If we know that the scalar epilogue is required to run, modify the CFG to end the middle block with an unconditional branch to scalar preheader. This is instead of a conditional branch to either the preheader or the exit block.
The motivation to do this is to support multiple exit blocks. Specifically, the current structure forces us to identify immediate dominators and *which* exit block to branch from in the middle terminator. For the multiple exit case - where we know require scalar will hold - these questions are ill formed.
This is the last change needed to support multiple exit loops, but since the diffs are already large enough, I'm going to land this, and then enable separately. You can think of this as being NFCI-ish prep work, but the changes are a bit too involved for me to feel comfortable tagging the change that way.
Differential Revision: https://reviews.llvm.org/D94892
These attributes were all incorrect or inappropriate for LLVM to infer:
- inaccessiblememonly is generally wrong; user replacement operator new
can access memory that's visible to the caller, as can a new_handler
function.
- willreturn is generally wrong; a custom new_handler is not guaranteed
to terminate.
- noalias is inappropriate: Clang has a flag to determine whether this
attribute should be present and adds it itself when appropriate.
- noundef and nonnull on the return value should be specified by the
frontend on all 'operator new' functions if we want them, not here.
In any case, inferring attributes on functions declared 'nobuiltin' (as
these are when Clang emits them) seems questionable.
Several of the new attributes here were incorrect, and even the ones
that are generally correct were being added even to nobuiltin calls.
This reverts commit bb3f169b59.
This patch extends the condition collection logic to allow adding
conditions from pre-headers to loop headers, by allowing cases where the
target block dominates some of its predecessors.
This patch detaches SampleProfileLoader from class
SampleCoverageTracker. We plan to move SampleProfileLoader
to a template class. This would remain SampleCoverageTracker
as a class.
Also make callsiteIsHot() as a file static function.
Differential Revision: https://reviews.llvm.org/D95823
This patch updates the induction value creation to use VPValues of
recipes to map the created values. This should bring is one step closer
to being able to optimize induction recipes directly in VPlan.
Currently widenIntOrFpInduction also generates vector values for a cast
of the induction, if it exists. Make this explicit by adding the cast
instruction to the values defined by the recipe.
Reviewed By: gilr
Differential Revision: https://reviews.llvm.org/D92284
This patch adds constructors to VPIteration as a cleaner way of
initialising the struct and replaces existing constructions of
the form:
{Part, Lane}
with
VPIteration(Part, Lane)
I have also added a default constructor, which is used by VPlan.cpp
when deciding whether to replicate a block or not.
This refactoring will be required in a later patch that adds more
members and functions to VPIteration.
Differential Revision: https://reviews.llvm.org/D95676
C identifier name input sections such as __llvm_prf_* are GC roots so
they cannot be discarded. In LLD, the SHF_LINK_ORDER flag overrides the
C identifier name semantics.
The !associated metadata may be attached to a global object declaration
with a single argument that references another global object, and it
gets lowered to SHF_LINK_ORDER flag. When a function symbol is discarded
by the linker, setting up !associated metadata allows linker to discard
counters, data and values associated with that function symbol.
Note that !associated metadata is only supported by ELF, it does not have
any effect on non-ELF targets.
Differential Revision: https://reviews.llvm.org/D76802
Sample re-annotation is required in LTO time to achieve a reasonable post-inline profile quality. However, we have seen that such LTO-time re-annotation degrades profile quality. This is mainly caused by preLTO code duplication that is done by passes such as loop unrolling, jump threading, indirect call promotion etc, where samples corresponding to a source location are aggregated multiple times due to the duplicates. In this change we are introducing a concept of distribution factor for pseudo probes so that samples can be distributed for duplicated probes scaled by a factor. We hope that optimizations duplicating code well-maintain the branch frequency information (BFI) based on which probe distribution factors are calculated. Distribution factors are updated at the end of preLTO pipeline to reflect an estimated portion of the real execution count.
This change also introduces a pseudo probe verifier that can be run after each IR passes to detect duplicated pseudo probes.
A saturated distribution factor stands for 1.0. A pesudo probe will carry a factor with the value ranged from 0.0 to 1.0. A 64-bit integral distribution factor field that represents [0.0, 1.0] is associated to each block probe. Unfortunately this cannot be done for callsite probes due to the size limitation of a 32-bit Dwarf discriminator. A 7-bit distribution factor is used instead.
Changes are also needed to the sample profile inliner to deal with prorated callsite counts. Call sites duplicated by PreLTO passes, when later on inlined in LTO time, should have the callees’s probe prorated based on the Prelink-computed distribution factors. The distribution factors should also be taken into account when computing hotness for inline candidates. Also, Indirect call promotion results in multiple callisites. The original samples should be distributed across them. This is fixed by adjusting the callisites' distribution factors.
Reviewed By: wmi
Differential Revision: https://reviews.llvm.org/D93264
Inlining sometimes maps different instructions to be inlined onto the same instruction.
We must ensure to only remap the noalias scopes once. Otherwise the scope might disappear (at best).
This patch ensures that we only replace scopes for which the mapping is known.
This approach is preferred over tracking which instructions we already handled in a SmallPtrSet,
as that one will need more memory.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D95862
Refactoring SampleProfileLoader::inlineHotFunctions to use helpers from CSSPGO inlining and reduce similar code in the inlining loop, plus minor cleanup for AFDO path.
This is resubmit of D95024, with build break and overtighten assertion fixed.
Test Plan:
This is a yet another hint that we will eventually need InstCombineInverter,
which would consistently sink inversions, but but for that we'll need
to consistently hoist inversions where possible, so let's do that here.
Example of a proof: https://alive2.llvm.org/ce/z/78SbDq
See https://bugs.llvm.org/show_bug.cgi?id=48995
This patch updates IRBuilder::CreateMaskedGather/Scatter to work
with ScalableVectorType and adds isLegalMaskedGather/Scatter functions
to AArch64TargetTransformInfo. In addition I've fixed up
isLegalMaskedLoad/Store to return true for supported scalar types,
since this is what the vectorizer asks for.
In LoopVectorize.cpp I've changed
LoopVectorizationCostModel::getInterleaveGroupCost to return an invalid
cost for scalable vectors, since currently this relies upon using shuffle
vector for reversing vectors. In addition, in
LoopVectorizationCostModel::setCostBasedWideningDecision I have assumed
that the cost of scalarising memory ops is infinitely expensive.
I have added some simple masked load/store and gather/scatter tests,
including cases where we use gathers and scatters for conditional invariant
loads and stores.
Differential Revision: https://reviews.llvm.org/D95350
Refactoring SampleProfileLoader::inlineHotFunctions to use helpers from CSSPGO inlining and reduce similar code in the inlining loop, plus minor cleanup for AFDO path.
Test Plan:
Differential Revision: https://reviews.llvm.org/D95024
This change implemented call site prioritized BFS profile guided inlining for sample profile loader. The new inlining strategy maximize the benefit of context-sensitive profile as mentioned in the follow up discussion of CSSPGO RFC. The change will not affect today's AutoFDO as it's opt-in. CSSPGO now defaults to the new FDO inliner, but can fall back to today's replay inliner using a switch (`-sample-profile-prioritized-inline=0`).
Motivation
With baseline AutoFDO, the inliner in sample profile loader only replays previous inlining, and the use of profile is only for pruning previous inlining that turned out to be cold. Due to the nature of replay, the FDO inliner is simple with hotness being the only decision factor. It has the following limitations that we're improving now for CSSPGO.
- It doesn't take inline candidate size into account. Since it's doing replay, the size growth is bounded by previous CGSCC inlining. With context-sensitive profile, FDO inliner is no longer limited by previous inlining, so we need to take size into account to avoid significant size bloat.
- The way it looks at hotness is not accurate. It uses total samples in an inlinee as proxy for hotness, while what really matters for an inline decision is the call site count. This is an unfortunate fall back because call site count and callee entry count are not reliable due to dwarf based correlation, especially for inlinees. Now paired with pseudo-probe, we have accurate call site count and callee's entry count, so we can use that to gauge hotness more accurately.
- It treats all call sites from a block as hot as long as there's one call site considered hot. This is normally true, but since total samples is used as hotness proxy, this transitiveness within block magnifies the inacurate hotness heuristic. With pseduo-probe and the change above, this is no longer an issue for CSSPGO.
New FDO Inliner
Putting all the requirement for CSSPGO together, we need a top-down call site prioritized BFS inliner. Here're reasons why each component is needed.
- Top-down: We need a top-down inliner to better leverage context-sensitive profile, so inlining is driven by accurate context profile, and post-inline is also accurate. This is already implemented in https://reviews.llvm.org/D70655.
- Size Cap: For top-down inliner, taking function size into account for inline decision alone isn't sufficient to control size growth. We also need to explicitly cap size growth because with top-down inlining, we can grow inliner size significantly with large number of smaller inlinees even if each individually passes the cost/size check.
- Prioritize call sites: With size cap, inlining order also becomes important, because if we stop inlining due to size budget limit, we'd want to use budget towards the most beneficial call sites.
- BFS inline: Same as call site prioritization, if we stop inlining due to size budget limit, we want a balanced inline tree, rather than going deep on one call path.
Note that the new inliner avoids repeatedly evaluating same set of call site, so it should help with compile time too. For this reason, we could transition today's FDO inliner to use a queue with equal priority to avoid wasted reevaluation of same call site (TODO).
Speculative indirect call promotion and inlining is also supported now with CSSPGO just like baseline AutoFDO.
Tunings and knobs
I created tuning knobs for size growth/cap control, and for hot threshold separate from CGSCC inliner. The default values are selected based on initial tuning with CSSPGO.
Results
Evaluated with an internal LLVM fork couple months ago, plus another change to adjust hot-threshold cutoff for context profile (will send up after this one), the new inliner show ~1% geomean perf win on spec2006 with CSSPGO, while reducing code size too. The measurement was done using train-train setup, MonoLTO w/ new pass manager and pseudo-probe. Note that this is just a starting point - we hope that the new inliner will open up more opportunity with CSSPGO, but it will certainly take more time and effort to make it fully calibrated and ready for bigger workloads (we're working on it).
Differential Revision: https://reviews.llvm.org/D94001
Extend applyLoopGuards() to take into account conditions/assumes proving some
value %v to be divisible by D by rewriting %v to (%v / D) * D. This lets the
loop unroller and the loop vectorizer identify more loops as not requiring
remainder loops.
Differential Revision: https://reviews.llvm.org/D95521
C identifier name input sections such as __llvm_prf_* are GC roots so
they cannot be discarded. In LLD, the SHF_LINK_ORDER flag overrides the
C identifier name semantics.
The !associated metadata may be attached to a global object declaration
with a single argument that references another global object, and it
gets lowered to SHF_LINK_ORDER flag. When a function symbol is discarded
by the linker, setting up !associated metadata allows linker to discard
counters, data and values associated with that function symbol.
Note that !associated metadata is only supported by ELF, it does not have
any effect on non-ELF targets.
Differential Revision: https://reviews.llvm.org/D76802
Fixing up a couple places where `getCallSiteIdentifier` is needed to support pseudo-probe-based callsites.
Also fixing an issue in the extbinary profile reader where the metadata section is not fully scanned based on the number of profiles loaded only for the current module.
Reviewed By: wmi, wenlei
Differential Revision: https://reviews.llvm.org/D95791
This is another step (see D95452) towards correcting fast-math-flags
bugs in vector reductions.
There are multiple bugs visible in the test diffs, and this is still
not working as it should. We still use function attributes (rather
than FMF) to drive part of the logic, but we are not checking for
the correct FP function attributes.
Note that FMF may not be propagated optimally on selects (example
in https://llvm.org/PR35607 ). That's why I'm proposing to union the
FMF of a fcmp+select pair and avoid regressions on existing vectorizer
tests.
Differential Revision: https://reviews.llvm.org/D95690
A == B map to A >= B && A <= B
(https://alive2.llvm.org/ce/z/_dwxKn).
This extends the constraint construction to return a list of
constraints, which can be used to properly de-compose nested AND & OR.
If the incoming block to a phi node is an EH pad, then we will
materialize into an EH pad, which is not supposed to happen. To fix
this, I added a check to see if incoming block of a phi node is an EH
pad before using it as the insertion point.
Differential Revision: https://reviews.llvm.org/D95019
The constant trunc/ext may not be the optimal pre-condition,
but I think that handles the common cases.
Example of Alive2 proof:
https://alive2.llvm.org/ce/z/sREeLC
This is another step towards canonicalizing to the intrinsics.
Narrowing was identified as source of potential regression for
abs(), so we need to handle this for min/max - see:
https://llvm.org/PR48816
If this is not enough, we could process intrinsics in
the trunc-driven matching in canEvaluateTruncated().
Instead of using ConstraintSystem::negate when adding new constraints,
flip the condition in IR.
The main advantage is that EQ predicates can be represented by 2
constraints, which makes negating based on the constraint tricky. The IR
condition can easily negated.
D90687 introduced a crash:
llvm::LoopVectorizationCostModel::computeMaxVF(llvm::ElementCount, unsigned int):
Assertion `WideningDecisions.empty() && Uniforms.empty() && Scalars.empty() &&
"No decisions should have been taken at this point"' failed.
when compiling the following C code:
typedef struct {
char a;
} b;
b *c;
int d, e;
int f() {
int g = 0;
for (; d; d++) {
e = 0;
for (; e < c[d].a; e++)
g++;
}
return g;
}
with:
clang -Os -target hexagon -mhvx -fvectorize -mv67 testcase.c -S -o -
This occurred since prior to D90687 computeFeasibleMaxVF would only be
called in computeMaxVF when a scalar epilogue was allowed, but now it's
always called. This causes the assert above since computeFeasibleMaxVF
collects all viable VFs larger than the default MaxVF, and for each VF
calculates the register usage which results in analysis being done the
assert above guards against. This can occur in computeFeasibleMaxVF if
TTI.shouldMaximizeVectorBandwidth and this target hook is implemented in
the hexagon backend to always return true.
Reported by @iajbar.
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D94869
If we determine that the invariant path through the loop has no effects,
we can directly branch to the exit block, instead to unswitching first.
Besides avoiding some extra work (unswitching first, then deleting the
loop again) this allows to be more aggressive than regular unswitching
with respect to cost-modeling. This approach should always be be
desirable.
This is similar in spirit to D93734, just that it uses the previously
added checks for loop-unswitching.
I tried to add the required no-op checks from scratch, as we only check
a subset of the loop. There is potential to unify the checks with
LoopDeletion, at the cost of adding a predicate whether a block should
be considered.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D95468
The reduction of a sanitizer build failure when enabling the dominance check (D95335) showed that loop peeling also needs to take care of scope duplication, just like loop unrolling (D92887).
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D95544
This patch fixes updating MemorySSA if the header contains memory
defs that do not clobber a duplicated instruction. We need to find the
first defining access outside the loop body and use that as defining
access of the duplicated instruction.
This fixes a crash caused by bee486851c.
Kazu Hirata