Summary: Added 6 new target hooks for the vectorizer in order to filter types, handle size constraints and decide how to split chains.
Reviewers: tstellarAMD, arsenm
Subscribers: arsenm, mzolotukhin, wdng, llvm-commits, nhaehnle
Differential Revision: https://reviews.llvm.org/D24727
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@283099 91177308-0d34-0410-b5e6-96231b3b80d8
This was first landed in rL283058 and subsequenlty reverted since a
change this depends on (rL283057) was buggy and had to be reverted.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@283079 91177308-0d34-0410-b5e6-96231b3b80d8
They've broken the sanitizer-bootstrap bots. Reverting while I investigate.
Original commit messages:
r283057: "[ConstantRange] Make getEquivalentICmp smarter"
r283058: "[SCEV] Rely on ConstantRange instead of custom logic; NFCI"
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@283062 91177308-0d34-0410-b5e6-96231b3b80d8
(Recommit after making sure IsVerbose gets properly initialized in
DiagnosticInfoOptimizationBase. See previous commit that takes care of
this.)
OptimizationRemarkAnalysis directly takes the role of the report that is
generated by LAA.
Then we need the magic to be able to turn an LAA remark into an LV
remark. This is done via a new OptimizationRemark ctor.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282813 91177308-0d34-0410-b5e6-96231b3b80d8
OptimizationRemarkAnalysis directly takes the role of the report that is
generated by LAA.
Then we need the magic to be able to turn an LAA remark into an LV
remark. This is done via a new OptimizationRemark ctor.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282758 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
When using llc with -compile-twice, module is generated twice, but getAnalysis<ProfileSummaryInfoWrapperPass>().getPSI will still get the old PSI with the original (invalidated) Module. This patch checks if the module has changed when calling getPSI, if yes, update the module and invalidate the Summary.
The bug does not show up in the current llc because PSI is not used in CodeGen yet. But with https://reviews.llvm.org/D24989, the bug will be exposed by test/CodeGen/PowerPC/pr26378.ll
Reviewers: eraman, davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D24993
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282616 91177308-0d34-0410-b5e6-96231b3b80d8
Pointers in different addrspaces can have different sizes, so it's not valid to look through addrspace cast calculating base and offset for a value.
This is similar to D13008.
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D24729
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282612 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
Instead of creating and destroying SCEVUnionPredicate instances (which
internally creates and destroys a DenseMap), use temporary SmallPtrSet
instances of remember the set of predicates that will get reified into a
SCEVUnionPredicate.
Reviewers: silviu.baranga, sbaranga
Subscribers: sanjoy, mcrosier, llvm-commits, mzolotukhin
Differential Revision: https://reviews.llvm.org/D25000
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282606 91177308-0d34-0410-b5e6-96231b3b80d8
Ever since LAA was split out into an analysis on its own, this function
stopped emitting the report directly. Instead it stores it to be
retrieved by the client which can then emit it as its own report
(e.g. -Rpass-analysis=loop-vectorize).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282561 91177308-0d34-0410-b5e6-96231b3b80d8
(Re-committed after moving the template specialization under the yaml
namespace. GCC was complaining about this.)
This allows various presentation of this data using an external tool.
This was first recommended here[1].
As an example, consider this module:
1 int foo();
2 int bar();
3
4 int baz() {
5 return foo() + bar();
6 }
The inliner generates these missed-optimization remarks today (the
hotness information is pulled from PGO):
remark: /tmp/s.c:5:10: foo will not be inlined into baz (hotness: 30)
remark: /tmp/s.c:5:18: bar will not be inlined into baz (hotness: 30)
Now with -pass-remarks-output=<yaml-file>, we generate this YAML file:
--- !Missed
Pass: inline
Name: NotInlined
DebugLoc: { File: /tmp/s.c, Line: 5, Column: 10 }
Function: baz
Hotness: 30
Args:
- Callee: foo
- String: will not be inlined into
- Caller: baz
...
--- !Missed
Pass: inline
Name: NotInlined
DebugLoc: { File: /tmp/s.c, Line: 5, Column: 18 }
Function: baz
Hotness: 30
Args:
- Callee: bar
- String: will not be inlined into
- Caller: baz
...
This is a summary of the high-level decisions:
* There is a new streaming interface to emit optimization remarks.
E.g. for the inliner remark above:
ORE.emit(DiagnosticInfoOptimizationRemarkMissed(
DEBUG_TYPE, "NotInlined", &I)
<< NV("Callee", Callee) << " will not be inlined into "
<< NV("Caller", CS.getCaller()) << setIsVerbose());
NV stands for named value and allows the YAML client to process a remark
using its name (NotInlined) and the named arguments (Callee and Caller)
without parsing the text of the message.
Subsequent patches will update ORE users to use the new streaming API.
* I am using YAML I/O for writing the YAML file. YAML I/O requires you
to specify reading and writing at once but reading is highly non-trivial
for some of the more complex LLVM types. Since it's not clear that we
(ever) want to use LLVM to parse this YAML file, the code supports and
asserts that we're writing only.
On the other hand, I did experiment that the class hierarchy starting at
DiagnosticInfoOptimizationBase can be mapped back from YAML generated
here (see D24479).
* The YAML stream is stored in the LLVM context.
* In the example, we can probably further specify the IR value used,
i.e. print "Function" rather than "Value".
* As before hotness is computed in the analysis pass instead of
DiganosticInfo. This avoids the layering problem since BFI is in
Analysis while DiagnosticInfo is in IR.
[1] https://reviews.llvm.org/D19678#419445
Differential Revision: https://reviews.llvm.org/D24587
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282539 91177308-0d34-0410-b5e6-96231b3b80d8
I don't expect `PendingLoopPredicates` to have very many
elements (e.g. when -O3'ing the sqlite3 amalgamation,
`PendingLoopPredicates` has at most 3 elements). So now we use a
`SmallPtrSet` for it instead of the more heavyweight `DenseSet`.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282511 91177308-0d34-0410-b5e6-96231b3b80d8
This allows various presentation of this data using an external tool.
This was first recommended here[1].
As an example, consider this module:
1 int foo();
2 int bar();
3
4 int baz() {
5 return foo() + bar();
6 }
The inliner generates these missed-optimization remarks today (the
hotness information is pulled from PGO):
remark: /tmp/s.c:5:10: foo will not be inlined into baz (hotness: 30)
remark: /tmp/s.c:5:18: bar will not be inlined into baz (hotness: 30)
Now with -pass-remarks-output=<yaml-file>, we generate this YAML file:
--- !Missed
Pass: inline
Name: NotInlined
DebugLoc: { File: /tmp/s.c, Line: 5, Column: 10 }
Function: baz
Hotness: 30
Args:
- Callee: foo
- String: will not be inlined into
- Caller: baz
...
--- !Missed
Pass: inline
Name: NotInlined
DebugLoc: { File: /tmp/s.c, Line: 5, Column: 18 }
Function: baz
Hotness: 30
Args:
- Callee: bar
- String: will not be inlined into
- Caller: baz
...
This is a summary of the high-level decisions:
* There is a new streaming interface to emit optimization remarks.
E.g. for the inliner remark above:
ORE.emit(DiagnosticInfoOptimizationRemarkMissed(
DEBUG_TYPE, "NotInlined", &I)
<< NV("Callee", Callee) << " will not be inlined into "
<< NV("Caller", CS.getCaller()) << setIsVerbose());
NV stands for named value and allows the YAML client to process a remark
using its name (NotInlined) and the named arguments (Callee and Caller)
without parsing the text of the message.
Subsequent patches will update ORE users to use the new streaming API.
* I am using YAML I/O for writing the YAML file. YAML I/O requires you
to specify reading and writing at once but reading is highly non-trivial
for some of the more complex LLVM types. Since it's not clear that we
(ever) want to use LLVM to parse this YAML file, the code supports and
asserts that we're writing only.
On the other hand, I did experiment that the class hierarchy starting at
DiagnosticInfoOptimizationBase can be mapped back from YAML generated
here (see D24479).
* The YAML stream is stored in the LLVM context.
* In the example, we can probably further specify the IR value used,
i.e. print "Function" rather than "Value".
* As before hotness is computed in the analysis pass instead of
DiganosticInfo. This avoids the layering problem since BFI is in
Analysis while DiagnosticInfo is in IR.
[1] https://reviews.llvm.org/D19678#419445
Differential Revision: https://reviews.llvm.org/D24587
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282499 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
This patch improves thinlto importer
by importing 3x larger functions that are called from hot block.
I compared performance with the trunk on spec, and there
were about 2% on povray and 3.33% on milc. These results seems
to be consistant and match the results Teresa got with her simple
heuristic. Some benchmarks got slower but I think they are just
noisy (mcf, xalancbmki, omnetpp)- running the benchmarks again with
more iterations to confirm. Geomean of all benchmarks including the noisy ones
were about +0.02%.
I see much better improvement on google branch with Easwaran patch
for pgo callsite inlining (the inliner actually inline those big functions)
Over all I see +0.5% improvement, and I get +8.65% on povray.
So I guess we will see much bigger change when Easwaran patch will land
(it depends on new pass manager), but it is still worth putting this to trunk
before it.
Implementation details changes:
- Removed CallsiteCount.
- ProfileCount got replaced by Hotness
- hot-import-multiplier is set to 3.0 for now,
didn't have time to tune it up, but I see that we get most of the interesting
functions with 3, so there is no much performance difference with higher, and
binary size doesn't grow as much as with 10.0.
Reviewers: eraman, mehdi_amini, tejohnson
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D24638
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282437 91177308-0d34-0410-b5e6-96231b3b80d8
Noticed due to the warning on this line. Sanjoy is on
a less-than-awesome internet connection, so committing on his behalf.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282380 91177308-0d34-0410-b5e6-96231b3b80d8
In a previous change I collapsed two different caches into one. When
doing that I noticed that ScalarEvolution's move constructor was not
moving those caches.
To keep the previous change simple, I've moved that bugfix into this
separate change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282376 91177308-0d34-0410-b5e6-96231b3b80d8
Both `loopHasNoSideEffects` and `loopHasNoAbnormalExits` involve walking
the loop and maintaining similar sorts of caches. This commit changes
SCEV to compute both the predicates via a single walk, and maintain a
single cache instead of two.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282375 91177308-0d34-0410-b5e6-96231b3b80d8
Specifically, it moves SCEVUnionPredicates from its input into its own
storage. Make this obvious at the type level.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282374 91177308-0d34-0410-b5e6-96231b3b80d8
SCEVUnionPredicate is a "heavyweight" structure, so it is beneficial to
store the (optional) data out of line.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282366 91177308-0d34-0410-b5e6-96231b3b80d8
This change simplifies a data structure optimization in the
`BackedgeTakenInfo` class for loops with exactly one computable exit.
I've sanity checked that this does not regress compile time performance,
using sqlite3's amalgamated build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282365 91177308-0d34-0410-b5e6-96231b3b80d8
There is no benefit in looking through assumptions on UndefValue to
guess known bits. Return early to avoid walking their use-lists, and
assert that all instances of ConstantData are handled here for similar
reasons (UndefValue was the only integer/pointer holdout).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282337 91177308-0d34-0410-b5e6-96231b3b80d8
Check and return early for ConstantPointerNull and UndefValue
specifically in isKnownNonNullAt, and assert that ConstantData never
make it to isKnownNonNullFromDominatingCondition.
This confirms that isKnownNonNullFromDominatingCondition never walks
through the use-list of an instance of ConstantData. Given that such
use-lists cross module boundaries, it never really made sense to do so,
and was potentially very expensive.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282333 91177308-0d34-0410-b5e6-96231b3b80d8
Summary: When identifying cold blocks, consider only the edge to the normal destination if the terminator is InvokeInst and let calcInvokeHeuristics() decide edge weights for the InvokeInst.
Reviewers: mcrosier, hfinkel, davidxl
Subscribers: mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D24868
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282262 91177308-0d34-0410-b5e6-96231b3b80d8
computeKnownBits() already works for integer vectors, so allow vector types when calling that from InstCombine.
I don't think the change to use m_APInt in computeKnownBits is strictly necessary because we do check for
ConstantVector later, but it's more efficient to handle the splat case without needing to loop on vector elements.
This should work with InstSimplify, but doesn't yet, so I made that a FIXME comment on the test for PR24942:
https://llvm.org/bugs/show_bug.cgi?id=24942
Differential Revision: https://reviews.llvm.org/D24677
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@281777 91177308-0d34-0410-b5e6-96231b3b80d8
LazyCallGraph to support repeated, stable iterations, even in the face
of graph updates.
This is particularly important to allow the CGSCC pass manager to walk
the RefSCCs (and thus everything else) in a module more than once. Lots
of unittests and other tests were hard or impossible to write because
repeated CGSCC pass managers which didn't invalidate the LazyCallGraph
would conclude the module was empty after the first one. =[ Really,
really bad.
The interesting thing is that in many ways this simplifies the code. We
can now re-use the same code for handling reference edge insertion
updates of the RefSCC graph as we use for handling call edge insertion
updates of the SCC graph. Outside of adapting to the shared logic for
this (which isn't trivial, but is *much* simpler than the DFS it
replaces!), the new code involves putting newly created RefSCCs when
deleting a reference edge into the cached list in the correct way, and
to re-formulate the iterator to be stable and effective even in the face
of these kinds of updates.
I've updated the unittests for the LazyCallGraph to re-iterate the
postorder sequence and verify that this all works. We even check for
using alternating iterators to trigger the lazy formation of RefSCCs
after mutation has occured.
It's worth noting that there are a reasonable number of likely
simplifications we can make past this. It isn't clear that we need to
keep the "LeafRefSCCs" around any more. But I've not removed that mostly
because I want this to be a more isolated change.
Differential Revision: https://reviews.llvm.org/D24219
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@281716 91177308-0d34-0410-b5e6-96231b3b80d8
The patch is to partially fix PR10584. Correlated Value Propagation queries LVI
to check non-null for pointer params of each callsite. If we know the def of
param is an alloca instruction, we know it is non-null and can return early from
LVI. Similarly, CVP queries LVI to check whether pointer for each mem access is
constant. If the def of the pointer is an alloca instruction, we know it is not
a constant pointer. These shortcuts can reduce the cost of CVP significantly.
Differential Revision: https://reviews.llvm.org/D18066
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@281586 91177308-0d34-0410-b5e6-96231b3b80d8
value is a pointer.
This patch is to fix PR30213. When expanding an expr based on ValueOffsetPair,
if the value is of pointer type, we can only create a getelementptr instead
of sub expr.
Differential Revision: https://reviews.llvm.org/D24088
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@281439 91177308-0d34-0410-b5e6-96231b3b80d8
The constant folder didn't know how to always fold bitcasts of constant integer
vectors. In particular, it was unable to handle the case where a constant vector
had some undef elements, and the resulting (i.e. bitcasted) vector type had more
elements than the original vector type.
Example:
%cast = bitcast <2 x i64><i64 undef, i64 2> to <4 x i32>
On a little endian target, %cast could have been folded to:
<4 x i32><i32 undef, i32 undef, i32 2, i32 0>
This patch improves the folding logic by teaching how to correctly propagate
undef elements in the folded vector.
Differential Revision: https://reviews.llvm.org/D24301
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@281343 91177308-0d34-0410-b5e6-96231b3b80d8
Convert the previous introduced is-a relationship between the LVICache and LVIImple clases into a has-a relationship and hide all the implementation details of the cache from the lazy query layer.
The only slightly concerning change here is removing the addition of a queried block into the SeenBlock set in LVIImpl::getBlockValue. As far as I can tell, this was effectively dead code. I think it *used* to be the case that getCachedValueInfo wasn't const and might end up inserting elements in the cache during lookup. That's no longer true and hasn't been for a while. I did fixup the const usage to make that more obvious.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@281272 91177308-0d34-0410-b5e6-96231b3b80d8
The only interesting bit here is the refactor of the handle callback and even that's pretty straight-forward.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@281267 91177308-0d34-0410-b5e6-96231b3b80d8
Seperate the caching logic from the implementation of the lazy analysis. For the moment, the lazy analysis impl has a is-a relationship with the cache; this will change to a has-a relationship shortly. This was done as two steps merely to keep the changes simple and the diff understandable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@281266 91177308-0d34-0410-b5e6-96231b3b80d8
make_scope_exit now that we have that utility.
This makes the code much more clear and readable by isolating the check.
It also makes it easy to go through and make sure all the interesting
update routines have a start and end verify so we don't slowly let the
graph drift into an invalid state.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280619 91177308-0d34-0410-b5e6-96231b3b80d8
a postorder-sequence based update after edge insertion into a generic
helper function.
This separates the SCC-specific logic into two fairly simple lambdas and
extracts the rest into a generic helper template function. I think this
is a net win on its own merits because it disentangles different pieces
of the algorithm. Now there is one place that does the two-step
partition to identify a set of newly connected components and at the
same time update the postorder sequence.
However, I'm also hoping to re-use this an upcoming patch to update
a cached post-order sequence of RefSCCs when doing the analogous update
to the RefSCC graph, and I don't want to have two copies.
The diff is quite messy but this really is just moving things around and
making types generic rather than specific.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280618 91177308-0d34-0410-b5e6-96231b3b80d8
We don't need to call `GetCompareTy(LHS)' every single time true or false is
returned from function SimplifyFCmpInst as suggested by Sanjay in review D24142.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280491 91177308-0d34-0410-b5e6-96231b3b80d8
This patch fixes a crash caused by an incorrect folding of an ordered comparison
between a packed floating point vector and a splat vector of NaN.
An ordered comparison between a vector and a constant vector of NaN, should
always be folded into a constant vector where each element is i1 false.
Since revision 266175, SimplifyFCmpInst folds the ordered fcmp into a scalar
'false'. Later on, this would cause an assertion failure, since the value type
of the folded value doesn't match the expected value type of the uses of the
original instruction: "Assertion failed: New->getType() == getType() &&
"replaceAllUses of value with new value of different type!".
This patch fixes the issue and adds a test case to the already existing test
InstSimplify/floating-point-compares.ll.
Differential Revision: https://reviews.llvm.org/D24143
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280488 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
Current implementation of LI verifier isn't ideal and fails to detect
some cases when LI is incorrect. For instance, it checks that all
recorded loops are in a correct form, but it has no way to check if
there are no more other (unrecorded in LI) loops in the function. This
patch adds a way to detect such bugs.
Reviewers: chandlerc, sanjoy, hfinkel
Subscribers: llvm-commits, silvas, mzolotukhin
Differential Revision: https://reviews.llvm.org/D23437
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280280 91177308-0d34-0410-b5e6-96231b3b80d8
There were paths where we wouldn't populate the visited set, causing us
to recurse forever if an SSA variable was defined in terms of itself.
This fixes PR30210.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280191 91177308-0d34-0410-b5e6-96231b3b80d8
Or they were not instantiated as expected;
llvm::InnerAnalysisManagerProxy<llvm::AnalysisManager<llvm::Function>, llvm::LazyCallGraph::SCC>::PassID
llvm::InnerAnalysisManagerProxy<llvm::AnalysisManager<llvm::Function>, llvm::LazyCallGraph::SCC>::PassID
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280105 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
Changed this code because it was not very readable.
The one question that I got after changing it is, should we
count calls to intrinsics? We don't add them to caller summary,
so maybe we shouldn't also count them?
Reviewers: tejohnson, eraman, mehdi_amini
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D23949
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280036 91177308-0d34-0410-b5e6-96231b3b80d8
Fixed a bug in run-time checks for possible memory conflicts inside loop.
The bug is in Low <-> High boundaries calculation. The High boundary should be calculated as "last memory access pointer + element size".
Differential revision: https://reviews.llvm.org/D23176
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279930 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
This is obviously an interesting case because it may motivate code
restructuring or LTO.
Reporting this requires instantiation of ORE in the loop where the call
sites are first gathered. I've checked compile-time
overhead *with* -Rpass-with-hotness and the worst slow-down was 6% in
mcf and quickly tailing off. As before without -Rpass-with-hotness
there is no overhead.
Because this could be a pretty noisy diagnostics, it is currently
qualified as 'verbose'. As of this patch, 'verbose' diagnostics are
only emitted with -Rpass-with-hotness, i.e. when the output is expected
to be filtered.
Reviewers: eraman, chandlerc, davidxl, hfinkel
Subscribers: tejohnson, Prazek, davide, llvm-commits
Differential Revision: https://reviews.llvm.org/D23415
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279860 91177308-0d34-0410-b5e6-96231b3b80d8
Summary: Dead store elimination gets very expensive when large numbers of instructions need to be analyzed. This patch limits the number of instructions analyzed per store to the value of the memdep-block-scan-limit parameter (which defaults to 100). This resulted in no observed difference in performance of the generated code, and no change in the statistics for the dead store elimination pass, but improved compilation time on some files by more than an order of magnitude.
Reviewers: dexonsmith, bruno, george.burgess.iv, dberlin, reames, davidxl
Subscribers: davide, chandlerc, dberlin, davidxl, eraman, tejohnson, mbodart, llvm-commits
Differential Revision: https://reviews.llvm.org/D15537
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279833 91177308-0d34-0410-b5e6-96231b3b80d8
r279696, which changed `LLVM_CONSTEXPR AliasAttr` to `const AliasAttr`,
made this comment make less sense.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279699 91177308-0d34-0410-b5e6-96231b3b80d8
This patch changes LLVM_CONSTEXPR variable declarations to const
variable declarations, since LLVM_CONSTEXPR expands to nothing if the
current compiler doesn't support constexpr. In all of the changed
cases, it looks like the code intended the variable to be const instead
of sometimes-constexpr sometimes-not.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279696 91177308-0d34-0410-b5e6-96231b3b80d8
manager, including both plumbing and logic to handle function pass
updates.
There are three fundamentally tied changes here:
1) Plumbing *some* mechanism for updating the CGSCC pass manager as the
CG changes while passes are running.
2) Changing the CGSCC pass manager infrastructure to have support for
the underlying graph to mutate mid-pass run.
3) Actually updating the CG after function passes run.
I can separate them if necessary, but I think its really useful to have
them together as the needs of #3 drove #2, and that in turn drove #1.
The plumbing technique is to extend the "run" method signature with
extra arguments. We provide the call graph that intrinsically is
available as it is the basis of the pass manager's IR units, and an
output parameter that records the results of updating the call graph
during an SCC passes's run. Note that "...UpdateResult" isn't a *great*
name here... suggestions very welcome.
I tried a pretty frustrating number of different data structures and such
for the innards of the update result. Every other one failed for one
reason or another. Sometimes I just couldn't keep the layers of
complexity right in my head. The thing that really worked was to just
directly provide access to the underlying structures used to walk the
call graph so that their updates could be informed by the *particular*
nature of the change to the graph.
The technique for how to make the pass management infrastructure cope
with mutating graphs was also something that took a really, really large
number of iterations to get to a place where I was happy. Here are some
of the considerations that drove the design:
- We operate at three levels within the infrastructure: RefSCC, SCC, and
Node. In each case, we are working bottom up and so we want to
continue to iterate on the "lowest" node as the graph changes. Look at
how we iterate over nodes in an SCC running function passes as those
function passes mutate the CG. We continue to iterate on the "lowest"
SCC, which is the one that continues to contain the function just
processed.
- The call graph structure re-uses SCCs (and RefSCCs) during mutation
events for the *highest* entry in the resulting new subgraph, not the
lowest. This means that it is necessary to continually update the
current SCC or RefSCC as it shifts. This is really surprising and
subtle, and took a long time for me to work out. I actually tried
changing the call graph to provide the opposite behavior, and it
breaks *EVERYTHING*. The graph update algorithms are really deeply
tied to this particualr pattern.
- When SCCs or RefSCCs are split apart and refined and we continually
re-pin our processing to the bottom one in the subgraph, we need to
enqueue the newly formed SCCs and RefSCCs for subsequent processing.
Queuing them presents a few challenges:
1) SCCs and RefSCCs use wildly different iteration strategies at
a high level. We end up needing to converge them on worklist
approaches that can be extended in order to be able to handle the
mutations.
2) The order of the enqueuing need to remain bottom-up post-order so
that we don't get surprising order of visitation for things like
the inliner.
3) We need the worklists to have set semantics so we don't duplicate
things endlessly. We don't need a *persistent* set though because
we always keep processing the bottom node!!!! This is super, super
surprising to me and took a long time to convince myself this is
correct, but I'm pretty sure it is... Once we sink down to the
bottom node, we can't re-split out the same node in any way, and
the postorder of the current queue is fixed and unchanging.
4) We need to make sure that the "current" SCC or RefSCC actually gets
enqueued here such that we re-visit it because we continue
processing a *new*, *bottom* SCC/RefSCC.
- We also need the ability to *skip* SCCs and RefSCCs that get merged
into a larger component. We even need the ability to skip *nodes* from
an SCC that are no longer part of that SCC.
This led to the design you see in the patch which uses SetVector-based
worklists. The RefSCC worklist is always empty until an update occurs
and is just used to handle those RefSCCs created by updates as the
others don't even exist yet and are formed on-demand during the
bottom-up walk. The SCC worklist is pre-populated from the RefSCC, and
we push new SCCs onto it and blacklist existing SCCs on it to get the
desired processing.
We then *directly* update these when updating the call graph as I was
never able to find a satisfactory abstraction around the update
strategy.
Finally, we need to compute the updates for function passes. This is
mostly used as an initial customer of all the update mechanisms to drive
their design to at least cover some real set of use cases. There are
a bunch of interesting things that came out of doing this:
- It is really nice to do this a function at a time because that
function is likely hot in the cache. This means we want even the
function pass adaptor to support online updates to the call graph!
- To update the call graph after arbitrary function pass mutations is
quite hard. We have to build a fairly comprehensive set of
data structures and then process them. Fortunately, some of this code
is related to the code for building the cal graph in the first place.
Unfortunately, very little of it makes any sense to share because the
nature of what we're doing is so very different. I've factored out the
one part that made sense at least.
- We need to transfer these updates into the various structures for the
CGSCC pass manager. Once those were more sanely worked out, this
became relatively easier. But some of those needs necessitated changes
to the LazyCallGraph interface to make it significantly easier to
extract the changed SCCs from an update operation.
- We also need to update the CGSCC analysis manager as the shape of the
graph changes. When an SCC is merged away we need to clear analyses
associated with it from the analysis manager which we didn't have
support for in the analysis manager infrsatructure. New SCCs are easy!
But then we have the case that the original SCC has its shape changed
but remains in the call graph. There we need to *invalidate* the
analyses associated with it.
- We also need to invalidate analyses after we *finish* processing an
SCC. But the analyses we need to invalidate here are *only those for
the newly updated SCC*!!! Because we only continue processing the
bottom SCC, if we split SCCs apart the original one gets invalidated
once when its shape changes and is not processed farther so its
analyses will be correct. It is the bottom SCC which continues being
processed and needs to have the "normal" invalidation done based on
the preserved analyses set.
All of this is mostly background and context for the changes here.
Many thanks to all the reviewers who helped here. Especially Sanjoy who
caught several interesting bugs in the graph algorithms, David, Sean,
and others who all helped with feedback.
Differential Revision: http://reviews.llvm.org/D21464
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279618 91177308-0d34-0410-b5e6-96231b3b80d8
And add a FIXME because the helper excludes folds for vectors. It's
not clear yet how many of these are actually testable (and therefore
necessary?) because later analysis uses computeKnownBits and other
methods to catch many of these cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279492 91177308-0d34-0410-b5e6-96231b3b80d8
Repeated inserts into AliasSetTracker have quadratic behavior - inserting a
pointer into AST is linear, since it requires walking over all "may" alias
sets and running an alias check vs. every pointer in the set.
We can avoid this by tracking the total number of pointers in "may" sets,
and when that number exceeds a threshold, declare the tracker "saturated".
This lumps all pointers into a single "may" set that aliases every other
pointer.
(This is a stop-gap solution until we migrate to MemorySSA)
This fixes PR28832.
Differential Revision: https://reviews.llvm.org/D23432
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279274 91177308-0d34-0410-b5e6-96231b3b80d8
directly produce the index as the value type result.
This requires making the index movable which is straightforward. It
greatly simplifies things by allowing us to completely avoid the builder
API and the layers of abstraction inherent there. Instead both pass
managers can directly construct these when run by value. They still
won't be constructed truly eagerly thanks to the optional in the legacy
PM. The code that directly builds the index can also just share a direct
function.
A notable change here is that the result type of the analysis for the
new PM is no longer a reference type. This was really problematic when
making changes to how we handle result types to make our interface
requirements *much* more strict and precise. But I think this is an
overall improvement.
Differential Revision: https://reviews.llvm.org/D23701
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279216 91177308-0d34-0410-b5e6-96231b3b80d8
number of assume intrinsics.
The classical way to have a cache-friendly vector style container when
we need queue semantics for BFS instead of stack semantics for DFS is to
use an ever-growing vector and an index. Erasing from the front requires
O(size) work, and unless we expect the worklist to grow *very* large,
its probably cheaper to just grow and race down the list.
But that makes it more bad that we're putting the assume intrinsics in
this at all. We end up looking at the (by definition empty) use list to
see if they're ephemeral (when we've already put them in that set), etc.
Instead, directly populate the worklist with the operands when we mark
the assume intrinsics as ephemeral. Also, test the visited set *before*
putting things into the worklist so we don't accumulate the same value
in the list 100s of times.
It would be nice to use a set-vector for this but I think its useful to
test the set earlier to avoid repeatedly querying whether the same
instruction is safe to speculate.
Hopefully with these changes the number of values pushed onto the
worklist is smaller, and we avoid quadratic work by letting it grow as
necessary.
Differential Revision: https://reviews.llvm.org/D23396
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@279099 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
This is part of the "NodeType* -> NodeRef" migration. Notice that since
GraphWriter prints object address as identity, I added a static_assert on
NodeRef to be a pointer type.
Reviewers: dblaikie
Subscribers: llvm-commits, MatzeB
Differential Revision: https://reviews.llvm.org/D23580
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@278966 91177308-0d34-0410-b5e6-96231b3b80d8
Refactored so that a LSRUse owns its fixups, as oppsed to letting the
LSRInstance own them. This makes it easier to rate formulas for
LSRUses, since the fixups are available directly. The Offsets vector
has been removed since it was no longer necessary.
New target hook isFoldableMemAccessOffset(), which is used during formula
rating.
For SystemZ, this is useful to express that loads and stores with
float or vector types with a big/negative offset should be avoided in
loops. Without this, LSR will generate a lot of negative offsets that
would require extra instructions for loading the address.
Updated tests:
test/CodeGen/SystemZ/loop-01.ll
Reviewed by: Quentin Colombet and Ulrich Weigand.
https://reviews.llvm.org/D19152
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@278927 91177308-0d34-0410-b5e6-96231b3b80d8
This is a mechanical change of comments in switches like fallthrough,
fall-through, or fall-thru to use the LLVM_FALLTHROUGH macro instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@278902 91177308-0d34-0410-b5e6-96231b3b80d8
When there's only one argument and it doesn't match one of the known
functions, return ARCInstKind::CallOrUser rather than falling through
to the two argument case. The old behaviour both incremented past and
dereferenced end().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@278881 91177308-0d34-0410-b5e6-96231b3b80d8
