This patch removes the IsPairwiseForm flag from the Reduction Cost TTI
hooks, along with some accompanying code for pattern matching reductions
from trees starting at extract elements. IsPairWise is now assumed to be
false, which was the predominant way that the value was used from both
the Loop and SLP vectorizers. Since the adjustments such as D93860, the
SLP vectorizer has not relied upon this distinction between paiwise and
non-pairwise reductions.
This also removes some code that was detecting reductions trees starting
from extract elements inside the costmodel. This case was
double-counting costs though, adding the individual costs on the
individual instruction _and_ the total cost of the reduction. Removing
it changes the costs in llvm/test/Analysis/CostModel/X86/reduction.ll to
not double count. The cost of reduction intrinsics is still tested
through the various tests in
llvm/test/Analysis/CostModel/X86/reduce-xyz.ll.
Differential Revision: https://reviews.llvm.org/D105484
This reverts commit e2d30846327c7ec5cc9d2a46aa9bcd9c2c4eff93.
MSVC doesn't seem to resolve the intended ambiguity in implicit
conversion contexts correctly: https://godbolt.org/z/ee16aqv4v
See bug for full details, but basically there's an upcoming ambiguity in
the conversion in `StringRef(SomeSmallString)` - either the implicit
conversion operator (SmallString::operator StringRef) could be used, or
the std::string_view range-based ctor (& then `StringRef(std::string_view)`
would be used)
To address this, make such a conversion invalid up-front - most uses are
more tersely written as `SomeSmallString.str()` anyway, or more clearly
written as `StringRef x = y;` rather than `StringRef x(y);` - so if you
hit this in out-of-tree code, please update in one of those ways.
Hopefully I've fixed everything in tree prior to this patch landing.
SelectionDAG's equivalents in ISD::InputArg/OutputArg track the
original argument index. Mips relies on this, and its currently
reinventing its own parallel CallLowering infrastructure which tracks
these indexes on the side. Add this to help move towards deleting the
custom mips handling.
There are two issues with the current implementation of computeBECount:
1. It doesn't account for the possibility that adding "Stride - 1" to
Delta might overflow. For almost all loops, it doesn't, but it's not
actually proven anywhere.
2. It doesn't account for the possibility that Stride is zero. If Delta
is zero, the backedge is never taken; the value of Stride isn't
relevant. To handle this, we have to make sure that the expression
returned by computeBECount evaluates to zero.
To deal with this, add two new checks:
1. Use a variety of tricks to try to prove that the addition doesn't
overflow. If the proof is impossible, use an alternate sequence which
never overflows.
2. Use umax(Stride, 1) to handle the possibility that Stride is zero.
Differential Revision: https://reviews.llvm.org/D105216
As pointed out in post-commit review on rG8e22539067d9, it's
necessary to call getScalarType() to support GEPs with a vector
base. Dropping that call was an oversight on my side.
This adds a new llvm::thread class with the same interface as std::thread
except there is an extra constructor that allows us to set the new thread's
stack size. On Darwin even the default size is boosted to 8MB to match the main
thread.
It also switches all users of the older C-style `llvm_execute_on_thread` API
family over to `llvm::thread` followed by either a `detach` or `join` call and
removes the old API.
Moved definition of DefaultStackSize into the .cpp file to hopefully
fix the build on some (GCC-6?) machines.
This adds a new llvm::thread class with the same interface as std::thread
except there is an extra constructor that allows us to set the new thread's
stack size. On Darwin even the default size is boosted to 8MB to match the main
thread.
It also switches all users of the older C-style `llvm_execute_on_thread` API
family over to `llvm::thread` followed by either a `detach` or `join` call and
removes the old API.
Several subclasses of User override operator new without also overriding
operator delete. This means that delete expressions fall back to using
operator delete of the base class, which would be User. However, this is
only allowed if the base class has a virtual destructor which is not the
case for User, so this is UB.
See also [expr.delete] (3) for the exact wording.
This is actually detected in some cases by GCC 11's
-Wmismatched-new-delete now which is how I found this error.
Differential Revision: https://reviews.llvm.org/D103143
ExecutorAddressRange depended on JITTargetAddress, but JITTargetAddress is
defined in ExecutionEngine, which OrcShared should not depend on.
This seems like as good a time as any to introduce a new ExecutorAddress type
to eventually replace JITTargetAddress. For now it's just another uint64_t
alias, but it will soon be changed to a class type to provide greater type
safety.
The computeNamedSymbolDependencies and computeLocalDeps methods on
ObjectLinkingLayerJITLinkContext are responsible for computing, for each symbol
in the current MaterializationResponsibility, the set of non-locally-scoped
symbols that are depended on. To calculate this we have to consider the effect
of chains of dependence through locally scoped symbols in the LinkGraph. E.g.
.text
.globl foo
foo:
callq bar ## foo depneds on external 'bar'
movq Ltmp1(%rip), %rcx ## foo depends on locally scoped 'Ltmp1'
addl (%rcx), %eax
retq
.data
Ltmp1:
.quad x ## Ltmp1 depends on external 'x'
In this example symbol 'foo' depends directly on 'bar', and indirectly on 'x'
via 'Ltmp1', which is locally scoped.
Performance of the existing implementations appears to have been mediocre:
Based on flame graphs posted by @drmeister (in #jit on the LLVM discord server)
the computeLocalDeps function was taking up a substantial amount of time when
starting up Clasp (https://github.com/clasp-developers/clasp).
This commit attempts to address the performance problems in three ways:
1. Using jitlink::Blocks instead of jitlink::Symbols as the nodes of the
dependencies-introduced-by-locally-scoped-symbols graph.
Using either Blocks or Symbols as nodes provides the same information, but since
there may be more than one locally scoped symbol per block the block-based
version of the dependence graph should always be a subgraph of the Symbol-based
version, and so faster to operate on.
2. Improved worklist management.
The older version of computeLocalDeps used a fixed worklist containing all
nodes, and iterated over this list propagating dependencies until no further
changes were required. The worklist was not sorted into a useful order before
the loop started.
The new version uses a variable work-stack, visiting nodes in DFS order and
only adding nodes when there is meaningful work to do on them.
Compared to the old version the new version avoids revisiting nodes which
haven't changed, and I suspect it converges more quickly (due to the DFS
ordering).
3. Laziness and caching.
Mappings of...
jitlink::Symbol* -> Interned Name (as SymbolStringPtr)
jitlink::Block* -> Immediate dependencies (as SymbolNameSet)
jitlink::Block* -> Transitive dependencies (as SymbolNameSet)
are all built lazily and cached while running computeNamedSymbolDependencies.
According to @drmeister these changes reduced Clasp startup time in his test
setup (averaged over a handful of starts) from 4.8 to 2.8 seconds (with
ORC/JITLink linking ~11,000 object files in that time), which seems like
enough to justify switching to the new algorithm in the absence of any other
perf numbers.
MachOJITDylibInitializers::SectionExtent represented the address range of a
section as an (address, size) pair. The new ExecutorAddressRange type
generalizes this to an address range (for any object, not necessarily a section)
represented as a (start-address, end-address) pair.
The aim is to express more of ORC (and the ORC runtime) in terms of simple types
that can be serialized/deserialized via SPS. This will simplify SPS-based RPC
involving arguments/return-values of these types.
These currently always require a type parameter. The bitcode reader
already upgrades old bitcode without the type parameter to use the
pointee type.
In cases where the caller does not have byval but the callee does, we
need to follow CallBase::paramHasAttr() and also look at the callee for
the byval type so that CallBase::isByValArgument() and
CallBase::getParamByValType() are in sync. Do the same for preallocated.
While we're here add a corresponding version for inalloca since we'll
need it soon.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D104663
The Legalizer expands the operations of urem/srem into a div+mul+sub or divrem
when those are legal/custom. This patch changes the cost-model to reflect that
cost.
Since there is no 'divrem' Instruction in LLVM IR, the cost of divrem
is assumed to be the same as div+mul+sub since the three operations will
need to be executed at runtime regardless.
Patch co-authored by David Sherwood (@david-arm)
Reviewed By: RKSimon, paulwalker-arm
Differential Revision: https://reviews.llvm.org/D103799
Before we replaced value by registering all their uses. However, as we
replace a value old uses become stale. We now replace values explicitly
and keep track of "new values" when doing so to avoid replacing only
uses in stale/old values but not their replacements.
We often need to deal with the value lattice that contains none and
undef as special values. A simple helper makes this much nicer.
Differential Revision: https://reviews.llvm.org/D103857
When we do simplification via AAPotentialValues or AAValueConstantRange
we need to simplify the operands of an instruction we deconstruct first.
This does not only improve the result, see for example range.ll, but is
required as we allow outside AAs to provide simplification rules via
callbacks. If we do ignore the simplification rules and base other
simplifications on the IR instead we can create an inconsistent state.
As part of making ScalarEvolution's handling of pointers consistent, we
want to forbid multiplying a pointer by -1 (or any other value). This
means we can't blindly subtract pointers.
There are a few ways we could deal with this:
1. We could completely forbid subtracting pointers in getMinusSCEV()
2. We could forbid subracting pointers with different pointer bases
(this patch).
3. We could try to ptrtoint pointer operands.
The option in this patch is more friendly to non-integral pointers: code
that works with normal pointers will also work with non-integral
pointers. And it seems like there are very few places that actually
benefit from the third option.
As a minimal patch, the ScalarEvolution implementation of getMinusSCEV
still ends up subtracting pointers if they have the same base. This
should eliminate the shared pointer base, but eventually we'll need to
rewrite it to avoid negating the pointer base. I plan to do this as a
separate step to allow measuring the compile-time impact.
This doesn't cause obvious functional changes in most cases; the one
case that is significantly affected is ICmpZero handling in LSR (which
is the source of almost all the test changes). The resulting changes
seem okay to me, but suggestions welcome. As an alternative, I tried
explicitly ptrtoint'ing the operands, but the result doesn't seem
obviously better.
I deleted the test lsr-undef-in-binop.ll becuase I couldn't figure out
how to repair it to test what it was actually trying to test.
Recommitting with fix to MemoryDepChecker::isDependent.
Differential Revision: https://reviews.llvm.org/D104806
As part of making ScalarEvolution's handling of pointers consistent, we
want to forbid multiplying a pointer by -1 (or any other value). This
means we can't blindly subtract pointers.
There are a few ways we could deal with this:
1. We could completely forbid subtracting pointers in getMinusSCEV()
2. We could forbid subracting pointers with different pointer bases
(this patch).
3. We could try to ptrtoint pointer operands.
The option in this patch is more friendly to non-integral pointers: code
that works with normal pointers will also work with non-integral
pointers. And it seems like there are very few places that actually
benefit from the third option.
As a minimal patch, the ScalarEvolution implementation of getMinusSCEV
still ends up subtracting pointers if they have the same base. This
should eliminate the shared pointer base, but eventually we'll need to
rewrite it to avoid negating the pointer base. I plan to do this as a
separate step to allow measuring the compile-time impact.
This doesn't cause obvious functional changes in most cases; the one
case that is significantly affected is ICmpZero handling in LSR (which
is the source of almost all the test changes). The resulting changes
seem okay to me, but suggestions welcome. As an alternative, I tried
explicitly ptrtoint'ing the operands, but the result doesn't seem
obviously better.
I deleted the test lsr-undef-in-binop.ll becuase I couldn't figure out
how to repair it to test what it was actually trying to test.
Differential Revision: https://reviews.llvm.org/D104806
This patch emits DBG_INSTR_REFs for two remaining flavours of variable
locations that weren't supported: copies, and inter-block VRegs. There are
still some locations that must be represented by DBG_VALUE such as
constants, but they're mostly independent of optimisations.
For variable locations that refer to values defined in different blocks,
vregs are allocated before isel begins, but the defining instruction
might not exist until late in isel. To get around this, emit
DBG_INSTR_REFs in a "half done" state, where the first operand refers to a
VReg. Then at the end of isel, patch these back up to refer to
instructions, using the finalizeDebugInstrRefs method.
Copies are something that I complained about the original RFC, and I
really don't want to have to put instruction numbers on copies. They don't
define a value: they move them. To address this isel, salvageCopySSA
interprets:
* COPYs,
* SUBREG_TO_REG,
* Anything that isCopyInstr thinks is a copy.
And follows chains of copies back to the defining instruction that they
read from. This relies on any physical registers that COPYs read being
defined in the same block, or being entry-block arguments. For the former
we can put an instruction number on the defining instruction; for the
latter we can drop a DBG_PHI that reads the incoming value.
Differential Revision: https://reviews.llvm.org/D88896
This patch adds a TTI function, isElementTypeLegalForScalableVector, to query
whether it is possible to vectorize a given element type. This is called by
isLegalToVectorizeInstTypesForScalable to reject scalable vectorization if
any of the instruction types in the loop are unsupported, e.g:
int foo(__int128_t* ptr, int N)
#pragma clang loop vectorize_width(4, scalable)
for (int i=0; i<N; ++i)
ptr[i] = ptr[i] + 42;
This example currently crashes if we attempt to vectorize since i128 is not a
supported type for scalable vectorization.
Reviewed By: sdesmalen, david-arm
Differential Revision: https://reviews.llvm.org/D102253
This patch implaments the load and reserve and store conditional
builtins for the PowerPC target, in order to have feature parody with
xlC on AIX.
Differential revision: https://reviews.llvm.org/D105236
This patch adds a new ShuffleKind SK_Splice and then handle the cost in
getShuffleCost, as in experimental.vector.reverse.
Differential Revision: https://reviews.llvm.org/D104630
Summary: The patch adds the StringTable dumping to
llvm-readobj. Currently only XCOFF is supported.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D104613
This API is not compatible with opaque pointers, the method
accepting an explicit pointer element type should be used instead.
Thankfully there were few in-tree users. The BPF case still ends
up using the pointer element type for now and needs something like
D105407 to avoid doing so.
Same as other CreateLoad-style APIs, these need an explicit type
argument to support opaque pointers.
Differential Revision: https://reviews.llvm.org/D105395
Compiling LLVM with Clang modules and libc++ identified that
`Support/Printable.h` and `ADL/SmallVector.h` were using features that
live in these headers.
Differential Revision: https://reviews.llvm.org/D105402
This reverts commit 8cd35ad854ab4458fd509447359066ea3578b494.
It breaks `TestMembersAndLocalsWithSameName.py` on GreenDragon and
Mikael Holmén points out in D104827 that bitcode files created with the
patch cannot be parsed with binaries built before it.
We're trying to match a few pointer computation patterns here for
re-association opportunities.
1) Isolating a constant operand to be on the RHS, e.g.:
G_PTR_ADD(BASE, G_ADD(X, C)) -> G_PTR_ADD(G_PTR_ADD(BASE, X), C)
2) Folding two constants in each sub-tree as long as such folding
doesn't break a legal addressing mode.
G_PTR_ADD(G_PTR_ADD(BASE, C1), C2) -> G_PTR_ADD(BASE, C1+C2)
AArch64 code size improvements on CTMark with -Os:
Program before after diff
pairlocalalign 251048 251044 -0.0%
consumer-typeset 421820 421812 -0.0%
kc 431348 431320 -0.0%
SPASS 413404 413300 -0.0%
clamscan 384396 384220 -0.0%
tramp3d-v4 370640 370412 -0.1%
lencod 432096 431772 -0.1%
bullet 479400 478796 -0.1%
sqlite3 288504 288072 -0.1%
7zip-benchmark 573796 570768 -0.5%
Geomean difference -0.1%
Differential Revision: https://reviews.llvm.org/D105069
This change yields an additional 2% size reduction on an internal search
binary, and an additional 0.5% size reduction on fuchsia.
Differential Revision: https://reviews.llvm.org/D104751
Add a flag so that target can choose to use AsmParser for parsing inline asm.
And set the flag by default for AIX.
-no-intergrated-as will override this default if specified explicitly.
Reviewed By: #powerpc, shchenz
Differential Revision: https://reviews.llvm.org/D105314
While this should not matter for most architectures (where the program
address space is 0), it is important for CHERI (and therefore Arm Morello).
We use address space 200 for all of our code pointers and without this
change we assert in the SelectionDAG handling of BlockAddress nodes.
It is also useful for AVR: previously programs targeting
AVR that attempt to read their own machine code
via a pointer to a label would instead read from RAM
using a pointer relative to the the start of program flash.
Reviewed By: dylanmckay, theraven
Differential Revision: https://reviews.llvm.org/D48803
Reland of 31859f896.
This change implements new DAG notes GLOBAL_GET/GLOBAL_SET, and
lowering methods for load and stores of reference types from IR
globals. Once the lowering creates the new nodes, tablegen pattern
matches those and converts them to Wasm global.get/set.
Differential Revision: https://reviews.llvm.org/D104797
