indvar, 1
1078 %exitcond = icmp eq i64 %indvar.next, %tmp22
1079 br i1 %exitcond, label %for.cond.for.end_crit_edge, label %for.body
1082 It is good that we hoisted the reloads of numf2's, and Y out of the loop and
1083 sunk the store to winner out.
1085 However, this is awful on several levels: the conditional truncate in the loop
1086 (-indvars at fault? why can't we completely promote the IV to i64?).
1088 Beyond that, we have a partially redundant load in the loop: if "winner" (aka
1089 %i.01718) isn't updated, we reload Y[winner].y the next time through the loop.
1090 Similarly, the addressing that feeds it (including the sext) is redundant. In
1091 the end we get this generated assembly:
1093 LBB0_2: ## %for.body
1094 ## =>This Inner Loop Header: Depth=1
1098 ucomisd (%rcx,%r8), %xmm0
1107 All things considered this isn't too bad, but we shouldn't need the movslq or
1108 the shlq instruction, or the load folded into ucomisd every time through the
1111 On an x86-specific topic, if the loop can't be restructure, the movl should be a
1114 //===---------------------------------------------------------------------===//
1118 GCC PR37810 is an interesting case where we should sink load/store reload
1119 into the if block and outside the loop, so we don't reload/store it on the
1140 We now hoist the reload after the call (Transforms/GVN/lpre-call-wrap.ll), but
1141 we don't sink the store. We need partially dead store sinking.
1143 //===---------------------------------------------------------------------===//
1145 [LOAD PRE CRIT EDGE SPLITTING]
1147 GCC PR37166: Sinking of loads prevents SROA'ing the "g" struct on the stack
1148 leading to excess stack traffic. This could be handled by GVN with some crazy
1149 symbolic phi translation. The code we get looks like (g is on the stack):
1153 %9 = getelementptr %struct.f* %g, i32 0, i32 0
1154 store i32 %8, i32* %9, align bel %bb3
1156 bb3: ; preds = %bb1, %bb2, %bb
1157 %c_addr.0 = phi %struct.f* [ %g, %bb2 ], [ %c, %bb ], [ %c, %bb1 ]
1158 %b_addr.0 = phi %struct.f* [ %b, %bb2 ], [ %g, %bb ], [ %b, %bb1 ]
1159 %10 = getelementptr %struct.f* %c_addr.0, i32 0, i32 0
1160 %11 = load i32* %10, align 4
1162 %11 is partially redundant, an in BB2 it should have the value %8.
1164 GCC PR33344 and PR35287 are similar cases.
1167 //===---------------------------------------------------------------------===//
1171 There are many load PRE testcases in testsuite/gcc.dg/tree-ssa/loadpre* in the
1172 GCC testsuite, ones we don't get yet are (checked through loadpre25):
1174 [CRIT EDGE BREAKING]
1175 loadpre3.c predcom-4.c
1177 [PRE OF READONLY CALL]
1180 [TURN SELECT INTO BRANCH]
1181 loadpre14.c loadpre15.c
1183 actually a conditional increment: loadpre18.c loadpre19.c
1185 //===---------------------------------------------------------------------===//
1187 [LOAD PRE / STORE SINKING / SPEC HACK]
1189 This is a chunk of code from 456.hmmer:
1191 int f(int M, int *mc, int *mpp, int *tpmm, int *ip, int *tpim, int *dpp,
1192 int *tpdm, int xmb, int *bp, int *ms) {
1194 for (k = 1; k <= M; k++) {
1195 mc[k] = mpp[k-1] + tpmm[k-1];
1196 if ((sc = ip[k-1] + tpim[k-1]) > mc[k]) mc[k] = sc;
1197 if ((sc = dpp[k-1] + tpdm[k-1]) > mc[k]) mc[k] = sc;
1198 if ((sc = xmb + bp[k]) > mc[k]) mc[k] = sc;
1203 It is very profitable for this benchmark to turn the conditional stores to mc[k]
1204 into a conditional move (select instr in IR) and allow the final store to do the
1205 store. See GCC PR27313 for more details. Note that this is valid to xform even
1206 with the new C++ memory model, since mc[k] is previously loaded and later
1209 //===---------------------------------------------------------------------===//
1212 There are many PRE testcases in testsuite/gcc.dg/tree-ssa/ssa-pre-*.c in the
1215 //===---------------------------------------------------------------------===//
1217 There are some interesting cases in testsuite/gcc.dg/tree-ssa/pred-comm* in the
1218 GCC testsuite. For example, we get the first example in predcom-1.c, but
1219 miss the second one:
1224 __attribute__ ((noinline))
1225 void count_averages(int n) {
1227 for (i = 1; i < n; i++)
1228 avg[i] = (((unsigned long) fib[i - 1] + fib[i] + fib[i + 1]) / 3) & 0xffff;
1231 which compiles into two loads instead of one in the loop.
1233 predcom-2.c is the same as predcom-1.c
1235 predcom-3.c is very similar but needs loads feeding each other instead of
1239 //===---------------------------------------------------------------------===//
1243 Type based alias analysis:
1244 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14705
1246 We should do better analysis of posix_memalign. At the least it should
1247 no-capture its pointer argument, at best, we should know that the out-value
1248 result doesn't point to anything (like malloc). One example of this is in
1249 SingleSource/Benchmarks/Misc/dt.c
1251 //===---------------------------------------------------------------------===//
1253 Interesting missed case because of control flow flattening (should be 2 loads):
1254 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26629
1255 With: llvm-gcc t2.c -S -o - -O0 -emit-llvm | llvm-as |
1256 opt -mem2reg -gvn -instcombine | llvm-dis
1257 we miss it because we need 1) CRIT EDGE 2) MULTIPLE DIFFERENT
1258 VALS PRODUCED BY ONE BLOCK OVER DIFFERENT PATHS
1260 //===---------------------------------------------------------------------===//
1262 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19633
1263 We could eliminate the branch condition here, loading from null is undefined:
1265 struct S { int w, x, y, z; };
1266 struct T { int r; struct S s; };
1267 void bar (struct S, int);
1268 void foo (int a, struct T b)
1276 //===---------------------------------------------------------------------===//
1278 simplifylibcalls should do several optimizations for strspn/strcspn:
1280 strcspn(x, "a") -> inlined loop for up to 3 letters (similarly for strspn):
1282 size_t __strcspn_c3 (__const char *__s, int __reject1, int __reject2,
1284 register size_t __result = 0;
1285 while (__s[__result] != '\0' && __s[__result] != __reject1 &&
1286 __s[__result] != __reject2 && __s[__result] != __reject3)
1291 This should turn into a switch on the character. See PR3253 for some notes on
1294 456.hmmer apparently uses strcspn and strspn a lot. 471.omnetpp uses strspn.
1296 //===---------------------------------------------------------------------===//
1298 "gas" uses this idiom:
1299 else if (strchr ("+-/*%|&^:[]()~", *intel_parser.op_string))
1301 else if (strchr ("<>", *intel_parser.op_string)
1303 Those should be turned into a switch.
1305 //===---------------------------------------------------------------------===//
1307 252.eon contains this interesting code:
1309 %3072 = getelementptr [100 x i8]* %tempString, i32 0, i32 0
1310 %3073 = call i8* @strcpy(i8* %3072, i8* %3071) nounwind
1311 %strlen = call i32 @strlen(i8* %3072) ; uses = 1
1312 %endptr = getelementptr [100 x i8]* %tempString, i32 0, i32 %strlen
1313 call void @llvm.memcpy.i32(i8* %endptr,
1314 i8* getelementptr ([5 x i8]* @"\01LC42", i32 0, i32 0), i32 5, i32 1)
1315 %3074 = call i32 @strlen(i8* %endptr) nounwind readonly
1317 This is interesting for a couple reasons. First, in this:
1319 The memcpy+strlen strlen can be replaced with:
1321 %3074 = call i32 @strlen([5 x i8]* @"\01LC42") nounwind readonly
1323 Because the destination was just copied into the specified memory buffer. This,
1324 in turn, can be constant folded to "4".
1326 In other code, it contains:
1328 %endptr6978 = bitcast i8* %endptr69 to i32*
1329 store i32 7107374, i32* %endptr6978, align 1
1330 %3167 = call i32 @strlen(i8* %endptr69) nounwind readonly
1332 Which could also be constant folded. Whatever is producing this should probably
1333 be fixed to leave this as a memcpy from a string.
1335 Further, eon also has an interesting partially redundant strlen call:
1337 bb8: ; preds = %_ZN18eonImageCalculatorC1Ev.exit
1338 %682 = getelementptr i8** %argv, i32 6 ; <i8**> [#uses=2]
1339 %683 = load i8** %682, align 4 ; <i8*> [#uses=4]
1340 %684 = load i8* %683, align 1 ; <i8> [#uses=1]
1341 %685 = icmp eq i8 %684, 0 ; <i1> [#uses=1]
1342 br i1 %685, label %bb10, label %bb9
1345 %686 = call i32 @strlen(i8* %683) nounwind readonly
1346 %687 = icmp ugt i32 %686, 254 ; <i1> [#uses=1]
1347 br i1 %687, label %bb10, label %bb11
1349 bb10: ; preds = %bb9, %bb8
1350 %688 = call i32 @strlen(i8* %683) nounwind readonly
1352 This could be eliminated by doing the strlen once in bb8, saving code size and
1353 improving perf on the bb8->9->10 path.
1355 //===---------------------------------------------------------------------===//
1357 I see an interesting fully redundant call to strlen left in 186.crafty:InputMove
1359 %movetext11 = getelementptr [128 x i8]* %movetext, i32 0, i32 0
1362 bb62: ; preds = %bb55, %bb53
1363 %promote.0 = phi i32 [ %169, %bb55 ], [ 0, %bb53 ]
1364 %171 = call i32 @strlen(i8* %movetext11) nounwind readonly align 1
1365 %172 = add i32 %171, -1 ; <i32> [#uses=1]
1366 %173 = getelementptr [128 x i8]* %movetext, i32 0, i32 %172
1369 br i1 %or.cond, label %bb65, label %bb72
1371 bb65: ; preds = %bb62
1372 store i8 0, i8* %173, align 1
1375 bb72: ; preds = %bb65, %bb62
1376 %trank.1 = phi i32 [ %176, %bb65 ], [ -1, %bb62 ]
1377 %177 = call i32 @strlen(i8* %movetext11) nounwind readonly align 1
1379 Note that on the bb62->bb72 path, that the %177 strlen call is partially
1380 redundant with the %171 call. At worst, we could shove the %177 strlen call
1381 up into the bb65 block moving it out of the bb62->bb72 path. However, note
1382 that bb65 stores to the string, zeroing out the last byte. This means that on
1383 that path the value of %177 is actually just %171-1. A sub is cheaper than a
1386 This pattern repeats several times, basically doing:
1391 where it is "obvious" that B = A-1.
1393 //===---------------------------------------------------------------------===//
1395 186.crafty has this interesting pattern with the "out.4543" variable:
1397 call void @llvm.memcpy.i32(
1398 i8* getelementptr ([10 x i8]* @out.4543, i32 0, i32 0),
1399 i8* getelementptr ([7 x i8]* @"\01LC28700", i32 0, i32 0), i32 7, i32 1)
1400 %101 = call@printf(i8* ... @out.4543, i32 0, i32 0)) nounwind
1402 It is basically doing:
1404 memcpy(globalarray, "string");
1405 printf(..., globalarray);
1407 Anyway, by knowing that printf just reads the memory and forward substituting
1408 the string directly into the printf, this eliminates reads from globalarray.
1409 Since this pattern occurs frequently in crafty (due to the "DisplayTime" and
1410 other similar functions) there are many stores to "out". Once all the printfs
1411 stop using "out", all that is left is the memcpy's into it. This should allow
1412 globalopt to remove the "stored only" global.
1414 //===---------------------------------------------------------------------===//
1418 define inreg i32 @foo(i8* inreg %p) nounwind {
1420 %tmp1 = ashr i8 %tmp0, 5
1421 %tmp2 = sext i8 %tmp1 to i32
1425 could be dagcombine'd to a sign-extending load with a shift.
1426 For example, on x86 this currently gets this:
1432 while it could get this:
1437 //===---------------------------------------------------------------------===//
1441 int test(int x) { return 1-x == x; } // --> return false
1442 int test2(int x) { return 2-x == x; } // --> return x == 1 ?
1444 Always foldable for odd constants, what is the rule for even?
1446 //===---------------------------------------------------------------------===//
1448 PR 3381: GEP to field of size 0 inside a struct could be turned into GEP
1449 for next field in struct (which is at same address).
1451 For example: store of float into { {{}}, float } could be turned into a store to
1454 //===---------------------------------------------------------------------===//
1456 The arg promotion pass should make use of nocapture to make its alias analysis
1457 stuff much more precise.
1459 //===---------------------------------------------------------------------===//
1461 The following functions should be optimized to use a select instead of a
1462 branch (from gcc PR40072):
1464 char char_int(int m) {if(m>7) return 0; return m;}
1465 int int_char(char m) {if(m>7) return 0; return m;}
1467 //===---------------------------------------------------------------------===//
1469 int func(int a, int b) { if (a & 0x80) b |= 0x80; else b &= ~0x80; return b; }
1473 define i32 @func(i32 %a, i32 %b) nounwind readnone ssp {
1475 %0 = and i32 %a, 128 ; <i32> [#uses=1]
1476 %1 = icmp eq i32 %0, 0 ; <i1> [#uses=1]
1477 %2 = or i32 %b, 128 ; <i32> [#uses=1]
1478 %3 = and i32 %b, -129 ; <i32> [#uses=1]
1479 %b_addr.0 = select i1 %1, i32 %3, i32 %2 ; <i32> [#uses=1]
1483 However, it's functionally equivalent to:
1485 b = (b & ~0x80) | (a & 0x80);
1487 Which generates this:
1489 define i32 @func(i32 %a, i32 %b) nounwind readnone ssp {
1491 %0 = and i32 %b, -129 ; <i32> [#uses=1]
1492 %1 = and i32 %a, 128 ; <i32> [#uses=1]
1493 %2 = or i32 %0, %1 ; <i32> [#uses=1]
1497 This can be generalized for other forms:
1499 b = (b & ~0x80) | (a & 0x40) << 1;
1501 //===---------------------------------------------------------------------===//
1503 These two functions produce different code. They shouldn't:
1507 uint8_t p1(uint8_t b, uint8_t a) {
1508 b = (b & ~0xc0) | (a & 0xc0);
1512 uint8_t p2(uint8_t b, uint8_t a) {
1513 b = (b & ~0x40) | (a & 0x40);
1514 b = (b & ~0x80) | (a & 0x80);
1518 define zeroext i8 @p1(i8 zeroext %b, i8 zeroext %a) nounwind readnone ssp {
1520 %0 = and i8 %b, 63 ; <i8> [#uses=1]
1521 %1 = and i8 %a, -64 ; <i8> [#uses=1]
1522 %2 = or i8 %1, %0 ; <i8> [#uses=1]
1526 define zeroext i8 @p2(i8 zeroext %b, i8 zeroext %a) nounwind readnone ssp {
1528 %0 = and i8 %b, 63 ; <i8> [#uses=1]
1529 %.masked = and i8 %a, 64 ; <i8> [#uses=1]
1530 %1 = and i8 %a, -128 ; <i8> [#uses=1]
1531 %2 = or i8 %1, %0 ; <i8> [#uses=1]
1532 %3 = or i8 %2, %.masked ; <i8> [#uses=1]
1536 //===---------------------------------------------------------------------===//
1538 IPSCCP does not currently propagate argument dependent constants through
1539 functions where it does not not all of the callers. This includes functions
1540 with normal external linkage as well as templates, C99 inline functions etc.
1541 Specifically, it does nothing to:
1543 define i32 @test(i32 %x, i32 %y, i32 %z) nounwind {
1545 %0 = add nsw i32 %y, %z
1548 %3 = add nsw i32 %1, %2
1552 define i32 @test2() nounwind {
1554 %0 = call i32 @test(i32 1, i32 2, i32 4) nounwind
1558 It would be interesting extend IPSCCP to be able to handle simple cases like
1559 this, where all of the arguments to a call are constant. Because IPSCCP runs
1560 before inlining, trivial templates and inline functions are not yet inlined.
1561 The results for a function + set of constant arguments should be memoized in a
1564 //===---------------------------------------------------------------------===//
1566 The libcall constant folding stuff should be moved out of SimplifyLibcalls into
1567 libanalysis' constantfolding logic. This would allow IPSCCP to be able to
1568 handle simple things like this:
1570 static int foo(const char *X) { return strlen(X); }
1571 int bar() { return foo("abcd"); }
1573 //===---------------------------------------------------------------------===//
1575 functionattrs doesn't know much about memcpy/memset. This function should be
1576 marked readnone rather than readonly, since it only twiddles local memory, but
1577 functionattrs doesn't handle memset/memcpy/memmove aggressively:
1579 struct X { int *p; int *q; };
1586 p = __builtin_memcpy (&x, &y, sizeof (int *));
1590 This can be seen at:
1591 $ clang t.c -S -o - -mkernel -O0 -emit-llvm | opt -functionattrs -S
1594 //===---------------------------------------------------------------------===//
1596 Missed instcombine transformation:
1597 define i1 @a(i32 %x) nounwind readnone {
1599 %cmp = icmp eq i32 %x, 30
1600 %sub = add i32 %x, -30
1601 %cmp2 = icmp ugt i32 %sub, 9
1602 %or = or i1 %cmp, %cmp2
1605 This should be optimized to a single compare. Testcase derived from gcc.
1607 //===---------------------------------------------------------------------===//
1609 Missed instcombine or reassociate transformation:
1610 int a(int a, int b) { return (a==12)&(b>47)&(b<58); }
1612 The sgt and slt should be combined into a single comparison. Testcase derived
1615 //===---------------------------------------------------------------------===//
1617 Missed instcombine transformation:
1619 %382 = srem i32 %tmp14.i, 64 ; [#uses=1]
1620 %383 = zext i32 %382 to i64 ; [#uses=1]
1621 %384 = shl i64 %381, %383 ; [#uses=1]
1622 %385 = icmp slt i32 %tmp14.i, 64 ; [#uses=1]
1624 The srem can be transformed to an and because if %tmp14.i is negative, the
1625 shift is undefined. Testcase derived from 403.gcc.
1627 //===---------------------------------------------------------------------===//
1629 This is a range comparison on a divided result (from 403.gcc):
1631 %1337 = sdiv i32 %1336, 8 ; [#uses=1]
1632 %.off.i208 = add i32 %1336, 7 ; [#uses=1]
1633 %1338 = icmp ult i32 %.off.i208, 15 ; [#uses=1]
1635 We already catch this (removing the sdiv) if there isn't an add, we should
1636 handle the 'add' as well. This is a common idiom with it's builtin_alloca code.
1639 int a(int x) { return (unsigned)(x/16+7) < 15; }
1641 Another similar case involves truncations on 64-bit targets:
1643 %361 = sdiv i64 %.046, 8 ; [#uses=1]
1644 %362 = trunc i64 %361 to i32 ; [#uses=2]
1646 %367 = icmp eq i32 %362, 0 ; [#uses=1]
1648 //===---------------------------------------------------------------------===//
1650 Missed instcombine/dagcombine transformation:
1651 define void @lshift_lt(i8 zeroext %a) nounwind {
1653 %conv = zext i8 %a to i32
1654 %shl = shl i32 %conv, 3
1655 %cmp = icmp ult i32 %shl, 33
1656 br i1 %cmp, label %if.then, label %if.end
1659 tail call void @bar() nounwind
1665 declare void @bar() nounwind
1667 The shift should be eliminated. Testcase derived from gcc.
1669 //===---------------------------------------------------------------------===//
1671 These compile into different code, one gets recognized as a switch and the
1672 other doesn't due to phase ordering issues (PR6212):
1674 int test1(int mainType, int subType) {
1677 else if (mainType == 9)
1679 else if (mainType == 11)
1684 int test2(int mainType, int subType) {
1694 //===---------------------------------------------------------------------===//
1696 The following test case (from PR6576):
1698 define i32 @mul(i32 %a, i32 %b) nounwind readnone {
1700 %cond1 = icmp eq i32 %b, 0 ; <i1> [#uses=1]
1701 br i1 %cond1, label %exit, label %bb.nph
1702 bb.nph: ; preds = %entry
1703 %tmp = mul i32 %b, %a ; <i32> [#uses=1]
1705 exit: ; preds = %entry
1709 could be reduced to:
1711 define i32 @mul(i32 %a, i32 %b) nounwind readnone {
1713 %tmp = mul i32 %b, %a
1717 //===---------------------------------------------------------------------===//
1719 We should use DSE + llvm.lifetime.end to delete dead vtable pointer updates.
1722 Another interesting case is that something related could be used for variables
1723 that go const after their ctor has finished. In these cases, globalopt (which
1724 can statically run the constructor) could mark the global const (so it gets put
1725 in the readonly section). A testcase would be:
1728 using namespace std;
1729 const complex<char> should_be_in_rodata (42,-42);
1730 complex<char> should_be_in_data (42,-42);
1731 complex<char> should_be_in_bss;
1733 Where we currently evaluate the ctors but the globals don't become const because
1734 the optimizer doesn't know they "become const" after the ctor is done. See
1735 GCC PR4131 for more examples.
1737 //===---------------------------------------------------------------------===//
1742 return x > 1 ? x : 1;
1745 LLVM emits a comparison with 1 instead of 0. 0 would be equivalent
1746 and cheaper on most targets.
1748 LLVM prefers comparisons with zero over non-zero in general, but in this
1749 case it choses instead to keep the max operation obvious.
1751 //===---------------------------------------------------------------------===//
1753 Take the following testcase on x86-64 (similar testcases exist for all targets
1756 define void @a(i64* nocapture %s, i64* nocapture %t, i64 %a, i64 %b,
1759 %0 = zext i64 %a to i128 ; <i128> [#uses=1]
1760 %1 = zext i64 %b to i128 ; <i128> [#uses=1]
1761 %2 = add i128 %1, %0 ; <i128> [#uses=2]
1762 %3 = zext i64 %c to i128 ; <i128> [#uses=1]
1763 %4 = shl i128 %3, 64 ; <i128> [#uses=1]
1764 %5 = add i128 %4, %2 ; <i128> [#uses=1]
1765 %6 = lshr i128 %5, 64 ; <i128> [#uses=1]
1766 %7 = trunc i128 %6 to i64 ; <i64> [#uses=1]
1767 store i64 %7, i64* %s, align 8
1768 %8 = trunc i128 %2 to i64 ; <i64> [#uses=1]
1769 store i64 %8, i64* %t, align 8
1788 //===---------------------------------------------------------------------===//
1790 Switch lowering generates less than ideal code for the following switch:
1791 define void @a(i32 %x) nounwind {
1793 switch i32 %x, label %if.end [
1794 i32 0, label %if.then
1795 i32 1, label %if.then
1796 i32 2, label %if.then
1797 i32 3, label %if.then
1798 i32 5, label %if.then
1801 tail call void @foo() nounwind
1808 Generated code on x86-64 (other platforms give similar results):
1821 The movl+movl+btq+jb could be simplified to a cmpl+jne.
1823 Or, if we wanted to be really clever, we could simplify the whole thing to
1824 something like the following, which eliminates a branch:
1832 //===---------------------------------------------------------------------===//
1836 int foo(int a) { return (a & (~15)) / 16; }
1840 define i32 @foo(i32 %a) nounwind readnone ssp {
1842 %and = and i32 %a, -16
1843 %div = sdiv i32 %and, 16
1847 but this code (X & -A)/A is X >> log2(A) when A is a power of 2, so this case
1848 should be instcombined into just "a >> 4".
1850 We do get this at the codegen level, so something knows about it, but
1851 instcombine should catch it earlier:
1859 //===---------------------------------------------------------------------===//
1861 This code (from GCC PR28685):
1863 int test(int a, int b) {
1873 define i32 @test(i32 %a, i32 %b) nounwind readnone ssp {
1875 %cmp = icmp slt i32 %a, %b
1876 br i1 %cmp, label %return, label %if.end
1878 if.end: ; preds = %entry
1879 %cmp5 = icmp eq i32 %a, %b
1880 %conv6 = zext i1 %cmp5 to i32
1883 return: ; preds = %entry
1889 define i32 @test__(i32 %a, i32 %b) nounwind readnone ssp {
1891 %0 = icmp sle i32 %a, %b
1892 %retval = zext i1 %0 to i32
1896 //===---------------------------------------------------------------------===//
1898 This code can be seen in viterbi:
1900 %64 = call noalias i8* @malloc(i64 %62) nounwind
1902 %67 = call i64 @llvm.objectsize.i64(i8* %64, i1 false) nounwind
1903 %68 = call i8* @__memset_chk(i8* %64, i32 0, i64 %62, i64 %67) nounwind
1905 llvm.objectsize.i64 should be taught about malloc/calloc, allowing it to
1906 fold to %62. This is a security win (overflows of malloc will get caught)
1907 and also a performance win by exposing more memsets to the optimizer.
1909 This occurs several times in viterbi.
1911 Note that this would change the semantics of @llvm.objectsize which by its
1912 current definition always folds to a constant. We also should make sure that
1913 we remove checking in code like
1915 char *p = malloc(strlen(s)+1);
1916 __strcpy_chk(p, s, __builtin_objectsize(p, 0));
1918 //===---------------------------------------------------------------------===//
1920 This code (from Benchmarks/Dhrystone/dry.c):
1922 define i32 @Func1(i32, i32) nounwind readnone optsize ssp {
1924 %sext = shl i32 %0, 24
1925 %conv = ashr i32 %sext, 24
1926 %sext6 = shl i32 %1, 24
1927 %conv4 = ashr i32 %sext6, 24
1928 %cmp = icmp eq i32 %conv, %conv4
1929 %. = select i1 %cmp, i32 10000, i32 0
1933 Should be simplified into something like:
1935 define i32 @Func1(i32, i32) nounwind readnone optsize ssp {
1937 %sext = shl i32 %0, 24
1938 %conv = and i32 %sext, 0xFF000000
1939 %sext6 = shl i32 %1, 24
1940 %conv4 = and i32 %sext6, 0xFF000000
1941 %cmp = icmp eq i32 %conv, %conv4
1942 %. = select i1 %cmp, i32 10000, i32 0
1948 define i32 @Func1(i32, i32) nounwind readnone optsize ssp {
1950 %conv = and i32 %0, 0xFF
1951 %conv4 = and i32 %1, 0xFF
1952 %cmp = icmp eq i32 %conv, %conv4
1953 %. = select i1 %cmp, i32 10000, i32 0
1956 //===---------------------------------------------------------------------===//
1958 clang -O3 currently compiles this code
1960 int g(unsigned int a) {
1961 unsigned int c[100];
1964 unsigned int b = c[10] + c[11];
1972 define i32 @g(i32 a) nounwind readnone {
1973 %add = shl i32 %a, 1
1974 %mul = shl i32 %a, 1
1975 %cmp = icmp ugt i32 %add, %mul
1976 %a.addr.0 = select i1 %cmp, i32 11, i32 15
1980 The icmp should fold to false. This CSE opportunity is only available
1981 after GVN and InstCombine have run.
1983 //===---------------------------------------------------------------------===//
1985 memcpyopt should turn this:
1987 define i8* @test10(i32 %x) {
1988 %alloc = call noalias i8* @malloc(i32 %x) nounwind
1989 call void @llvm.memset.p0i8.i32(i8* %alloc, i8 0, i32 %x, i32 1, i1 false)
1993 into a call to calloc. We should make sure that we analyze calloc as
1994 aggressively as malloc though.
1996 //===---------------------------------------------------------------------===//
1998 clang -O3 doesn't optimize this:
2000 void f1(int* begin, int* end) {
2001 std::fill(begin, end, 0);
2004 into a memset. This is PR8942.
2006 //===---------------------------------------------------------------------===//
2008 clang -O3 -fno-exceptions currently compiles this code:
2011 std::vector<int> v(N);
2013 extern void sink(void*); sink(&v);
2018 define void @_Z1fi(i32 %N) nounwind {
2020 %v2 = alloca [3 x i32*], align 8
2021 %v2.sub = getelementptr inbounds [3 x i32*]* %v2, i64 0, i64 0
2022 %tmpcast = bitcast [3 x i32*]* %v2 to %"class.std::vector"*
2023 %conv = sext i32 %N to i64
2024 store i32* null, i32** %v2.sub, align 8, !tbaa !0
2025 %tmp3.i.i.i.i.i = getelementptr inbounds [3 x i32*]* %v2, i64 0, i64 1
2026 store i32* null, i32** %tmp3.i.i.i.i.i, align 8, !tbaa !0
2027 %tmp4.i.i.i.i.i = getelementptr inbounds [3 x i32*]* %v2, i64 0, i64 2
2028 store i32* null, i32** %tmp4.i.i.i.i.i, align 8, !tbaa !0
2029 %cmp.i.i.i.i = icmp eq i32 %N, 0
2030 br i1 %cmp.i.i.i.i, label %_ZNSt12_Vector_baseIiSaIiEEC2EmRKS0_.exit.thread.i.i, label %cond.true.i.i.i.i
2032 _ZNSt12_Vector_baseIiSaIiEEC2EmRKS0_.exit.thread.i.i: ; preds = %entry
2033 store i32* null, i32** %v2.sub, align 8, !tbaa !0
2034 store i32* null, i32** %tmp3.i.i.i.i.i, align 8, !tbaa !0
2035 %add.ptr.i5.i.i = getelementptr inbounds i32* null, i64 %conv
2036 store i32* %add.ptr.i5.i.i, i32** %tmp4.i.i.i.i.i, align 8, !tbaa !0
2037 br label %_ZNSt6vectorIiSaIiEEC1EmRKiRKS0_.exit
2039 cond.true.i.i.i.i: ; preds = %entry
2040 %cmp.i.i.i.i.i = icmp slt i32 %N, 0
2041 br i1 %cmp.i.i.i.i.i, label %if.then.i.i.i.i.i, label %_ZNSt12_Vector_baseIiSaIiEEC2EmRKS0_.exit.i.i
2043 if.then.i.i.i.i.i: ; preds = %cond.true.i.i.i.i
2044 call void @_ZSt17__throw_bad_allocv() noreturn nounwind
2047 _ZNSt12_Vector_baseIiSaIiEEC2EmRKS0_.exit.i.i: ; preds = %cond.true.i.i.i.i
2048 %mul.i.i.i.i.i = shl i64 %conv, 2
2049 %call3.i.i.i.i.i = call noalias i8* @_Znwm(i64 %mul.i.i.i.i.i) nounwind
2050 %0 = bitcast i8* %call3.i.i.i.i.i to i32*
2051 store i32* %0, i32** %v2.sub, align 8, !tbaa !0
2052 store i32* %0, i32** %tmp3.i.i.i.i.i, align 8, !tbaa !0
2053 %add.ptr.i.i.i = getelementptr inbounds i32* %0, i64 %conv
2054 store i32* %add.ptr.i.i.i, i32** %tmp4.i.i.i.i.i, align 8, !tbaa !0
2055 call void @llvm.memset.p0i8.i64(i8* %call3.i.i.i.i.i, i8 0, i64 %mul.i.i.i.i.i, i32 4, i1 false)
2056 br label %_ZNSt6vectorIiSaIiEEC1EmRKiRKS0_.exit
2058 This is just the handling the construction of the vector. Most surprising here
2059 is the fact that all three null stores in %entry are dead (because we do no
2062 Also surprising is that %conv isn't simplified to 0 in %....exit.thread.i.i.
2063 This is a because the client of LazyValueInfo doesn't simplify all instruction
2064 operands, just selected ones.
2066 //===---------------------------------------------------------------------===//
2068 clang -O3 -fno-exceptions currently compiles this code:
2070 void f(char* a, int n) {
2071 __builtin_memset(a, 0, n);
2072 for (int i = 0; i < n; ++i)
2078 define void @_Z1fPci(i8* nocapture %a, i32 %n) nounwind {
2080 %conv = sext i32 %n to i64
2081 tail call void @llvm.memset.p0i8.i64(i8* %a, i8 0, i64 %conv, i32 1, i1 false)
2082 %cmp8 = icmp sgt i32 %n, 0
2083 br i1 %cmp8, label %for.body.lr.ph, label %for.end
2085 for.body.lr.ph: ; preds = %entry
2086 %tmp10 = add i32 %n, -1
2087 %tmp11 = zext i32 %tmp10 to i64
2088 %tmp12 = add i64 %tmp11, 1
2089 call void @llvm.memset.p0i8.i64(i8* %a, i8 0, i64 %tmp12, i32 1, i1 false)
2092 for.end: ; preds = %entry
2096 This shouldn't need the ((zext (%n - 1)) + 1) game, and it should ideally fold
2097 the two memset's together. The issue with %n seems to stem from poor handling
2098 of the original loop.
2100 To simplify this, we need SCEV to know that "n != 0" because of the dominating
2101 conditional. That would turn the second memset into a simple memset of 'n'.
2103 //===---------------------------------------------------------------------===//
2105 clang -O3 -fno-exceptions currently compiles this code:
2108 unsigned short m1, m2;
2109 unsigned char m3, m4;
2113 std::vector<S> v(N);
2114 extern void sink(void*); sink(&v);
2117 into poor code for zero-initializing 'v' when N is >0. The problem is that
2118 S is only 6 bytes, but each element is 8 byte-aligned. We generate a loop and
2119 4 stores on each iteration. If the struct were 8 bytes, this gets turned into
2122 In order to handle this we have to:
2123 A) Teach clang to generate metadata for memsets of structs that have holes in
2125 B) Teach clang to use such a memset for zero init of this struct (since it has
2126 a hole), instead of doing elementwise zeroing.
2128 //===---------------------------------------------------------------------===//
2130 clang -O3 currently compiles this code:
2132 extern const int magic;
2133 double f() { return 0.0 * magic; }
2137 @magic = external constant i32
2139 define double @_Z1fv() nounwind readnone {
2141 %tmp = load i32* @magic, align 4, !tbaa !0
2142 %conv = sitofp i32 %tmp to double
2143 %mul = fmul double %conv, 0.000000e+00
2147 We should be able to fold away this fmul to 0.0. More generally, fmul(x,0.0)
2148 can be folded to 0.0 if we can prove that the LHS is not -0.0, not a NaN, and
2149 not an INF. The CannotBeNegativeZero predicate in value tracking should be
2150 extended to support general "fpclassify" operations that can return
2151 yes/no/unknown for each of these predicates.
2153 In this predicate, we know that uitofp is trivially never NaN or -0.0, and
2154 we know that it isn't +/-Inf if the floating point type has enough exponent bits
2155 to represent the largest integer value as < inf.
2157 //===---------------------------------------------------------------------===//
2159 When optimizing a transformation that can change the sign of 0.0 (such as the
2160 0.0*val -> 0.0 transformation above), it might be provable that the sign of the
2161 expression doesn't matter. For example, by the above rules, we can't transform
2162 fmul(sitofp(x), 0.0) into 0.0, because x might be -1 and the result of the
2163 expression is defined to be -0.0.
2165 If we look at the uses of the fmul for example, we might be able to prove that
2166 all uses don't care about the sign of zero. For example, if we have:
2168 fadd(fmul(sitofp(x), 0.0), 2.0)
2170 Since we know that x+2.0 doesn't care about the sign of any zeros in X, we can
2171 transform the fmul to 0.0, and then the fadd to 2.0.
2173 //===---------------------------------------------------------------------===//
2175 We should enhance memcpy/memcpy/memset to allow a metadata node on them
2176 indicating that some bytes of the transfer are undefined. This is useful for
2177 frontends like clang when lowering struct copies, when some elements of the
2178 struct are undefined. Consider something like this:
2184 void foo(struct x*P);
2185 struct x testfunc() {
2193 We currently compile this to:
2194 $ clang t.c -S -o - -O0 -emit-llvm | opt -scalarrepl -S
2197 %struct.x = type { i8, [4 x i32] }
2199 define void @testfunc(%struct.x* sret %agg.result) nounwind ssp {
2201 %V1 = alloca %struct.x, align 4
2202 call void @foo(%struct.x* %V1)
2203 %tmp1 = bitcast %struct.x* %V1 to i8*
2204 %0 = bitcast %struct.x* %V1 to i160*
2205 %srcval1 = load i160* %0, align 4
2206 %tmp2 = bitcast %struct.x* %agg.result to i8*
2207 %1 = bitcast %struct.x* %agg.result to i160*
2208 store i160 %srcval1, i160* %1, align 4
2212 This happens because SRoA sees that the temp alloca has is being memcpy'd into
2213 and out of and it has holes and it has to be conservative. If we knew about the
2214 holes, then this could be much much better.
2216 Having information about these holes would also improve memcpy (etc) lowering at
2217 llc time when it gets inlined, because we can use smaller transfers. This also
2218 avoids partial register stalls in some important cases.
2220 //===---------------------------------------------------------------------===//
2222 Some missed instcombine xforms (from GCC PR14753):
2226 void mask_gt (unsigned int a) {
2227 /* This is equivalent to a > 15. */
2232 void neg_eq_cst(unsigned int a) {
2237 void minus_cst(unsigned int a) {
2242 void rotate_cst (unsigned a) {
2243 a = (a << 10) | (a >> 22);
2248 //===---------------------------------------------------------------------===//