[SelectionDAG] In PromoteFloatRes_BITCAST, insert a bitcast before the fp16_to_fp in case the input type isn't an i16.

The bitcast can be further legalized as needed.

Fixes PR38533.

llvm-svn: 339533

llvm-svn: 339532

[InstCombine] Replace call to haveNoCommonBitsSet in visitXor with just the special case that doesn't use computeKnownBits.

Summary: computeKnownBits is expensive. The cases that would be detected by the computeKnownBits portion of haveNoCommonBitsSet were already handled by the earlier call to SimplifyDemandedInstructionBits.

Reviewers: spatel, lebedev.ri

Reviewed By: lebedev.ri

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D50604

llvm-svn: 339531

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D50605

llvm-svn: 339530

[X86] Add constant folding for AVX512 versions of scalar floating point to integer conversion intrinsics.

Summary:
We've supported constant folding for sse versions for many years. This patch adds support for the avx512 versions including unsigned with the default rounding mode. We could probably do more with other roundings modes and SAE in the future.

The test cases are largely based on the sse.ll test cases. But I did add some test cases to ensure the unsigned versions don't accept negative values. Also checked the bounds of f64->i32 conversions to make sure unsigned has a larger positive range than signed.

Reviewers: RKSimon, spatel, chandlerc

Reviewed By: RKSimon

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D50553

llvm-svn: 339529

Summary: Found by GCC's -Wunused-function.

Patch by Kim Gräsman

Reviewers: ab, dsanders, llvm-commits

Reviewed By: dsanders

Subscribers: rovka, kristof.beyls

Differential Revision: https://reviews.llvm.org/D50611

llvm-svn: 339528

This form makes more sense (it is a range over constant arguments) and is most consistent with const_arg_iterator (there are zero instances of arg_const_iterator).

llvm-svn: 339527

Also add some more tests in preparation for
a future patch.

llvm-svn: 339526

Addresses fixme, although this should still be checking individual
operand flags.

llvm-svn: 339525

llvm-svn: 339524

llvm-svn: 339523

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D50606

llvm-svn: 339522

Summary:
The as<T>() method would trigger the following warning on GCC <7:

   warning: dereferencing type-punned pointer will break
       strict-aliasing rules [-Wstrict-aliasing]

     return *reinterpret_cast<T *>(Union.buffer);
                                               ^

Union.buffer is guaranteed to be aligned to whatever types it contains,
and json::Value maintains the invariant that it only calls as<T>() for a
T it has previously placement-newed into Union.buffer. This should
follow the rules for strict aliasing.

Using two static_cast via void * instead of reinterpret_cast
silences the warning and presumably makes GCC understand that no
strict-aliasing violation is happening.

No functional change intended.

Patch by: kimgr (Kim Gräsman)

Reviewers: sammccall, xiangzhai, HaoLiu, llvm-commits, xbolva00

Reviewed By: sammccall, xbolva00

Subscribers: xbolva00

Differential Revision: https://reviews.llvm.org/D50608

llvm-svn: 339521

Summary:
Basic version was merged - https://reviews.llvm.org/D49954

This adds support for FP & non-commutative opcodes

Precommited tests: https://reviews.llvm.org/rL338727

Reviewers: spatel, lebedev.ri

Reviewed By: spatel

Subscribers: jfb

Differential Revision: https://reviews.llvm.org/D50190

llvm-svn: 339520

  (X * Z) + (Y * Z) --> (X + Y) * Z
  (X * Z) - (Y * Z) --> (X - Y) * Z
  (X / Z) + (Y / Z) --> (X + Y) / Z
  (X / Z) - (Y / Z) --> (X - Y) / Z

The existing code that implemented these folds failed to 
optimize vectors, and it transformed code with multiple 
uses when it should not have.

llvm-svn: 339519

llvm-svn: 339518

Adding the readability module to the list of dependencies for the C++ Core Guidelines module. Amends r339516 for a failing bot.

llvm-svn: 339517

Add a new check to the readability module that flags uses of "magic numbers" (both floating-point and integral).

Patch by Florin Iucha <florin@signbit.net>

llvm-svn: 339516

These are always UB, but can happen for large integer inputs. Testing it
is very fragile as -simplifycfg will nuke the UB top-down.

llvm-svn: 339515

This adds a coverage for the following early continue:
https://github.com/llvm-mirror/lld/blob/master/ELF/CallGraphSort.cpp#L200

llvm-svn: 339514

I'm not sure the exact nsz flag combination that
is OK. I think as long as it's on either, this is OK.
For now just check it on the omod multiply.

llvm-svn: 339513

If one of the elements is undef, use the canonicalized constant
from the other element instead of 0.

Splat vectors are more useful for other optimizations, such
as matching vector clamps. This was breaking on clamps
of half3 from the undef 4th component.

llvm-svn: 339512

llvm-svn: 339511

llvm-svn: 339510

We were checking for all bits being Known by checking Known.Zero|Known.One, but if all the bits are known then the value should be a Constant and we can just check for that instead.

llvm-svn: 339509

llvm-svn: 339508

The use of the or_is_add predicate already gives enough of a complexity boost to get the patterns ordered properly.

llvm-svn: 339507

Reviewers: a.sidorin, a_sidorin

Reviewed By: a_sidorin

Subscribers: a_sidorin, martong, cfe-commits

Differential Revision: https://reviews.llvm.org/D50550

llvm-svn: 339506

Reviewers: a.sidorin, a_sidorin

Reviewed By: a_sidorin

Subscribers: a_sidorin, martong, cfe-commits

Differential Revision: https://reviews.llvm.org/D50552

llvm-svn: 339505

Summary:
Instead of iterating over our vector of functions, we might as well use a map here to
directly get the function we need.

Thanks to Vedant for pointing this out.

Reviewers: vsk

Reviewed By: vsk

Subscribers: mgrang, lldb-commits

Differential Revision: https://reviews.llvm.org/D50225

llvm-svn: 339504

This was missed in SVN r337754.

llvm-svn: 339503

Summary: After converting all existing passes to use the new DomTreeUpdater interface, there isn't any usage of the original DeferredDominance class. Thus, we can safely remove it from the codebase.

Reviewers: kuhar, brzycki, dmgreen, davide, grosser

Reviewed By: kuhar, brzycki

Subscribers: mgorny, llvm-commits

Differential Revision: https://reviews.llvm.org/D49747

llvm-svn: 339502

Try to improve the computed counts when it has been explicitly set by a pragma
or command line option. This moves the code around, so that first call to
computeUnrollCount to get a sensible count and override that if explicit unroll
and jam counts are specified.

Also added some extra debug messages for when unroll and jamming is disabled.

Differential Revision: https://reviews.llvm.org/D50075

llvm-svn: 339501

Pulled out a separate function for some code that calculates
if an inner loop iteration count is invariant to it's outer
loop.

Differential Revision: https://reviews.llvm.org/D50063

llvm-svn: 339500

[X86] Remove the AL/AX/EAX/RAX short immediate forms from the macro fusion shouldScheduleAdjacent. NFC

These instructions are only created by the backend during MCInst lowering.

llvm-svn: 339499

Unlike the other arithmetic instructions the mem-reg form of compare is just a load and not a RMW operation. According to the Intel optimization manual, this form is also supported by macro fusion.

llvm-svn: 339498

The are RMW of memory operations. They aren't eligible for macro fusion.

llvm-svn: 339497

[X86] Change the MOV32ri64 pseudo instruction to def a GR64 directly instead of wrapping it in a SUBREG_TO_REG.

Now we switch to the subregister in expandPostRAPseudos where we already switched the opcode.

This simplifies a few isel patterns that used the pseudo directly. And magically seems to have improved our ability to CSE it in the undef-label.ll test.

llvm-svn: 339496

Treat the stack variants of control instructions the same as regular
instructions.  Otherwise, the vector ControlFlowStack will be the wrong
size and have out-of-bounds access.  This was detected by MemorySanitizer.

llvm-svn: 339495

Make the section boundary checks on Windows not depend on the order as they are emitted in reverse when the compiler is built by Visual C++.

llvm-svn: 339494