Based off the codegen reports on PR51075 - hopefully we can handle some of this in SLP or VectorCombine, but we usually have to leave load combining until the backend so at least some of these patterns will still appear even then.

This is part of an effort to reduce the differences between the custom C++ bindings used right now by polly in `lib/External/isl/include/isl/isl-noxceptions.h` and the official isl C++ interface.

Changes made:
 - Stop generating `isl::union_set` and isl::union_map` from `isl::space` and instead generate them from `isl::ctx`
 - Disable clang-format on `isl-noexceptions.h`
 - Removed `isl::union_{set,map}` generator from `isl::space` from `isl-noexceptions.h`
 - `isl-noexceptions.h` has been generated by this https://github.com/patacca/isl/commit/87c3413b6f1d62ca3dddf716352f90a0b8533353

Reviewed By: Meinersbur

Differential Revision: https://reviews.llvm.org/D106059

Insert ops replacing pad_tensor in front of the associated tansfer_write / insert_slice op. Otherwise we may end up with invalid ir if one of the remaining tansfer_write / insert_slice operands is defined after the pad_tensor op.

Reviewed By: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D106162

Finds function calls where the call arguments might be provided in an
incorrect order, based on the comparison (via string metrics) of the
parameter names and the argument names against each other.

A diagnostic is emitted if an argument name is similar to a *different*
parameter than the one currently passed to, and it is sufficiently
dissimilar to the one it **is** passed to currently.

False-positive warnings from this check are useful to indicate bad
naming convention issues, even if a swap isn't necessary.
This check does not generate FixIts.

Originally implemented by @varjujan as his Master's Thesis work.
The check was subsequently taken over by @barancsuk who added type
conformity checks to silence false positive matches.
The work by @whisperity involved driving the check's review and fixing
some more bugs in the process.

Reviewed By: aaron.ballman, alexfh

Differential Revision: http://reviews.llvm.org/D20689



Co-authored-by: János Varjú <varjujanos2@gmail.com>
Co-authored-by: Lilla Barancsuk <barancsuklilla@gmail.com>

Replace code which identifies induction phi with helper function
getInductionVariable to improve robustness.

Differential Revision: https://reviews.llvm.org/D106045

This patch adds support for the following instructions:

    SCLAMP, UCLAMP, REV, DUP (predicate)

The reference can be found here:
https://developer.arm.com/documentation/ddi0602/2021-06

Reviewed By: kmclaughlin

Differential Revision: https://reviews.llvm.org/D105577

This relaxes the VMLAV and VADDV reduction recognition code to handle
smaller than legal types, extending them as needed. That was already
handled for some reductions, this extends it to more types in a more
generic way. If a smaller than legal value is found it is extended to
the legal type as needed.

Differential Revision: https://reviews.llvm.org/D106051

The case for nxv2f32/nxv2i32 was already covered by D104573.
This patch builds on top of that by making the mechanism work for
nxv2[b]f16/nxv2i16, nxv4[b]f16/nxv4i16 as well.

Reviewed By: efriedma

Differential Revision: https://reviews.llvm.org/D106138

If a file has no symbols, perhaps because it is a linked executable,
synthesize some symbols by walking the code section.  Otherwise the
disassembler will try to treat the whole code section as a function,
which won't parse.  Fixes https://bugs.llvm.org/show_bug.cgi?id=50957.

Differential Revision: https://reviews.llvm.org/D105539

D106236 added a new CMake argument for `libomptarget` test, but when user's
input contains white spaces, CMake will add escape char to the final lit command,
which leads to an error. This patch converts the user's input `LIBOMPTARGET_LIT_ARGS`
into a local array, and then passes the array to the function.

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D106247

Apparently there was a latent bug here.

Reviewed By: jpienaar

Differential Revision: https://reviews.llvm.org/D105883

This reverts commit 2a419a0b.

The result of a shufflevector must not propagate poison from any element
other than the one noted in the shuffle mask.

The regressions outside of fptoui-may-overflow.ll can probably be
recovered some other way; for example, using isGuaranteedNotToBePoison.

See discussion on https://reviews.llvm.org/D106053 for more background.

Differential Revision: https://reviews.llvm.org/D106222

The `__CUDA__` macro is already defined for openmp/nvptx and is not used by
`__clang_cuda_complex_builtins.h`, so dropping that macro slightly simplifies
nvptx and avoids defining it on amdgcn (where it is likely to be harmful).

Also dropped a cplusplus test from a C++ header as compilation will have
failed on cmath earlier if it was included from C.

Reviewed By: jdoerfert, fodinabor

Differential Revision: https://reviews.llvm.org/D105221

The inttoptr/ptrtoint roundtrip optimization is not always correct.
We are working towards removing this optimization and adding support
to specific cases where this optimization works. This patch is the
first one on this line.

Consider the example:

    %i = ptrtoint i8* %X to i64
    %p = inttoptr i64 %i to i16*
    %cmp = icmp eq i8* %load, %p

In this specific case, the inttoptr/ptrtoint optimization is correct
as it only compares the pointer values. In this patch, we fold
inttoptr/ptrtoint to a bitcast (if src and dest types are different).

Differential Revision: https://reviews.llvm.org/D105088

This patch relands https://reviews.llvm.org/D104799, but fixes the
memory handling causing leak sanitizer failures.

This reverts commit a56fe117.

Keeping a marker for Changed might have made sense before
this code was refactored, but we never touch that variable
after initialization now.

As discussed on D106058 - we were failing to keep the common flags. This matches the behaviour in InstCombinerImpl::foldSelectOpOp.

[OpenMP][Offloading] Add a CMake argument LIBOMPTARGET_LIT_ARGS to control behavior of libomptarget lit test

By default, `lit` uses all threads to invoke tests, which  can easily cause out
of memory on GPUs because most of OpenMP offloading test usually take about 1GB
GPU memory, but a typical GPU only has 4-8GB memory. This patch introduce a
CMake argument `LIBOMPTARGET_LIT_ARGS` to allow users to control the behavior of
`libomptarget` tests, similar to `LLVM_LIT_ARGS`.

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D106236

Shows failure to fold sel(sra(X,C1),sra(Y,C1)) -> sra(sel(X,Y),C1) (and to retain the flags)

The last use was removed on Apr 28, 2014 in commit
c5a3139e.

As discussed on D106058 - we're not propagating the common flags to the merged binop

Differential Revision: https://reviews.llvm.org/D103750

Broke ASAN buildbot, will reland with fixes

This reverts commit b5a6ad8c.

I'm going to extend the functionality started in D106058 so move the folds into their own method to reduce the amount of code in DAGCombiner::visitSELECT

Currently when we compile the project in debug mode, `-g` will not be added to
compilation flag. The bc files generated in different mode are of different size.
When using GPU debuggers like `cuda-gdb`, it is expected to provide more info
with a debug version of bc lib.

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D106229

[X86][SSE] matchShuffleWithPACK - avoid poison pollution from bitcasting multiple elements together.

D106053 exposed that we've not been taking into account that by bitcasting smaller elements together and then performing a ComputeKnownBits on the result we'd be allowing a poison element to influence other neighbouring elements being used in the pack. Instead we now peek through any existing bitcast to ensure that the source type already matches the width source of the pack node we're trying to match.

This has also been a chance to stop matchShuffleWithPACK creating unused nodes on the fly which could affect oneuse tests during shuffle lowering/combining.

The only regression we're seeing is due to being unable to peek through a bitcast as its on the other side of a extract_subvector - which should go away once we finally allow shuffle combining across different vector widths (by making matchShuffleWithPACK using const SelectionDAG& we've gotten closer to this - see PR45974).

At most these use the StringRef/Twine wrappers and don't have any implicit uses of std::string.

Move the include down to any cpp implementation where std::string is actually used.

This pattern is visible in unrolled and vectorized loops.
Although the backend seems to be able to reassociate to
ideal form in the examples I looked at, we might as well
do that in IR for efficiency.