If computeKnownBits encounters a phi node, and we fail to determine any known bits through direct analysis, see if the incoming value is part of a branch condition feeding the phi.
Handle cases where icmp(IncomingValue PRED Constant) is driving a branch instruction feeding that phi node - at the moment this only handles EQ/ULT/ULE predicate cases as they are the most straightforward to handle and most likely for branch-loop 'max upper bound' cases - we can extend this if/when necessary.
I investigated a more general icmp(LHS PRED RHS) KnownBits system, but the hard limits we put on value tracking depth through phi nodes meant that we were mainly catching constants anyhow.
Fixes the pointless vectorization in PR38280 / Issue #37628 (excessive unrolling still needs handling though)
Differential Revision: https://reviews.llvm.org/D131838
We call tail-call-elim near the beginning of the pipeline,
but that is too early to annotate calls that get added later.
In the motivating case from issue #47852, the missing 'tail'
on memset leads to sub-optimal codegen.
I experimented with removing the early instance of
tail-call-elim instead of just adding another pass, but that
appears to be slightly worse for compile-time:
+0.15% vs. +0.08% time.
"tailcall" shows adding the pass; "tailcall2" shows moving
the pass to later, then adding the original early pass back
(so 1596886802 is functionally equivalent to 180b0439dc ):
https://llvm-compile-time-tracker.com/index.php?config=NewPM-O3&stat=instructions&remote=rotateright
Note that there was an effort to split the tail call functionality
into 2 passes - that could help reduce compile-time if we find
that this change costs more in compile-time than expected based
on the preliminary testing:
D60031
Differential Revision: https://reviews.llvm.org/D130374
When F calls G calls H, G is nounwind, and G is inlined into F, then the
inlined call-site to H should be effectively nounwind so as not to lose
information during inlining.
If H itself is nounwind (which often happens when H is an intrinsic), we
no longer mark the callsite explicitly as nounwind. Previously, there
were cases where the inlined call-site of H differs from a pre-existing
call-site of H in F *only* in the explicitly added nounwind attribute,
thus preventing common subexpression elimination.
v2:
- just check CI->doesNotThrow
v3 (resubmit after revert at 344378808778c61d5599f4e0ac783ef7e6f8ed05):
- update Clang tests
Differential Revision: https://reviews.llvm.org/D129860
When F calls G calls H, G is nounwind, and G is inlined into F, then the
inlined call-site to H should be effectively nounwind so as not to lose
information during inlining.
If H itself is nounwind (which often happens when H is an intrinsic), we
no longer mark the callsite explicitly as nounwind. Previously, there
were cases where the inlined call-site of H differs from a pre-existing
call-site of H in F *only* in the explicitly added nounwind attribute,
thus preventing common subexpression elimination.
v2:
- just check CI->doesNotThrow
Differential Revision: https://reviews.llvm.org/D129860
Compiling with '-ffast-math' tuns on all the FastMathFlags (FMF), as
expected, and that enables FP reassociation. Only the two FMF flags
'reassoc' and 'nsz' are technically required to perform reassociation,
but disabling other unrelated FMF bits is needlessly suppressing the
optimization.
This patch fixes that needless suppression, and makes appropriate
adjustments to test-cases, fixing some outstanding TODOs in the process.
Fixes: #56483
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D129523
After replacing a loop phi with the preheader value, it's usually
possible to simplify some of the using instructions, so do that as
part of replaceLoopPHINodesWithPreheaderValues().
Doing this as part of IndVars is valuable, because it may make GEPs
in the loop have constant offsets and allow the following SROA run
to succeed (as demonstrated in the PhaseOrdering test).
Differential Revision: https://reviews.llvm.org/D129293
I looked at canonicalizing in the other direction, but that causes
many potential regressions and infinite loops because we already
(possibly wrongly) canonicalize "trunc X to i1" into an and+icmp.
This has a data layout restriction to avoid creating illegal
mask instructions, but we could remove that if we can show
that the backend can undo this when needed.
The motivating example from issue #56119 is modeled by the
PhaseOrdering test.
The ArgumentPromotion pass uses Mem2Reg promotion at the end to cutting
down generated alloca instructions as well as meaningless stores and
this behavior can leave unused (dead) arguments.
The test shows that the arguments are not removed in the current
optimization pipeline.
This enabled opaque pointers by default in LLVM. The effect of this
is twofold:
* If IR that contains *neither* explicit ptr nor %T* types is passed
to tools, we will now use opaque pointer mode, unless
-opaque-pointers=0 has been explicitly passed.
* Users of LLVM as a library will now default to opaque pointers.
It is possible to opt-out by calling setOpaquePointers(false) on
LLVMContext.
A cmake option to toggle this default will not be provided. Frontends
or other tools that want to (temporarily) keep using typed pointers
should disable opaque pointers via LLVMContext.
Differential Revision: https://reviews.llvm.org/D126689
Now that SimpleLoopUnswitch and other transforms no longer introduce
branch on poison, enable the -branch-on-poison-as-ub option by
default. The practical impact of this is mostly better flag
preservation in SCEV, and some freeze instructions no longer being
necessary.
Differential Revision: https://reviews.llvm.org/D125299
This option was added in D89854. It prevents GVN from performing
load PRE in a loop, if doing so would require critical edge
splitting on the backedge. From the review:
> I know that GVN Load PRE negatively impacts peeling,
> loop predication, so the passes expecting that latch has
> a conditional branch.
In the PhaseOrdering test in this patch, splitting the backedge
negatively affects vectorization: After critical edge splitting,
the loop gets rotated, effectively peeling off the first loop
iteration. The effect is that the first element is handled
separately, then the bulk of the elements use a vectorized
reduction (but using unaligned, off-by-one memory accesses) and
then a tail of 15 elements is handled separately again.
It's probably worth noting that the loop load PRE from D99926 is
not affected by this change (as it does not need backedge
splitting). This is about normal load PRE that happens to occur
inside a loop.
Differential Revision: https://reviews.llvm.org/D126382
Use IRBuilder so that the newly created freeze instructions
automatically gets inserted back into the IC worklist.
The changed worklist processing order leads to some cosmetic
differences in tests.
Fixes https://github.com/llvm/llvm-project/issues/55619.
When computing the BECount for multi-exit loops, we need to combine
individual exit counts using umin_seq rather than umin. This is
because an earlier exit may exit on the first iteration, in which
case later exit expressions will not be evaluated and could be
poisonous. We cannot propagate potential poison values from later
exits.
In particular, this avoids the introduction of "branch on poison"
UB when optimizing multi-exit loops.
Differential Revision: https://reviews.llvm.org/D124910
The transform was wrong in 3 ways:
1. It created an extra instruction when the source and dest types don't match.
2. It did not account for an extra use of the icmp, so could create 2 extra insts.
3. It favored bit hacks over icmp (icmp generally has better analysis).
This fixes#54692 (modeled by the PhaseOrdering tests).
This is a minimal step to fix the bug, but we should likely invert
this and the sibling transform for the "is negative" pattern too.
The backend should be able to invert this back to a shift if that
leads to better codegen.
This is a reduced try of 3794cc0e9964 - that was reverted because
it could cause infinite loops by conflicting with the related
transforms in this block that create shifts.
This reverts commit 3794cc0e996481e10307b67c8436aa44e0d65d22.
This change is suspected of causing bots to hang at stage 2
compiles, so reverting to confirm and investigate.
The existing transform was wrong in 3 ways:
1. It created an extra instruction when the source and dest types don't match.
2. It did not account for an extra use of the icmp, so could create 2 extra insts.
3. It favored bit hacks over icmp (icmp generally has better analysis).
This fixes#54692 (modeled by the PhaseOrdering tests).
This is a minimal step to fix the bug, but we should likely invert
the sibling transform for the "is negative" pattern too.
The backend should be able to invert this back to a shift if that
leads to better codegen.
The tests (see C++ source in #54692) have multiple potential
optimizations/canonicalizations, but we should be consistent
since they are logically identical.
This patch adds initial support for a pointer diff based runtime check
scheme for vectorization. This scheme requires fewer computations and
checks than the existing full overlap checking, if it is applicable.
The main idea is to only check if source and sink of a dependency are
far enough apart so the accesses won't overlap in the vector loop. To do
so, it is sufficient to compute the difference and compare it to the
`VF * UF * AccessSize`. It is sufficient to check
`(Sink - Src) <u VF * UF * AccessSize` to rule out a backwards
dependence in the vector loop with the given VF and UF. If Src >=u Sink,
there is not dependence preventing vectorization, hence the overflow
should not matter and using the ULT should be sufficient.
Note that the initial version is restricted in multiple ways:
1. Pointers must only either be read or written, by a single
instruction (this allows re-constructing source/sink for
dependences with the available information)
2. Source and sink pointers must be add-recs, with matching steps
3. The step must be a constant.
3. abs(step) == AccessSize.
Most of those restrictions can be relaxed in the future.
See https://github.com/llvm/llvm-project/issues/53590.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D119078
Currently SLP vectorizer walks through the instructions and selects
3 main classes of values: 1) reduction operations - instructions with same
reduction opcode (add, mul, min/max, etc.), which build the reduction,
2) reduced values - instructions with the same opcodes, but different
from the reduction opcode, 3) extra arguments - all other values,
instructions from the different basic block rather than the root node,
instructions with to many/less uses.
This scheme is not very efficient. It excludes some instructions and all
non-instruction values from the reductions (constants, proficient
gathers), to many possibly reduced values are marked as extra arguments.
Patch improves this process by introducing a bit extended analysis
stage. During this stage, we still try to select 3 classes of the
values: 1) reduction operations - same as before, 2) possibly reduced
values - all instructions from the current block/non-instructions, which
may build a vectorization tree, 3) extra arguments - instructions from
the different basic blocks. Additionally, an extra sorting of the
possibly reduced values occurs to build the scalar sequences which
highly likely will bed vectorized, e.g. loads are grouped by the
distance between them, constants are grouped together, cmp instructions
are sorted by their compare types and predicates, extractelement
instructions are sorted by the vector operand, etc. Also, these groups
are reordered by their length so the longest group is the first in the
list of the possibly reduced values.
The vectorization process tries to emit the reductions for all these
groups. These reductions, remaining non-vectorized possible reduced
values and extra arguments are then combined into the final expression
just like it was before.
Differential Revision: https://reviews.llvm.org/D114171
Given a shuffle with 4 elements size 16 or 32, we can use the costs
directly from the PerfectShuffle tables to get a slightly more accurate
cost for the resulting shuffle.
Differential Revision: https://reviews.llvm.org/D123409
Currently SLP vectorizer walks through the instructions and selects
3 main classes of values: 1) reduction operations - instructions with same
reduction opcode (add, mul, min/max, etc.), which build the reduction,
2) reduced values - instructions with the same opcodes, but different
from the reduction opcode, 3) extra arguments - all other values,
instructions from the different basic block rather than the root node,
instructions with to many/less uses.
This scheme is not very efficient. It excludes some instructions and all
non-instruction values from the reductions (constants, proficient
gathers), to many possibly reduced values are marked as extra arguments.
Patch improves this process by introducing a bit extended analysis
stage. During this stage, we still try to select 3 classes of the
values: 1) reduction operations - same as before, 2) possibly reduced
values - all instructions from the current block/non-instructions, which
may build a vectorization tree, 3) extra arguments - instructions from
the different basic blocks. Additionally, an extra sorting of the
possibly reduced values occurs to build the scalar sequences which
highly likely will bed vectorized, e.g. loads are grouped by the
distance between them, constants are grouped together, cmp instructions
are sorted by their compare types and predicates, extractelement
instructions are sorted by the vector operand, etc. Also, these groups
are reordered by their length so the longest group is the first in the
list of the possibly reduced values.
The vectorization process tries to emit the reductions for all these
groups. These reductions, remaining non-vectorized possible reduced
values and extra arguments are then combined into the final expression
just like it was before.
Differential Revision: https://reviews.llvm.org/D114171
RequireAnalysis<GlobalsAA> doesn't actually recompute GlobalsAA.
GlobalsAA isn't invalidated (unless specifically invalidated) because
it's self-updating via ValueHandles, but can be imprecise during the
self-updates.
Rather than invalidating GlobalsAA, which would invalidate AAManager and
any analyses that use AAManager, create a new pass that recomputes
GlobalsAA.
Fixes#53131.
Differential Revision: https://reviews.llvm.org/D121167
This is a revert of cfcc42bdc. The analysis is wrong as shown by
the minimal tests for instcombine:
https://alive2.llvm.org/ce/z/y9Dp8A
There may be a way to salvage some of the other tests,
but that can be done as follow-ups. This avoids a miscompile
and fixes#54311.
