[ConstantRange] Remove 'Of' from name of ConstantRange::isSizeStrictlySmallerThanOf so that it reads better. NFC

llvm-svn: 302383

This patch fixes test failures that occur on Windows because
the tests attempt to generate two distinct temp file names but
get the same name both time.

The fix for this is to create the first temp file before requesting
a second temporary file name. This ensures that the second name
will be unique.

llvm-svn: 302382

llvm-svn: 302381

On Windows the function template `template <class T> void test()` has
the same mangled name when instantiated with the distinct types `void()`
and `void() noexcept`. When this occurs Clang emits an error. This error
was causing two type-traits tests to fail.

However this can be worked around by using class templates instead of
function templates, which is what this patch does to fix the errors.

llvm-svn: 302380

Summary:
When compiling for GPU, one can now choose to compile for OpenCL or CUDA,
with the corresponding polly-gpu-runtime flag (libopencl / libcudart). The
GPURuntime library (GPUJIT) has been extended with the OpenCL Runtime library
for that purpose, correctly choosing the corresponding library calls to the
option chosen when compiling (via different initialization calls).

Additionally, a specific GPU Target architecture can now be chosen with -polly-gpu-arch (only nvptx64 implemented thus far).

Reviewers: grosser, bollu, Meinersbur, etherzhhb, singam-sanjay

Reviewed By: grosser, Meinersbur

Subscribers: singam-sanjay, llvm-commits, pollydev, nemanjai, mgorny, yaxunl, Anastasia

Tags: #polly

Differential Revision: https://reviews.llvm.org/D32431

llvm-svn: 302379

Account for subvector extraction/insertion, helps prevent the vectorizers from selecting 256-bit vectors that will have to be split anyhow on AVX1 targets. 

llvm-svn: 302378

llvm-svn: 302377

Summary: Function argument naming changed to reflect capitalization standards.

Reviewers: grosser, Meinersbur

Reviewed By: grosser

Differential Revision: https://reviews.llvm.org/D32854

llvm-svn: 302376

Summary: Errors previously printed to stdout now get printed to stderr.

Reviewers: grosser, Meinersbur

Reviewed By: grosser

Differential Revision: https://reviews.llvm.org/D32852

llvm-svn: 302375

llvm-svn: 302374

Summary:
Re-applying r301766 with a fix to a typo and a regression test.

The log message for r301766 was:
==================================================================================
    InstructionSimplify: Canonicalize shuffle operands. NFC-ish.

    Summary:
     Apply canonicalization rules:
        1. Input vectors with no elements selected from can be replaced with undef.
        2. If only one input vector is constant it shall be the second one.

    This allows constant-folding to cover more ad-hoc simplifications that
    were in place and avoid duplication for RHS and LHS checks.

    There are more rules we may want to add in the future when we see a
    justification. e.g. mask elements that select undef elements can be
    replaced with undef.
==================================================================================

Reviewers: spatel, RKSimon

Reviewed By: spatel

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D32863

llvm-svn: 302373

Currently llvm-rtdyld in -check mode will map sections to back-to-back 4k
aligned slabs starting at 0x1000. Automatically remapping sections by default is
helpful because it quickly exposes relocation bugs due to use of local addresses
rather than load addresses (these would silently pass if the load address was
not remapped). These mappings can be explicitly overridden on a per-section
basis using llvm-rtdlyd's -map-section option. This patch extends this scheme to
also preserve any mappings made by RuntimeDyld itself. Preserving RuntimeDyld's
automatic mappings allows us to write test cases to verify that these automatic
mappings have been applied.

This will allow the fix in https://reviews.llvm.org/D32899 to be tested with
llvm-rtdyld -check.

llvm-svn: 302372

Summary: This makes setRange take ConstantRange by rvalue reference since most callers were passing an unnamed temporary ConstantRange. We can then move that ConstantRange into the DenseMap caches. For the callers that weren't passing a temporary, I've added std::move to to the local variable being passed.

Reviewers: sanjoy, mzolotukhin, efriedma

Reviewed By: sanjoy

Subscribers: takuto.ikuta, llvm-commits

Differential Revision: https://reviews.llvm.org/D32943

llvm-svn: 302371

We can simplify (or (icmp X, C1), (icmp X, C2)) to 'true' or one of the icmps in many cases.
I had to check some of these with Alive to prove to myself it's right, but everything seems 
to check out. Eg, the deleted code in instcombine was completely ignoring predicates with
mismatched signedness.

This is a follow-up to:
https://reviews.llvm.org/rL301260
https://reviews.llvm.org/D32143

llvm-svn: 302370

The variable Proto is moved at the beginning of the codegen() function.
According to the comment above, the pointed object should be used due the
reference P.

Differential Revision: https://reviews.llvm.org/D32939

llvm-svn: 302369

llvm-svn: 302368

llvm-svn: 302367

llvm-svn: 302366

Add the ARM64 COFF relocation types.  This will be needed to add support
for the AArch64 Windows object file emission support.

llvm-svn: 302365

llvm-svn: 302364

llvm-svn: 302363

Summary:
In https://bugs.freebsd.org/207918, Daniel McRobb describes how using
std::showbase with ostreams can cause truncation of unsigned long long
when output format is octal.  In fact, this can even happen with
unsigned int and unsigned long.

To ensure this does not happen, add one additional character to the
do_put buffers if std::showbase is on.  Also add a test case.

Reviewers: EricWF, mclow.lists

Reviewed By: EricWF

Subscribers: cfe-commits, emaste

Differential Revision: https://reviews.llvm.org/D32670

llvm-svn: 302362

llvm-svn: 302361

llvm-svn: 302360

Extend NoVLX targets to use the 512-bit versions

llvm-svn: 302359

This test was introduced in r302339. It works on my system, but breaks on the
buildbots.

llvm-svn: 302358

[X86] Reduce code for setting operations actions by merging into loops across multiple types/ops. NFCI.

llvm-svn: 302357

Use the ISD::ABS opcode directly

Differential Revision: https://reviews.llvm.org/D32944

llvm-svn: 302356

rL294581 broke unnecessary register dependencies on partial v16i8/v8i16 BUILD_VECTORs, but on SSE41 we (currently) use insertion for full BUILD_VECTORs as well. By allowing full insertion to occur on SSE41 targets we can break register dependencies here as well.

llvm-svn: 302355

Summary:
When writing a loop pass I made a mistake and hit the assertion
"Unreachable block in loop". Later, I hit an assertion when I called
`BasicBlock::eraseFromParent()` incorrectly: "Use still stuck around
after Def is destroyed". This latter assertion, however, printed out
exactly which value is being deleted and what uses remain, which helped
me debug the issue.

To help people debugging their loop passes in the future, print out
exactly which basic block is unreachable in a loop.

Reviewers: sanjoy, hfinkel, mehdi_amini

Reviewed By: mehdi_amini

Subscribers: mzolotukhin

Differential Revision: https://reviews.llvm.org/D32878

llvm-svn: 302354

Enabled emulated TLS on WOA for mingw
Fix <windows.h> include for mingw

Reviewed By: chapuni, mstorsjo

Subscribers: compnerd, llvm-commits

Differential Revision: https://reviews.llvm.org/D32681

llvm-svn: 302340

Extend the Knowledge class to store information about the contents
of array elements and which values are written. Two knowledges do
not conflict the known content is the same. The content information
if computed from writes to and loads from the array elements, and
represented by "ValInst": isl spaces that compare equal if the value
represented is the same.

Differential Revision: https://reviews.llvm.org/D31247

llvm-svn: 302339

Many more to come...

llvm-svn: 302338

Remove an extra canonicalization step if ISD::ABS is going to be used anyway.

Updated x86 abs combine to check that we are lowering from both canonicalizations.

llvm-svn: 302337

Allow using a system's install jsoncpp library instead of the bundled
one with the setting POLLY_BUNDLED_JSONCPP=OFF.

This fixes llvm.org/PR32929

Differential Revision: https://reviews.llvm.org/D32922

llvm-svn: 302336

This changes one parameter to be a const APInt& since we only read from it. Use std::move on local APInts once they are no longer needed so we can reuse their allocations. Lastly, use operator+=(uint64_t) instead of adding 1 to an APInt twice creating a new APInt each time.

llvm-svn: 302335

llvm-svn: 302334

llvm-svn: 302333

llvm-svn: 302332

Summary:
ConstantRange contains two APInts which can allocate memory if their width is larger than 64-bits. So we shouldn't copy it when we can avoid it.

This changes LVILatticeVal::getConstantRange() to return its internal ConstantRange by reference. This allows many places that just need a ConstantRange reference to avoid making a copy.

Several places now capture the return value of getConstantRange() by reference so they can call methods on it that don't need a new object.

Lastly it adds std::move in one place to capture to move a local ConstantRange into an LVILatticeVal.

Reviewers: reames, dberlin, sanjoy, anna

Reviewed By: reames

Subscribers: grandinj, llvm-commits

Differential Revision: https://reviews.llvm.org/D32884

llvm-svn: 302331