Previously, we just always returned 1. But that ignores that we have to do the operation for each subvector or scalar component.
Differential Revision: https://reviews.llvm.org/D78824
As reported here: https://reviews.llvm.org/D75153#1987272
Before, each instance of llvm-cov was creating one thread per hardware core, which wasn't needed probably because the number of inputs were small. This was probably causing a thread rlimit issue on large core count systems.
After this patch, the previous behavior is restored (to what was before rG8404aeb5):
If --num-threads is not specified, we create one thread per input, up to num.cores.
When specified, --num-threads indicates any number of threads, with no upper limit.
Differential Revision: https://reviews.llvm.org/D78408
When using reversedInstructionsWithoutDebug to construct a range from a
pair of MachineInstrBundleIterators, the range unexpectedly leaves out an
element. This results in mis-optimization as @mstorsjo points out in
https://reviews.llvm.org/D78157.
The problem is that when we convert a MachineInstrBundleIterator to a
reverse iterator, the result gets incremented:
MachineInstrBundleIterator(++I.getReverse())
The comment there explains that the "resulting iterator will dereference
... to the previous node, which is somewhat unexpected; but converting
the two endpoints in a range will give the same range in reverse". This
makes it hard to understand what reversedInstructionsWithoutDebug will
do: I've removed the helper to prevent similar mistakes in the future.
Summary:
- Whether or not a vector is scalable is a function of its type. Since
all instances of ScalableVectorType will have true for this value and
all instances of FixedVectorType will have false for this value, there
is no need to store it as a class member.
Reviewers: efriedma, fpetrogalli, kmclaughlin
Reviewed By: fpetrogalli
Subscribers: tschuett, hiraditya, rkruppe, psnobl, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78601
Summary:
This simplifies testing in scenarios where we want to set up module-wide
analyses for inlining. The patch enables treating inlining and its
function cleanups, as a module pass. The alternative would be for tests
to describe the pipeline, which is tedious and adds maintenance
overhead.
Reviewers: davidxl, dblaikie, jdoerfert, sstefan1
Subscribers: hiraditya, steven_wu, dexonsmith, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78512
This patch upstreams support for the Armv8.6-a Matrix Multiplication
Extension. A summary of the features can be found here:
https://community.arm.com/developer/ip-products/processors/b/processors-ip-blog/posts/arm-architecture-developments-armv8-6-a
This patch includes:
- Command line options to enable these features with +i8mm, +f32mm, or f64mm
Note: +f32mm and +f64mm are optional and so are not enabled by default
This is part of a patch series, starting with BFloat16 support and
the other components in the armv8.6a extension (in previous patches
linked in phabricator)
Based on work by:
- Luke Geeson
- Oliver Stannard
- Luke Cheeseman
Reviewers: t.p.northover, DavidSpickett
Reviewed By: DavidSpickett
Subscribers: DavidSpickett, ostannard, kristof.beyls, danielkiss,
cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D77875
This patch upstreams support for the Armv8.6-a Matrix Multiplication
Extension. A summary of the features can be found here:
https://community.arm.com/developer/ip-products/processors/b/processors-ip-blog/posts/arm-architecture-developments-armv8-6-a
This patch includes:
- Assembly support for AArch32
- Intrinsics Support for AArch32 Neon Intrinsics for Matrix
Multiplication
Note: these extensions are optional in the 8.6a architecture and so have
to be enabled by default
No additional IR types or C Types are needed for this extension.
This is part of a patch series, starting with BFloat16 support and
the other components in the armv8.6a extension (in previous patches
linked in phabricator)
Based on work by:
- Luke Geeson
- Oliver Stannard
- Luke Cheeseman
Reviewers: t.p.northover, miyuki
Reviewed By: miyuki
Subscribers: miyuki, ostannard, kristof.beyls, hiraditya, danielkiss,
cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D77872
This patch upstreams support for the Armv8.6-a Matrix Multiplication
Extension. A summary of the features can be found here:
https://community.arm.com/developer/ip-products/processors/b/processors-ip-blog/posts/arm-architecture-developments-armv8-6-a
This patch includes:
- Assembly support for AArch64 only (no SVE or Neon)
- Intrinsics Support for AArch64 Armv8.6a Matrix Multiplication Instructions (No bfloat16 matrix multiplication)
No IR types or C Types are needed for this extension.
This is part of a patch series, starting with BFloat16 support and
the other components in the armv8.6a extension (in previous patches
linked in phabricator)
Based on work by:
- Luke Geeson
- Oliver Stannard
- Luke Cheeseman
Reviewers: ostannard, t.p.northover, rengolin, kmclaughlin
Reviewed By: kmclaughlin
Subscribers: kmclaughlin, kristof.beyls, hiraditya, danielkiss,
cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D77871
There's an ABI breakage here if LLVM is compiled in C++14 without
aligned allocation and a user tries to use the result with aligned
allocation. If DenseMap or unique_function is used across that ABI
boundary it will break (PR45413). Moving it out of line is a bit of
a band-aid and LLVM doesn't really give ABI guarantees at this level,
but given the number of complaints I've received over this it still
seems worth fixing.
One of transforms the loop vectorizer makes is LCSSA formation. In some cases it
is the only transform it makes. We should not drop CFG analyzes if only LCSSA was
formed and no actual CFG changes was made.
We should think of expanding this logic to other passes as well, and maybe make
it a part of PM framework.
Reviewed By: Florian Hahn
Differential Revision: https://reviews.llvm.org/D78360
MCELFObjectWriter::setRType## methods are always used altogether to
build complete MIPS N64 ABI "chain" of relocations. Using single
function for this task makes code less verbose.
Since every AbstractAttribute so far, and for the foreseeable future,
corresponds to a single IRPosition we can simplify the class structure.
We already did this for IRAttribute but there is no reason to stop
there.
Summary:
* This is a property of the instance of VectorType. If
isa<ScalableVectorType>(T) is true, then T->isScalable() would have
returned true and vice-versa. Most code that checks this function uses
the result to bail out if a vector is a scalable vector. This code will
be cleaner if it just calls isa<ScalableVectorType>(T)
Reviewers: efriedma, craig.topper, huntergr, sdesmalen
Reviewed By: sdesmalen
Subscribers: tschuett, rkruppe, psnobl, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77690
Summary:
Given a VL=14 that is enveloped by a proper VL=16, splitting the
masked load using the enveloping halving VL=8/8 should yields
should eventually yield V=8/5. This fixes various assert failures
in getHalfNumVectorElementsVT() and IncrementMemoryAddress().
Note, I suspect similar fixes will be needed for other masked
operations, but for now I send out a fix for masked load only.
Bugzilla issue 45563
https://bugs.llvm.org/show_bug.cgi?id=45563
Reviewers: craig.topper, mehdi_amini, nicolasvasilache
Reviewed By: craig.topper
Subscribers: hiraditya, dmgreen, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78608
This is a code clean up of the PropertyAttributeKind and
ObjCPropertyAttributeKind enums in ObjCPropertyDecl and ObjCDeclSpec that are
exactly identical. This non-functional change consolidates these enums
into one. The changes are to many files across clang (and comments in LLVM) so
that everything refers to the new consolidated enum in DeclObjCCommon.h.
2nd Landing Attempt...
Differential Revision: https://reviews.llvm.org/D77233
Summary:
llvm::getInlineCost starts off by determining whether inlining should
happen or not because of user directives or easily determinable
unviability. This CL refactors this functionality as a reusable API.
Reviewers: davidxl, eraman
Reviewed By: davidxl, eraman
Subscribers: hiraditya, haicheng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73825
Summary:
Piggy-back off of TypeSize's STRICT_FIXED_SIZE_VECTORS flag and:
- if it is defined, assert that the vector is not scalable
- if it is not defined, complain if the vector is scalable
Reviewers: efriedma, sdesmalen, c-rhodes
Reviewed By: sdesmalen
Subscribers: hiraditya, mgorny, tschuett, rkruppe, psnobl, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78576
It currently requires that the condition match the shape of the selected value, but this is only really useful for things like masks. This revision allows for the use of i1 to mean that all of the vector/tensor is selected. This also matches the behavior of LLVM select. A benefit of this change is that transformations that want to generate selects, like those on the CFG, don't have to special case vector/tensor. Previously the only way to generate a select from an i1 was to use a splat, but that doesn't support dynamically shaped/unranked tensors.
Differential Revision: https://reviews.llvm.org/D78690
This revision adds support for canonicalizing the following:
```
cond_br %cond, ^bb1(A, ..., N), ^bb1(A, ..., N)
br ^bb1(A, ..., N)
```
If the operands to the successor are different and the cond_br is the only predecessor, we emit selects for the branch operands.
```
cond_br %cond, ^bb1(A), ^bb1(B)
%select = select %cond, A, B
br ^bb1(%select)
```
Differential Revision: https://reviews.llvm.org/D78682
The motivation is to be able to play with the option and change if it is required.
Reviewers: fedor.sergeev, apilipenko, rnk, jdoerfert
Reviewed By: fedor.sergeev
Subscribers: hiraditya, dantrushin, llvm-commits
Differential Revision: https://reviews.llvm.org/D78624
This patch changes the FP conversion intrinsics to take a predicate
that matches the number of lanes for the vector with the widest element
type as opposed to using <vscale x 16 x i1>.
For example:
```<vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float>, <vscale x 4 x i1>, <vscale x 8 x half>)```
now uses <vscale x 4 x i1> instead of <vscale x 16 x i1>
And similar for:
```<vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float>, <vscale x 2 x i1>, <vscale x 2 x double>)```
where the predicate now matches the wider type, so <vscale x 2 x i1>.
Reviewers: efriedma, SjoerdMeijer, paulwalker-arm, rengolin
Reviewed By: efriedma
Tags: #clang
Differential Revision: https://reviews.llvm.org/D78402
This is a code clean up of the PropertyAttributeKind and
ObjCPropertyAttributeKind enums in ObjCPropertyDecl and ObjCDeclSpec that are
exactly identical. This non-functional change consolidates these enums
into one. The changes are to many files across clang (and comments in LLVM) so
that everything refers to the new consolidated enum in DeclObjCCommon.h.
Differential Revision: https://reviews.llvm.org/D77233
Summary:
This fixes several issues where the presence of debug instructions could
disable certain combines, due to dominance queries finding uses/defs that
don't actually exist.
Reviewers: dsanders, fhahn, paquette, aemerson
Subscribers: hiraditya, arphaman, aprantl, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78253
Summary:
These helpers are exercised by follow-up commits in this patch series,
which is all about removing CodeGen differences with vs. without debug
info in the AArch64 backend.
Reviewers: fhahn, aprantl, jpaquette, paquette
Subscribers: kristof.beyls, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78260
Summary:
Make GenericScheduler compute SchedDFSResult on initialization if
the policy is set. This makes it possible to create classes
that extend GenericScheduler and rely on the results of SchedDFSResult,
e.g. to perform subtree scheduling.
NFC unless the policy is set.
Subscribers: MatzeB, hiraditya, javed.absar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78432
Summary:
This is RFC for fixes in poison-related functions of ValueTracking.
These functions assume that a value can be poison bitwisely, but the semantics
of bitwise poison is not clear at the moment.
Allowing a value to have bitwise poison adds complexity to reasoning about
correctness of optimizations.
This patch makes the analysis functions simply assume that a value is
either fully poison or not, which has been used to understand the correctness
of a few previous optimizations.
The bitwise poison semantics seems to be only used by these functions as well.
In terms of implementation, using value-wise poison concept makes existing
functions do more precise analysis, which is what this patch contains.
Reviewers: spatel, lebedev.ri, jdoerfert, reames, nikic, nlopes, regehr
Reviewed By: nikic
Subscribers: fhahn, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78503
Summary:
This is RFC for fixes in poison-related functions of ValueTracking.
These functions assume that a value can be poison bitwisely, but the semantics
of bitwise poison is not clear at the moment.
Allowing a value to have bitwise poison adds complexity to reasoning about
correctness of optimizations.
This patch makes the analysis functions simply assume that a value is
either fully poison or not, which has been used to understand the correctness
of a few previous optimizations.
The bitwise poison semantics seems to be only used by these functions as well.
In terms of implementation, using value-wise poison concept makes existing
functions do more precise analysis, which is what this patch contains.
Reviewers: spatel, lebedev.ri, jdoerfert, reames, nikic, nlopes, regehr
Reviewed By: nikic
Subscribers: fhahn, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78503
This patch changes Block::removeEdge to return a valid iterator to the new next
element, and uses this to update the edge removal algorithm in
LinkGraph::splitBlock.
Add support for relocations for MachO to ObjectYAML / yaml2obj / obj2yaml.
Test plan: make check-all
Differential revision: https://reviews.llvm.org/D77844
Context:
/// Double the size of the allocated memory, guaranteeing space for at
/// least one more element or MinSize if specified.
void grow(size_t MinSize = 0) { this->grow_pod(MinSize, sizeof(T)); }
void push_back(const T &Elt) {
if (LLVM_UNLIKELY(this->size() >= this->capacity()))
this->grow();
memcpy(reinterpret_cast<void *>(this->end()), &Elt, sizeof(T));
this->set_size(this->size() + 1);
}
When grow is called in push_back() without a MinSize specified, this is
relying on the guarantee of space for at least one more element.
There is an edge case bug where the SmallVector is already at its maximum size
and push_back() calls grow() with default MinSize of zero. Grow is unable to
provide space for one more element, but push_back() assumes the additional
element it will be available. This can result in silent memory corruption, as
this->end() will be an invalid pointer and the program may continue executing.
Another alternative to fix would be to remove the default argument from
grow(), which would mean several changing grow() to grow(this->size()+1)
in several places.
No test case added because it would require allocating ~4GB.
Reviewers: echristo
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77601
the tests pass on Linux.
Summary:
This change implements readFromExe, and calculating VA and RVA, which
are some of the functionalities that will be used for native PDB reading
for llvm symbolizer.
bug: https://bugs.llvm.org/show_bug.cgi?id=41795
Summary:
* VectorType::getBitWidth() is just an unsafe version of
getPrimitiveSizeInBits() that assumes all vectors are fixed width.
Reviewers: efriedma, sdesmalen, huntergr, craig.topper
Reviewed By: efriedma
Subscribers: tschuett, hiraditya, rkruppe, psnobl, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77833
If we have a dependence between an abstract attribute A to an abstract
attribute B such hat changes in A should trigger an update of B, we do
not need to keep the dependence around once the update was triggered. If
the dependence is still required the update will reinsert it into the
dependence map, if it is not we avoid triggering B in the future. This
replaces the "recompute interval" mechanism we used before to prune
stale dependences.
Number of required iterations is generally down, compile time for the
module pass (not really the CGSCC pass) is down quite a bit.
There is one test change which looks like an artifact in the undefined
behavior AA that needs to be looked at.
The old command line option `-attributor-disable` was too coarse grained
as we want to measure the effects of the module or cgscc pass without
the other as well.
Since `none` is the default there is no real functional change.
Reviewed By: lebedev.ri
Differential Revision: https://reviews.llvm.org/D78571
Summary:
This change implements readFromExe, and calculating VA and RVA, which
are some of the functionalities that will be used for native PDB reading
for llvm symbolizer.
bug: https://bugs.llvm.org/show_bug.cgi?id=41795
Reviewers: hans, amccarth, rnk
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78128
Summary:
Machine Block Frequency Info (MBFI) is being computed but unused in AsmPrinter.
MBFI computation was introduced with PGO change D71149 and then its use was
removed in D71106. No need to keep computing it.
Reviewers: MaskRay, jyknight, skan, yamauchi, davidxl, efriedma, huihuiz
Reviewed By: MaskRay, skan, yamauchi
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78526
xray_instr_map contains absolute addresses of sleds, which are relocated
by `R_*_RELATIVE` when linked in -pie or -shared mode.
By making these addresses relative to PC, we can avoid the dynamic
relocations and remove the SHF_WRITE flag from xray_instr_map. We can
thus save VM pages containg xray_instr_map (because they are not
modified).
This patch changes x86-64 and bumps the sled version to 2. Subsequent
changes will change powerpc64le and AArch64.
Reviewed By: dberris, ianlevesque
Differential Revision: https://reviews.llvm.org/D78082
AbstractAttribute::initialize is used to initialize the deduction and
the object we do not always call it. To make sure we have the option to
initialize the object even if initialize is not called we pass the
Attributor to AbstractAttribute constructors now.
Summary:
This constructor allows us to create a new DWARFDataExtractor which will
only present a subrange of an entire debug section. Since debug sections
typically consist of multiple contributions, it is expected that one
will create a new data extractor for each contribution in order to
avoid unexpectedly running off into the next one.
This is very useful for unifying the flows for detecting parse errors.
Without it, the code needs to consider two very different scenarios:
1. If there is another contribution after the current one, the
DataExtractor functions will just start reading from there. This is
detectable by comparing the current offset against the known
end-of-contribution offset.
2. If this is the last contribution, the data extractor will just start
returning zeroes (or other default values). This situation can *not*
be detected by checking the parsing offset, as this will not be
advanced in case of errors.
Using a truncated data extractor simplifies the code (and reduces
cognitive load) by making these two cases behave identically -- a
running off the end of a contribution will _always_ produce an EOF error
(if one uses error-aware parsing methods) or return default values.
Reviewers: dblaikie, probinson, jhenderson, ikudrin
Subscribers: aprantl, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77556
This API call has been used recently with, a very valid, expectation
that it would do something useful but it doesn't actually query any
backend information. So, remove this method and merge its
functionality into getUserCost. As well as that, also use
getCastInstrCost to get a proper cost from the backend for the
concerned instructions though we only currently return the answer if
it's considered free. The default implementation now also checks
int/ptr conversions too, as well as truncs and bitcasts.
Differential Revision: https://reviews.llvm.org/D76124
This code was added in 887efa51c1 to
fix reverse iteration.
The call to InsertIntoBucket/InsertIntoBucketWithLookup can change
the number of buckets which will invalidate the BucketEnd. So
don't cache it and calculate it when creating the iterator.
Summary:
Before this patch, `relaxInstruction` takes three arguments, the first
argument refers to the instruction before relaxation and the third
argument is the output instruction after relaxation. There are two quite
strange things:
1) The first argument's type is `const MCInst &`, the third
argument's type is `MCInst &`, but they may be aliased to the same
variable
2) The backends of ARM, AMDGPU, RISC-V, Hexagon assume that the third
argument is a fresh uninitialized `MCInst` even if `relaxInstruction`
may be called like `relaxInstruction(Relaxed, STI, Relaxed)` in a
loop.
In this patch, we drop the thrid argument, and let `relaxInstruction`
directly modify the given instruction. Also, this patch fixes the bug https://bugs.llvm.org/show_bug.cgi?id=45580, which is introduced by D77851, and
breaks the assumption of ARM, AMDGPU, RISC-V, Hexagon.
Reviewers: Razer6, MaskRay, jyknight, asb, luismarques, enderby, rtaylor, colinl, bcain
Reviewed By: Razer6, MaskRay, bcain
Subscribers: bcain, nickdesaulniers, nathanchance, wuzish, annita.zhang, arsenm, dschuff, jyknight, dylanmckay, sdardis, nemanjai, jvesely, nhaehnle, tpr, sbc100, jgravelle-google, kristof.beyls, hiraditya, aheejin, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, kerbowa, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78364
We now also use the BumpPtrAllocator from the Attributor in the
InformationCache. The lifetime of objects in either is pretty much the
same and it should result in consistently good performance regardless of
the allocator.
Doing so requires to call more constructors manually but so far that
does not seem to be problematic or messy.
---
Single run of the Attributor module and then CGSCC pass (oldPM)
for SPASS/clause.c (~10k LLVM-IR loc):
Before:
```
calls to allocation functions: 615359 (368257/s)
temporary memory allocations: 83315 (49859/s)
peak heap memory consumption: 75.64MB
peak RSS (including heaptrack overhead): 163.43MB
total memory leaked: 269.04KB
```
After:
```
calls to allocation functions: 613042 (359555/s)
temporary memory allocations: 83322 (48869/s)
peak heap memory consumption: 75.64MB
peak RSS (including heaptrack overhead): 162.92MB
total memory leaked: 269.04KB
```
Difference:
```
calls to allocation functions: -2317 (-68147/s)
temporary memory allocations: 7 (205/s)
peak heap memory consumption: 2.23KB
peak RSS (including heaptrack overhead): 0B
total memory leaked: 0B
---
This patch gets the asserts working correctly when LLVM_REVERSE_ITERATION=On by fixing the iterators returned by the DenseMap::find* methods so that they return well-formed iterators that work with reverse iteration, and satisfy the assertions.
A recent change (4e86e5eedc), broke `LLVM_REVERSE_ITERATION` for DenseMaps by adding an assert. It is valid to de-reference and increment one step behind `End` when reverse iteration is enabled because `End` is actually the start of the pointer bucket.
With clang option -funique-internal-linkage-symbols, symbols with
internal linkage get names with the module hash appended.
Differential Revision: https://reviews.llvm.org/D78243
- Adding changes to support comments on outlined functions with outlining for the conditions through which it was outlined (e.g. Thunks, Tail calls)
- Adapts the emitFunctionHeader to print out a comment next to the header if the target specifies it based on information in MachineFunctionInfo
- Adds mir test for function annotiation
Differential Revision: https://reviews.llvm.org/D78062
Following the rule of five, declare move constructor and move
assignment operator for ValueLatticeElement. This allows moving
the ConstantRange rather than copying it.
This does not matter in most cases, where we're dealing with
APInts <= 64 bits. It does avoid unnecessary copies of allocations
for larger APInts.
Additionally we change the implementation approach to make the
copy/move assignment operators make use of the copy/move constructors,
rather than the other way around. The constructors are the more
primitive operations.
Differential Revision: https://reviews.llvm.org/D78425
Use Instruction::comesBefore() instead of OrderedInstructions
inside InstructionPrecedenceTracking. This also removes the
dominator tree dependency.
Differential Revision: https://reviews.llvm.org/D78461
Summary:
Similar to the CallLowering class used for lowering LLVM IR calls to MIR calls,
we introduce a separate class for lowering LLVM IR inline asm to MIR INLINEASM.
There is no functional change yet, all existing tests should pass.
Reviewers: arsenm, dsanders, aemerson, volkan, t.p.northover, paquette
Reviewed By: aemerson
Subscribers: gargaroff, wdng, mgorny, rovka, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78316
There are few `std::vector<std::string>` members in
`FileCheckRequest`. This patch changes these arrays to `std::vector<StringRef>`
and refactors the code related to cleanup/improve/simplify it.
Differential revision: https://reviews.llvm.org/D78202
The API for shuffles and reductions uses generic Type parameters,
instead of VectorType, and so assertions and casts are used a lot.
This patch makes those types explicit, which means that the clients
can't be lazy, but results in less ambiguity, and that can only be a
good thing.
Bugzilla: https://bugs.llvm.org/show_bug.cgi?id=45562
Differential Revision: https://reviews.llvm.org/D78357
Some includes are not required and forward declarations can be used
instead. This also exposed a few places that were not directly including
required files.
Most of the includes in LoopUtils.h are not required in the header and
they can be replaced by forward declarations.
Unfortunately includes of TargetTransformInfo.h and IVDescriptors.h pull
in a bunch of additional things, but there is no easy way to get rid of
them at the moment I think.
sched_getaffinity (Linux specific) has been available
* in glibc since 2002-08-08 (commit 972e719e8154eec5f543b027e2a08dfa285d55d5)
* in musl since the initial check-in.
We were only including MachineInstr.h for DebugLoc.h. This exposes an implicit include dependency in BTFDebug.h where I've had to add the MachineInstr.h include.
This makes it easier to extend the merge options in the future and also
reduces the risk of accidentally setting a wrong option.
Reviewers: efriedma, nikic, reames, davide
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D78368
Summary:
Add getBase accessor so that underlying tree can be
manipulated in a similar manner to MachineDominatorTree.
Reviewers: kuhar, arsenm, hliao, nhaehnle
Reviewed By: kuhar
Subscribers: lkail, mgorny, hiraditya, wdng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77967
LLJIT::defineAbsolute did not mangle its Name argument, which is inconsistent
with the behavior of other LLJIT methods (e.g. lookup). Since it is currently
unused anyway, this commit replaces it with a generic 'define' convenience
method for adding MaterializationUnits to the main JITDylib. This simplifies
use of the generic absoluteSymbols function (as well as the symbolAlias,
reexports and other functions that generate MaterializationUnits) with LLJIT.
ConstVal is not an owned pointer, so setting it to nullptr is not
actually doing anything. If we switch to a state that does not use
ConstVal, the value does not matter.
Split out from D78425.
When running IPSCCP on a module with many small functions, memory
usage is dominated by PredicateInfo, which is a huge structure
(partially due to some unfortunate nested SmallVector use). However,
most of it is actually only temporary state needed to build
predicate info, and does not need to be retained after initial
construction.
This patch factors out the predicate building logic and state
into a separate PrediceInfoBuilder, with the extra bonus that
it does not need to live in the header anymore.
Differential Revision: https://reviews.llvm.org/D78326
Cost-modeling decisions are tied to the compute interleave groups
(widening decisions, scalar and uniform values). When invalidating the
interleave groups, those decisions also need to be invalidated.
Otherwise there is a mis-match during VPlan construction.
VPWidenMemoryRecipes created initially are left around w/o converting them
into VPInterleave recipes. Such a conversion indeed should not take place,
and these gather/scatter recipes may in fact be right. The crux is leaving around
obsolete CM_Interleave (and dependent) markings of instructions along with
their costs, instead of recalculating decisions, costs, and recipes.
Alternatively to forcing a complete recompute later on, we could try
to selectively invalidate the decisions connected to the interleave
groups. But we would likely need to run the uniform/scalar value
detection parts again anyways and the extra complexity is probably not
worth it.
Fixes PR45572.
Reviewers: gilr, rengolin, Ayal, hsaito
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D78298
BitVectors and SmallBitVectors with equal contents but different
capacities were getting different hashes.
Reviewed By: aganea
Differential Revision: https://reviews.llvm.org/D77038
SmallVector currently uses 32bit integers for size and capacity to reduce
sizeof(SmallVector). This limits the number of elements to UINT32_MAX.
For a SmallVector<char>, this limits the SmallVector size to only 4GB.
Buffering bitcode output uses SmallVector<char>, but needs >4GB output.
This changes SmallVector size and capacity to conditionally use word-size
integers if the element type is small (<4 bytes). For larger elements types,
the vector size can reach ~16GB with 32bit size.
Making this conditional on the element type provides both the smaller
sizeof(SmallVector) for larger types which are unlikely to grow so large,
and supports larger capacities for smaller element types.
This change also includes a fix for the bug where a SmallVector with 32bit
size has reached UINT32_MAX elements, and cannot provide guaranteed growth.
Context:
// Double the size of the allocated memory, guaranteeing space for at
// least one more element or MinSize if specified.
void grow(size_t MinSize = 0) { this->grow_pod(MinSize, sizeof(T)); }
void push_back(const T &Elt) {
if (LLVM_UNLIKELY(this->size() >= this->capacity()))
this->grow();
memcpy(reinterpret_cast<void *>(this->end()), &Elt, sizeof(T));
this->set_size(this->size() + 1);
}
When grow is called in push_back() without a MinSize specified, this is
relying on the guarantee of space for at least one more element.
There is an edge case bug where the SmallVector is already at its maximum size
and push_back() calls grow() with default MinSize of zero. Grow is unable to
provide space for one more element, but push_back() assumes the additional
element it will be available. This can result in silent memory corruption, as
this->end() will be an invalid pointer and the program may continue executing.
An alternative to this fix would be to remove the default argument from
grow(), which would mean several changing grow() to grow(this->size()+1)
in several places.
No test case added because it would require allocating a large ammount.
Differential Revision: https://reviews.llvm.org/D77621
Summary:
Remove asserting vector getters from Type in preparation for the
VectorType refactor. The existence of these functions complicates the
refactor while adding little value.
Reviewers: dexonsmith, sdesmalen, efriedma
Reviewed By: efriedma
Subscribers: cfe-commits, hiraditya, llvm-commits
Tags: #llvm, #clang
Differential Revision: https://reviews.llvm.org/D77278
Summary:
Use a SmallSetVector instead of a SmallPtrSet when collecting
and storing Roots.
The iteration order for a SmallPtrSet is not deterministic,
so in the past the order of items inserted in the WorkList
inside walkBackwards has been non-deterministic. This patch
intends to make the order of rewrites done in Float2Int
deterministic by changing the container for the Roots set.
The semantics result of the transformation should not be
any different afaict. But at least naming of IR variables
(when outputting the result as an ll file) should be more
stable now.
Reviewers: craig.topper, spatel, cameron.mcinally
Reviewed By: spatel
Subscribers: mgrang, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D74534
There are also some adjustments to use MaybeAlign in here due
to CallBase::getParamAlignment() being deprecated. It would
be a little cleaner if getOrEnforceKnownAlignment was migrated
to Align/MaybeAlign.
Differential Revision: https://reviews.llvm.org/D78345
There are also some adjustments to use MaybeAlign in here due
to CallBase::getParamAlignment() being deprecated. It would
be cleaner if getOrEnforceKnownAlignment was migrated
to Align/MaybeAlign.
Differential Revision: https://reviews.llvm.org/D78345
PredicateInfo takes up a large amount of memory during IPSCCP
with many functions. And a large part of that space seems to
be going completely to waste here...
Add initial support for PC Relative addressing for global values that
require GOT indirect addressing. This patch adds PCRelative support for
global addresses that may not be known at link time and may require
access through the GOT.
Differential Revision: https://reviews.llvm.org/D76064
Users of ValueLatticeElement currently have to ensure constant ranges
are not extended indefinitely. For example, in SCCP, mergeIn goes to
overdefined if a constantrange value is repeatedly merged with larger
constantranges. This is a simple form of widening.
In some cases, this leads to an unnecessary loss of information and
things can be improved by allowing a small number of extensions in the
hope that a fixed point is reached after a small number of steps.
To make better decisions about widening, it is helpful to keep track of
the number of range extensions. That state is tied directly to a
concrete ValueLatticeElement and some unused bits in the class can be
used. The current patch preserves the existing behavior by default:
CheckWiden defaults to false and if CheckWiden is true, a single change
to the range is allowed.
Follow-up patches will slightly increase the threshold for widening.
Reviewers: efriedma, davide, mssimpso
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D78145
Summary:
AIX symbol have qualname and unqualified name. The stock getSymbol
could only return unqualified name, which leads us to patch many
caller side(lowerConstant, getMCSymbolForTOCPseudoMO).
So we should try to address this problem in the callee
side(getSymbol) and clean up the caller side instead.
Note: this is a "mostly" NFC patch, with a fix for the original
lowerConstant behavior.
Differential Revision: https://reviews.llvm.org/D78045
This allows targets to know exactly which operands are contributing to
the dependency, which is required for targets with per-operand
scheduling models.
Differential Revision: https://reviews.llvm.org/D77135
Craig Topper