DAGCombiner::reduceBuildVecConvertToConvertBuildVec() was making two
mistakes:
1. It was checking the legality of scalar INT_TO_FP nodes and then generating
vector nodes.
2. It was passing the result value type to
TargetLoweringInfo::getOperationAction() when it should have been
passing the value type of the first operand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171420 91177308-0d34-0410-b5e6-96231b3b80d8
register. In most cases we actually compare or select YMM-sized registers
and mixing the two types creates horrible code. This commit optimizes
some of the transition sequences.
PR14657.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171148 91177308-0d34-0410-b5e6-96231b3b80d8
When these instructions are encoded in VEX (on AVX) there is no such requirement. This changes the folding
tables and removes the alignment restrictions from VEX-encoded instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171024 91177308-0d34-0410-b5e6-96231b3b80d8
pmuludq is slow, but it turns out that all the unpacking and packing of the
scalarized mul is even slower. 10% speedup on loop-vectorized paq8p.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170985 91177308-0d34-0410-b5e6-96231b3b80d8
Also loosen the SSSE3 dependency a bit, expanded pshufb + psra is still better
than scalarized loads. Fixes PR14590.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170984 91177308-0d34-0410-b5e6-96231b3b80d8
The only way to read the eflags is using push and pop. If we don't
adjust the stack then we run over the first frame index. This is
not something that we want to do, so we have to make sure that
our machine function does not copy the flags. If it does then
we have to emit the prolog that adjusts the stack.
rdar://12896831
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170961 91177308-0d34-0410-b5e6-96231b3b80d8
We match the pattern "x >= y ? x-y : 0" into "subus x, y" and two special cases
if y is a constant. DAGCombiner canonicalizes those so we first have to undo the
canonicalization for those cases. The pattern occurs in gzip when the loop
vectorizer is enabled. Part of PR14613.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170273 91177308-0d34-0410-b5e6-96231b3b80d8
try to reduce the width of this load, and would end up transforming:
(truncate (lshr (sextload i48 <ptr> as i64), 32) to i32)
to
(truncate (zextload i32 <ptr+4> as i64) to i32)
We lost the sext attached to the load while building the narrower i32
load, and replaced it with a zext because lshr always zext's the
results. Instead, bail out of this combine when there is a conflict
between a sextload and a zext narrowing. The rest of the DAG combiner
still optimize the code down to the proper single instruction:
movswl 6(...),%eax
Which is exactly what we wanted. Previously we read past the end *and*
missed the sign extension:
movl 6(...), %eax
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169802 91177308-0d34-0410-b5e6-96231b3b80d8
This shouldn't affect codegen for -O0 compiles as tail call markers are not
emitted in unoptimized compiles. Testing with the external/internal nightly
test suite reveals no change in compile time performance. Testing with -O1,
-O2 and -O3 with fast-isel enabled did not cause any compile-time or
execution-time failures. All tests were performed on my x86 machine.
I'll monitor our arm testers to ensure no regressions occur there.
In an upcoming clang patch I will be marking the objc_autoreleaseReturnValue
and objc_retainAutoreleaseReturnValue as tail calls unconditionally. While
it's theoretically true that this is just an optimization, it's an
optimization that we very much want to happen even at -O0, or else ARC
applications become substantially harder to debug.
Part of rdar://12553082
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169796 91177308-0d34-0410-b5e6-96231b3b80d8
1. Teach it to use overlapping unaligned load / store to copy / set the trailing
bytes. e.g. On 86, use two pairs of movups / movaps for 17 - 31 byte copies.
2. Use f64 for memcpy / memset on targets where i64 is not legal but f64 is. e.g.
x86 and ARM.
3. When memcpy from a constant string, do *not* replace the load with a constant
if it's not possible to materialize an integer immediate with a single
instruction (required a new target hook: TLI.isIntImmLegal()).
4. Use unaligned load / stores more aggressively if target hooks indicates they
are "fast".
5. Update ARM target hooks to use unaligned load / stores. e.g. vld1.8 / vst1.8.
Also increase the threshold to something reasonable (8 for memset, 4 pairs
for memcpy).
This significantly improves Dhrystone, up to 50% on ARM iOS devices.
rdar://12760078
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169791 91177308-0d34-0410-b5e6-96231b3b80d8
the VSRI instruction before it since it does not affect the MSB.
Thanks Craig Topper for suggesting this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169638 91177308-0d34-0410-b5e6-96231b3b80d8
check if loads that happen in between stores alias with the first store in the
chain, only with the second store onwards.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169516 91177308-0d34-0410-b5e6-96231b3b80d8
This is much simpler to reason about, more efficient, and
fixes some corner cases involving implicit super-register defs.
Fixed rdar://12797931.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169425 91177308-0d34-0410-b5e6-96231b3b80d8
The count attribute is more accurate with regards to the size of an array. It
also obviates the upper bound attribute in the subrange. We can also better
handle an unbound array by setting the count to -1 instead of the lower bound to
1 and upper bound to 0.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169312 91177308-0d34-0410-b5e6-96231b3b80d8
The count field is necessary because there isn't a difference between the 'lo'
and 'hi' attributes for a one-element array and a zero-element array. When the
count is '0', we know that this is a zero-element array. When it's >=1, then
it's a normal constant sized array. When it's -1, then the array is unbounded.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@169218 91177308-0d34-0410-b5e6-96231b3b80d8
This is a simple, cheap infrastructure for analyzing the shape of a
DAG. It recognizes uniform DAGs that take the shape of bottom-up
subtrees, such as the included matrix multiplication example. This is
useful for heuristics that balance register pressure with ILP. Two
canonical expressions of the heuristic are implemented in scheduling
modes: -misched-ilpmin and -misched-ilpmax.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@168773 91177308-0d34-0410-b5e6-96231b3b80d8
This fixes a hole in the "cheap" alias analysis logic implemented within
the DAG builder itself, regardless of whether proper alias analysis is
enabled. It now handles this pattern produced by LSR+CodeGenPrepare.
%sunkaddr1 = ptrtoint * %obj to i64
%sunkaddr2 = add i64 %sunkaddr1, %lsr.iv
%sunkaddr3 = inttoptr i64 %sunkaddr2 to i32*
store i32 %v, i32* %sunkaddr3
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@168768 91177308-0d34-0410-b5e6-96231b3b80d8
when the destination register is wider than the memory load.
These load instructions load from m32 or m64 and set the upper bits to zero,
while the folded instructions may accept m128.
rdar://12721174
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@168710 91177308-0d34-0410-b5e6-96231b3b80d8
This pass was conservative in that it always reserved the FP to enable dynamic
stack realignment, which allowed the RA to use aligned spills for vector
registers. This happens even when spills were not necessary. The RA has
since been improved to use unaligned spills when necessary.
The new behavior is to realign the stack if the frame pointer was already
reserved for some other reason, but don't reserve the frame pointer just
because a function contains vector virtual registers.
Part of rdar://12719844
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@168627 91177308-0d34-0410-b5e6-96231b3b80d8
The stack realignment code was fixed to work when there is stack realignment and
a dynamic alloca is present so this shouldn't cause correctness issues anymore.
Note that this also enables generation of AVX instructions for memset
under the assumptions:
- Unaligned loads/stores are always fast on CPUs supporting AVX
- AVX is not slower than SSE
We may need some tweaked heuristics if one of those assumptions turns out not to
be true.
Effectively reverts r58317. Part of PR2962.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167967 91177308-0d34-0410-b5e6-96231b3b80d8
temporarily as it is breaking the gdb bots.
This reverts commit r167806/e7ff4c14b157746b3e0228d2dce9f70712d1c126.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167886 91177308-0d34-0410-b5e6-96231b3b80d8
chain is correctly setup.
As an example, if the original load must happen before later stores, we need
to make sure the constructed VZEXT_LOAD is constrained to be before the stores.
rdar://12684358
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167859 91177308-0d34-0410-b5e6-96231b3b80d8
If we have a type 'int a[1]' and a type 'int b[0]', the generated DWARF is the
same for both of them because we use the 'upper_bound' attribute. Instead use
the 'count' attrbute, which gives the correct number of elements in the array.
<rdar://problem/12566646>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167806 91177308-0d34-0410-b5e6-96231b3b80d8
- Fix operand order for atomic sub, where the minuend is the value
loaded from memory and the subtrahend is the parameter specified.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167718 91177308-0d34-0410-b5e6-96231b3b80d8
- Add RTM code generation support throught 3 X86 intrinsics:
xbegin()/xend() to start/end a transaction region, and xabort() to abort a
tranaction region
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167573 91177308-0d34-0410-b5e6-96231b3b80d8
misched is disabled by default. With -enable-misched, these heuristics
balance the schedule to simultaneously avoid saturating processor
resources, expose ILP, and minimize register pressure. I've been
analyzing the performance of these heuristics on everything in the
llvm test suite in addition to a few other benchmarks. I would like
each heuristic check to be verified by a unit test, but I'm still
trying to figure out the best way to do that. The heuristics are still
in considerable flux, but as they are refined we should be rigorous
about unit testing the improvements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167527 91177308-0d34-0410-b5e6-96231b3b80d8
The adc/sbb optimization is to able to convert following expression
into a single adc/sbb instruction:
(ult) ... = x + 1 // where the ult is unsigned-less-than comparison
(ult) ... = x - 1
This change is to flip the "x >u y" (i.e. ugt comparison) in order
to expose the adc/sbb opportunity.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167180 91177308-0d34-0410-b5e6-96231b3b80d8
the first source operand is tied to the destination operand.
This is to accurately model the corresponding instructions where the upper
bits are unmodified.
rdar://12558838
PR14221
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167064 91177308-0d34-0410-b5e6-96231b3b80d8
incorrect instruction sequence due to it not being aware that an
inline assembly instruction may reference memory.
This patch fixes the problem by causing the scheduler to always assume that any
inline assembly code instruction could access memory. This is necessary because
the internal representation of the inline instruction does not include
any information about memory accesses.
This should fix PR13504.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166929 91177308-0d34-0410-b5e6-96231b3b80d8
- If more than 1 elemennts are defined and target supports the vectorized
conversion, use the vectorized one instead to reduce the strength on
conversion operation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166546 91177308-0d34-0410-b5e6-96231b3b80d8
- As there's no 64-bit GPRs in 32-bit mode, a custom conversion from v2u32 to
v2f32 is added to improve the efficiency of the code generated.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166545 91177308-0d34-0410-b5e6-96231b3b80d8
the difference from "int x" (which should go in registers and
"struct y {int x;}" (which should not).
Clang will be updated in the next patches.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166536 91177308-0d34-0410-b5e6-96231b3b80d8
- Check index being extracted to be constant 0 before simplfiying.
Otherwise, retain the original sequence.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166504 91177308-0d34-0410-b5e6-96231b3b80d8
- If INSERT_VECTOR_ELT is supported (above SSE2, either by custom
sequence of legal insn), transform BUILD_VECTOR into SHUFFLE +
INSERT_VECTOR_ELT if most of elements could be built from SHUFFLE with few
(so far 1) elements being inserted.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166288 91177308-0d34-0410-b5e6-96231b3b80d8
When merging stack slots, if StackColoring::remapInstructions gets a
value back from GetUnderlyingObject that it does not know about or is
not itself a stack slot, clear the memory operand in case it aliases
the merged slot. This prevents the introduction of incorrect aliasing
information.
Author: Matthew Curtis <mcurtis@codeaurora.org>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166216 91177308-0d34-0410-b5e6-96231b3b80d8
- Folding (trunc (concat ... X )) to (concat ... (trunc X) ...) is valid
when '...' are all 'undef's.
- r166125 relies on this transformation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166155 91177308-0d34-0410-b5e6-96231b3b80d8
- If the extracted vector has the same type of all vectored being concatenated
together, it should be simplified directly into v_i, where i is the index of
the element being extracted.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166125 91177308-0d34-0410-b5e6-96231b3b80d8
- MBB address is only valid as an immediate value in Small & Static
code/relocation models. On other models, LEA is needed to load IP address of
the restore MBB.
- A minor fix of MBB in MC lowering is added as well to enable target
relocation flag being propagated into MC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166084 91177308-0d34-0410-b5e6-96231b3b80d8
PR14098 contains an example where we would rematerialize a MOV8ri
immediately after the original instruction:
%vreg7:sub_8bit<def> = MOV8ri 9; GR32_ABCD:%vreg7
%vreg22:sub_8bit<def> = MOV8ri 9; GR32_ABCD:%vreg7
Besides being pointless, it is also wrong since the original instruction
only redefines part of the register, and the value read by the new
instruction is wrong.
The problem was the LiveRangeEdit::allUsesAvailableAt() didn't
special-case OrigIdx == UseIdx and found the wrong SSA value.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166068 91177308-0d34-0410-b5e6-96231b3b80d8
Original message:
The attached is the fix to radar://11663049. The optimization can be outlined by following rules:
(select (x != c), e, c) -> select (x != c), e, x),
(select (x == c), c, e) -> select (x == c), x, e)
where the <c> is an integer constant.
The reason for this change is that : on x86, conditional-move-from-constant needs two instructions;
however, conditional-move-from-register need only one instruction.
While the LowerSELECT() sounds to be the most convenient place for this optimization, it turns out to be a bad place. The reason is that by replacing the constant <c> with a symbolic value, it obscure some instruction-combining opportunities which would otherwise be very easy to spot. For that reason, I have to postpone the change to last instruction-combining phase.
The change passes the test of "make check-all -C <build-root/test" and "make -C project/test-suite/SingleSource".
Original message since r165661:
My previous change has a bug: I negated the condition code of a CMOV, and go ahead creating a new CMOV using the *ORIGINAL* condition code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166017 91177308-0d34-0410-b5e6-96231b3b80d8
This is a medium term workaround until we have a more robust solution
in the form of a register liveness utility for postRA passes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166001 91177308-0d34-0410-b5e6-96231b3b80d8
- Besides used in SjLj exception handling, __builtin_setjmp/__longjmp is also
used as a light-weight replacement of setjmp/longjmp which are used to
implementation continuation, user-level threading, and etc. The support added
in this patch ONLY addresses this usage and is NOT intended to support SjLj
exception handling as zero-cost DWARF exception handling is used by default
in X86.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165989 91177308-0d34-0410-b5e6-96231b3b80d8
X86 doesn't have i8 cmovs so isel would emit a branch. Emitting branches at this
level is often not a good idea because it's too late for many optimizations to
kick in. This solution doesn't add any extensions (truncs are free) and tries
to avoid introducing partial register stalls by filtering direct copyfromregs.
I'm seeing a ~10% speedup on reading a random .png file with libpng15 via
graphicsmagick on x86_64/westmere, but YMMV depending on the microarchitecture.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165868 91177308-0d34-0410-b5e6-96231b3b80d8
Completely update one interval at a time instead of collecting live
range fragments to be updated. This avoids building data structures,
except for a single SmallPtrSet of updated intervals.
Also share code between handleMove() and handleMoveIntoBundle().
Add support for moving dead defs across other live values in the
interval. The MI scheduler can do that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165824 91177308-0d34-0410-b5e6-96231b3b80d8
PHIElimination inserts IMPLICIT_DEF instructions to guarantee that all
PHI predecessors have a live-out value. These IMPLICIT_DEF values are
not considered to be real interference when coalescing virtual
registers:
%vreg1 = IMPLICIT_DEF
%vreg2 = MOV32r0
When joining %vreg1 and %vreg2, the IMPLICIT_DEF instruction and its
value number should simply be erased since the %vreg2 value number now
provides a live-out value for the PHI predecesor block.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165813 91177308-0d34-0410-b5e6-96231b3b80d8
Not all instructions define a virtual register in their first operand.
Specifically, INLINEASM has a different format.
<rdar://problem/12472811>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165721 91177308-0d34-0410-b5e6-96231b3b80d8
Original message:
The attached is the fix to radar://11663049. The optimization can be outlined by following rules:
(select (x != c), e, c) -> select (x != c), e, x),
(select (x == c), c, e) -> select (x == c), x, e)
where the <c> is an integer constant.
The reason for this change is that : on x86, conditional-move-from-constant needs two instructions;
however, conditional-move-from-register need only one instruction.
While the LowerSELECT() sounds to be the most convenient place for this optimization, it turns out to be a bad place. The reason is that by replacing the constant <c> with a symbolic value, it obscure some instruction-combining opportunities which would otherwise be very easy to spot. For that reason, I have to postpone the change to last instruction-combining phase.
The change passes the test of "make check-all -C <build-root/test" and "make -C project/test-suite/SingleSource".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165661 91177308-0d34-0410-b5e6-96231b3b80d8
- Due to the current matching vector elements constraints in
ISD::FP_ROUND, rounding from v2f64 to v4f32 (after legalization from
v2f32) is scalarized. Add a customized v2f32 widening to convert it
into a target-specific X86ISD::VFPROUND to work around this
constraints.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165631 91177308-0d34-0410-b5e6-96231b3b80d8
When the CFG contains a loop with multiple entry blocks, the traces
computed by MachineTraceMetrics don't always have the same nice
properties. Loop back-edges are normally excluded from traces, but
MachineLoopInfo doesn't recognize loops with multiple entry blocks, so
those back-edges may be included.
Avoid asserting when that happens by adding an isEarlierInSameTrace()
function that accurately determines if a dominating block is part of the
same trace AND is above the currrent block in the trace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165434 91177308-0d34-0410-b5e6-96231b3b80d8
in the Intel syntax.
The MC layer supports emitting in the Intel syntax, but this would require the
inline assembly MachineInstr to be lowered to an MCInst before emission. This
is potential future work, but for now emitting directly from the MachineInstr
suffices.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165173 91177308-0d34-0410-b5e6-96231b3b80d8
multiple stores with a single load. We create the wide loads and stores (and their chains)
before we remove the scalar loads and stores and fix the DAG chain. We attempted to merge
loads with a different chain. When that happened, the assumption that it is safe to RAUW
broke and a cycle was introduced.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165148 91177308-0d34-0410-b5e6-96231b3b80d8
is not profitable in many cases because modern processors perform multiple stores
in parallel and merging stores prior to merging requires extra work. We handle two main cases:
1. Store of multiple consecutive constants:
q->a = 3;
q->4 = 5;
In this case we store a single legal wide integer.
2. Store of multiple consecutive loads:
int a = p->a;
int b = p->b;
q->a = a;
q->b = b;
In this case we load/store either ilegal vector registers or legal wide integer registers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165125 91177308-0d34-0410-b5e6-96231b3b80d8
Enable the pass by default for targets that request it, and change the
-enable-early-ifcvt to the opposite -disable-early-ifcvt.
There are still some x86 regressions when enabling early if-conversion
because of the missing machine models. Disable the pass for x86 until
machine models are added.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165075 91177308-0d34-0410-b5e6-96231b3b80d8
X86DAGToDAGISel::PreprocessISelDAG(), isel is moving load inside
callseq_start / callseq_end so it can be folded into a call. This can
create a cycle in the DAG when the call is glued to a copytoreg. We
have been lucky this hasn't caused too many issues because the pre-ra
scheduler has special handling of call sequences. However, it has
caused a crash in a specific tailcall case.
rdar://12393897
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165072 91177308-0d34-0410-b5e6-96231b3b80d8
the add/sub case since in the case of multiplication you also have to check that
the operation in the larger type did not overflow.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165017 91177308-0d34-0410-b5e6-96231b3b80d8
- Update maximal stack alignment when stack arguments are prepared before a
call.
- Test cases are enhanced to show it's not a Win32 specific issue but a generic
one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@164946 91177308-0d34-0410-b5e6-96231b3b80d8