Use "enable-vla-vectorization=vla" to generate a vector length agnostic
loops during vectorization. This option works for vectorization strategy 2.
Differential Revision: https://reviews.llvm.org/D118379
The way vector.create_mask is currently lowered is
vector-length-dependent, and therefore incompatible with scalable vector
types. This patch adds an alternative lowering path for create_mask
operations that return a scalable vector mask.
Differential Revision: https://reviews.llvm.org/D118248
On GFX10.3 targets, the following instruction sequence
v_cmp_* SGPR, ...
s_and_saveexec ..., SGPR
leads to a fairly long stall caused by a VALU write to a SGPR and having the
following SALU wait for the SGPR.
An equivalent sequence is to save the exec mask manually instead of letting
s_and_saveexec do the work and use a v_cmpx instruction instead to do the
comparison.
This patch modifies the SIOptimizeExecMasking pass as this is the last position
where s_and_saveexec instructions are inserted. It does the transformation by
trying to find the pattern, extracting the operands and generating the new
instruction sequence.
It also changes some existing lit tests and introduces a few new tests to show
the changed behavior on GFX10.3 targets.
Same as D119696 including a buildbot and MIR test fix.
Reviewed By: critson
Differential Revision: https://reviews.llvm.org/D122332
In D120129 we enhanced vectorization options of byval parameters. This patch
removes code duplication when handling byval and non-byval cases.
Differential Revision: https://reviews.llvm.org/D122381
This is the first in a series of patches that introduce C++20 importable
header units.
These differ from clang header modules in that:
(a) they are identifiable by an internal name
(b) they represent the top level source for a single header - although
that might include or import other headers.
We name importable header units with the path by which they are specified
(although that need not be the absolute path for the file).
So "foo/bar.h" would have a name "foo/bar.h". Header units are made a
separate module type so that we can deal with diagnosing places where they
are permitted but a named module is not.
Differential Revision: https://reviews.llvm.org/D121095
The way the alignment is determined here is not compatible with
opaque pointers -- we'd have to thread it through using either
align or elementtype attributes.
However, as far as I can tell this alignment is actually never
used for this particular intrinsic, so I've dropped the assignment
entirely and converted RecordAlignment to MaybeAlign, so we get an
assertion failure if it does end up being used.
Differential Revision: https://reviews.llvm.org/D122379
Follow-up of https://reviews.llvm.org/D121488 to ensure all descriptors
created inline complies with LBOUND requirement that the lower bound is
`1` when the related dimension extent is zero.
Both fir.xrebox and fir.xembox codegen is updated to enforce this
constraint.
Also upstream the "normalized lower bound" attribute that was added in fir-dev
since embox codegen was upstreamed, it is conflicting with this patch
otherwise.
Differential Revision: https://reviews.llvm.org/D122419
GetLowerBoundHelper rewrite in https://reviews.llvm.org/D121488 was
incorrect with POINTER/ALLOCATABLE components. The rewrite created a
descriptor inquiry to the component symbol only instead of the whole
named entity. The base information was lost, and not retrievable.
LBOUND(a(10)%p) became LBOUND(p).
Fix this regression, and also update DescriptorInquiry unparsing to
carry the kind information. DescriptorInquiries are KIND 8 expressions,
while LBOUND/SIZE/RANK, %LEN are default kind expressions.
This caused `print *,lbound(x,kind=8)` to unparse as `print*,lbound(x)` which is not
semantically the same (this unparsing issue was not an issue for
lowering, but I noticed it while writing my regression test).
Differential Revision: https://reviews.llvm.org/D122406
At Sony we are developing llvm-dva
https://lists.llvm.org/pipermail/llvm-dev/2020-August/144174.html
For its PDB support, it requires functionality already present in
llvm-pdbutil.
We intend to move that functionaly into the PDB library to be
shared by both tools. That change will be done in 2 steps, that
will be submitted as 2 patches:
(1) Replace 'ExitOnError' with explicit error handling.
(2) Move the intended shared code to the PDB library.
Patch for step (1): https://reviews.llvm.org/D121801
This patch is for step (2).
Move InputFile.cpp[h], FormatUtil.cpp[h] and LinePrinter.cpp[h]
files to the debug PDB library.
It exposes the following functionality that can be used by tools:
- Open a PDB file.
- Get module debug stream.
- Traverse module sections.
- Traverse module subsections.
Most of the needed functionality is in InputFile, but there are
dependencies from LinePrinter and FormatUtil.
Some other functionality is in the following functions in
DumpOutputStyle.cpp file:
- iterateModuleSubsections
- getModuleDebugStream
- iterateOneModule
- iterateSymbolGroups
- iterateModuleSubsections
Only these specific functions from DumpOutputStyle are moved to
the PDB library.
Reviewed By: aganea, dblaikie, rnk
Differential Revision: https://reviews.llvm.org/D122226
This patch adds the lowering for the `mvbits`
intrinsic.
This patch is part of the upstreaming effort from fir-dev branch.
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D122412
Co-authored-by: V Donaldson <vdonaldson@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
It's unusual that BLAKE3/CMakeLists.txt just defines a list of
files that it injects into its parent scope. The list should either
be defined in llvm/lib/Support/CMakeLists.txt, or
llvm/lib/Support/BLAKE3/CMakeLists.txt should define an object
library.
This does the latter. It makes llvm/lib/Support/BLAKE3/CMakeLists.txt
more self-contained.
No behavior change.
Differential Revision: https://reviews.llvm.org/D122428
Update `WeakUndeclaredIdentifiers` to hold a collection of weak
aliases per identifier instead of only one.
This also allows the "used" state to be removed from `WeakInfo`
because it is really only there as an alternative to removing
processed map entries, and we can represent that using an empty set
now. The serialization code is updated for the removal of the field.
Additionally, a PCH test is added for the new functionality.
The records are grouped by the "target" identifier, which was already
being used as a key for lookup purposes. We also store only one record
per alias name; combined, this means that diagnostics are grouped by
the "target" and limited to one per alias (which should be acceptable).
Fixes PR28611.
Fixesllvm/llvm-project#28985.
Reviewed By: aaron.ballman, cebowleratibm
Differential Revision: https://reviews.llvm.org/D121927
Co-authored-by: Rachel Craik <rcraik@ca.ibm.com>
Co-authored-by: Jamie Schmeiser <schmeise@ca.ibm.com>
This patch is a cleanup patch that merges PresburgerLocalSpace and
PresburgerSpace. Asserting that there are no locals is shifted to the
users of PresburgerSpace themselves.
The reasoning for this patch is that PresburgerLocalSpace did not contribute
much and only introduced additional complexity as locals could still be present
in PresburgerSpace, just not writable. This could introduce problems if a
PresburgerSpace with locals was copied to PresburgerLocalSpace which expected
no locals in a PresburgerSpace.
Reviewed By: arjunp
Differential Revision: https://reviews.llvm.org/D122353
In C++20 the type trait `type_identity` was introduced. For the same purpose there is `__identity` for pre-C++20 code. The name is confusing, because since C++20 there is also `identity`, which isn't a type trait.
Reviewed By: ldionne, Mordante, #libc
Spies: EricWF, libcxx-commits
Differential Revision: https://reviews.llvm.org/D122017
This is a re-commit of 98fd3b3598. The
newly added test was failing on the bots, and I've fixed the test now so
that it doesn't actually invoke the linker.
A previous patch removed the compiler generating offloading entries
for variables that were declared on the device but were internal or
hidden. This allowed us to compile programs but turns any attempt to run
'#pragma omp target update' on one of those variables a silent failure.
This patch adds a check in the semantic analysis for if the user is
attempting the update a variable on the device from the host that is not
externally visible.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D122403
This transformation allow to break up a reduction dimension in a
parallel and a reduction dimension. This is followed by a separate
reduction op. This allows to generate tree reduction which is beneficial
on target allowing to take advantage parallelism.
Differential Revision: https://reviews.llvm.org/D122045
When iterating over all Platforms looking for the best one, on a Mac the
Simulator platforms (iOS, tvOS, watchOS) will first find their SDK
directory by calling xcrun, then decide if they should activate or not.
When that SDK is absent, the call to xcrun to find it can be very slow.
This patch delays that directory search until we know we're activating
this platform, so non-simulator environments don't pay a perf cost ever
time they go through the list of platforms.
Differential Revision: https://reviews.llvm.org/D122373
rdar://87960090
To make it easier to find things that are not yet implemented, I'm changing the
messages that appear in the compiler's output to all have the string "not yet
implemented:".
These changes apply to files in the front end. I have another set of changes
to files in the lowering code.
Differential Revision: https://reviews.llvm.org/D122355
Reduce the range-reduction table size from 128 entries down to 64 entries, and
reduce the polynomial's degree from 6 down to 4.
Currently we use a degree-6 minimax polynomial on an interval of length 2^-7
around 0 to compute exp2f. Based on the suggestion of @santoshn and the RLIBM
project (https://github.com/rutgers-apl/rlibm-prog/blob/main/libm/float/exp2.c)
it is possible to have a good polynomial of degree-4 on a subinterval of length
2^(-6) to approximate 2^x.
We did try to either reduce the degree of the polynomial down to 3 or increase
the interval size to 2^(-5), but in both cases the number of exceptional values
exploded. So we settle with using a degree-4 polynomial of the interval of
size 2^(-6) around 0.
Reviewed By: michaelrj, sivachandra, zimmermann6, santoshn
Differential Revision: https://reviews.llvm.org/D122346
This is legacy PM-specific, which is deprecated.
Uses of this should be replaced with a corresponding `-passes='print<foo>'`.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D122420
Current clang generates extra set of simd variant function attribute
with extra 'v' encoding.
For example:
_ZGVbN2v__Z5add_1Pf vs _ZGVbN2vv__Z5add_1Pf
The problem is due to declaration of ParamAttrs following:
llvm::SmallVector<ParamAttrTy, 8> ParamAttrs(ParamPositions.size());
where ParamPositions.size() is grown after following assignment:
Pos = ParamPositions[PVD];
So the PVD is not find in ParamPositions.
The problem is ParamPositions need to set for each FD decl. To fix this
Move ParamPositions's init inside while loop for each FD.
Differential Revision: https://reviews.llvm.org/D122338
Since there is a table introduced for MAI instructions extend it
to use for DGEMM classification.
Differential Revision: https://reviews.llvm.org/D122337
Aakanksha