I wrote this patch before seeing the comment in:
https://reviews.llvm.org/D27114
...that suggests we should actually be canonicalizing the other way.
So just in case we decide this is the right way, we might as well
have a cleaner implementation.
llvm-svn: 290912
Summary:
Regardless how the loop body weight is distributed, we should preserve
total loop body weight. i.e. we should have same weight reaching the body of the loop
or its duplicates in peeled and unpeeled case.
Reviewers: mkuper, davidxl, anemet
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28179
llvm-svn: 290833
Apparently my suggestion of using ternary doesn't really work
as clang complains about incompatible types on LHS and RHS. Some
GCC versions happen to accept the code but clang behaviour is
correct here.
llvm-svn: 290822
Summary:
This avoids the very fragile code for null expressions. We could also use a denseset that tracks which things have null expressions instead, but that seems pretty fragile and premature optimization.
This resolves a number of infinite loop cases, test reductions coming.
Reviewers: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28193
llvm-svn: 290816
Summary: Previously, we tried to fix up the equivalences during symbolic evaluation. This does not work. Now, we change the equivalences during congruence finding, where it belongs. We also initialize the equivalence table to give a maximal answer.
Reviewers: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28192
llvm-svn: 290815
CVP doesn't care about the order of blocks visited, but by using a pre-order traversal over the graph we can a) not visit unreachable blocks and b) optimize as we go so that analysis of later blocks produce slightly more precise results.
I noticed this via inspection and don't have a concrete example which points to the issue.
llvm-svn: 290760
This is similar to the allocfn case - if an alloca is not captured, then it's
necessarily thread-local.
Differential Revision: https://reviews.llvm.org/D28170
llvm-svn: 290738
Summary:
The current loop complete unroll algorithm checks if unrolling complete will reduce the runtime by a certain percentage. If yes, it will apply a fixed boosting factor to the threshold (by discounting cost). The problem for this approach is that the threshold abruptly. This patch makes the boosting factor a function of runtime reduction percentage, capped by a fixed threshold. In this way, the threshold changes continuously.
The patch also simplified the code by reducing one parameter in UP.
The patch only affects code-gen of two speccpu2006 benchmark:
445.gobmk binary size decreases 0.08%, no performance change.
464.h264ref binary size increases 0.24%, no performance change.
Reviewers: mzolotukhin, chandlerc
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D26989
llvm-svn: 290737
"Changed" doesn't actually change within the loop, so there's
no reason to keep track of it - we always return false during
analysis and true after the transformation is made.
llvm-svn: 290735
We correctly canonicalized (add (sext x), (sext y)) to (sext (add x, y))
where possible. However, we didn't perform the same canonicalization
for zexts or for muls.
llvm-svn: 290733
This moves the exit block and insertion point computation to be eager,
instead of after seeing the first scalar we can promote.
The cost is relatively small (the computation happens anyway, see discussion
on D28147), and the code is easier to follow, and can bail out earlier
if there's a catchswitch present.
llvm-svn: 290729
We would check whether we have a prehader *or* dedicated exit blocks,
and go into the promotion loop. Then, for each alias set we'd check
if we have a preheader *and* dedicated exit blocks, and bail if not.
Instead, bail immediately if we don't have both.
llvm-svn: 290728
We want to recompute LCSSA only when we actually promoted a value.
This means we only need to look at changes made by promotion when
deciding whether to recompute it or not, not at regular sinking/hoisting.
(This was what the code was documented as doing, just not what it did)
Hopefully NFC.
llvm-svn: 290726
We bypassed the intrinsic and returned the passthru operand, but we should also add the intrinsic to the worklist since its now dead. This can allow DCE to find it sooner and remove it. Similar was done for InsertElement when the inserted element isn't demanded.
llvm-svn: 290704
Summary:
The optimal iteration order for this problem is RPO order. We want to
process as many preds of a backedge as we can before we process the
backedge.
At the same time, as we add predicate handling, we want to be able to
touch instructions that are dominated by a given block by
ranges (because a change in value numbering a predicate possibly
affects all users we dominate that are using that predicate).
If we don't do it this way, we can't do value inference over
backedges (the paper covers this in depth).
The newgvn branch currently overshoots the last part, and guarantees
that it will touch *at least* the right set of instructions, but it
does touch more. This is because the bitvector instruction ranges are
currently generated in RPO order (so we take the max and the min of
the ranges of dominated blocks, which means there are some in the
middle we didn't have to touch that we did).
We can do better by sorting the dominator tree, and then just using
dominator tree order.
As a preliminary, the dominator tree has some RPO guarantees, but not
enough. It guarantees that for a given node, your idom must come
before you in the RPO ordering. It guarantees no relative RPO ordering
for siblings. We add siblings in whatever order they appear in the module.
So that is what we fix.
We sort the children array of the domtree into RPO order, and then use
the dominator tree for ordering, instead of RPO, since the dominator
tree is now a valid RPO ordering.
Note: This would help any other pass that iterates a forward problem
in dominator tree order. Most of them are single pass. It will still
maximize whatever result they compute. We could also build the
dominator tree in this order, but our incremental updates would still
put it out of sort order, and recomputing the sort order is almost as
hard as general incremental updates of the domtree.
Also note that the sorting does not affect any tests, etc. Nothing
depends on domtree order, including the verifier, the equals
functions for domtree nodes, etc.
How much could this matter, you ask?
Here are the current numbers.
This is generated by running NewGVN over all files in LLVM.
Note that once we propagate equalities, the differences go up by an
order of magnitude or two (IE instead of 29, the max ends up in the
thousands, since the worst case we add a factor of N, where N is the
number of branch predicates). So while it doesn't look that stark for
the default ordering, it gets *much much* worse. There are also
programs in the wild where the difference is already pretty stark
(2 iterations vs hundreds).
RPO ordering:
759040 Number of iterations is 1
112908 Number of iterations is 2
Default dominator tree ordering:
755081 Number of iterations is 1
116234 Number of iterations is 2
603 Number of iterations is 3
27 Number of iterations is 4
2 Number of iterations is 5
1 Number of iterations is 7
Dominator tree sorted:
759040 Number of iterations is 1
112908 Number of iterations is 2
<yay!>
Really bad ordering (sort domtree siblings in postorder. not quite the
worst possible, but yeah):
754008 Number of iterations is 1
21 Number of iterations is 10
8 Number of iterations is 11
6 Number of iterations is 12
5 Number of iterations is 13
2 Number of iterations is 14
2 Number of iterations is 15
3 Number of iterations is 16
1 Number of iterations is 17
2 Number of iterations is 18
96642 Number of iterations is 2
1 Number of iterations is 20
2 Number of iterations is 21
1 Number of iterations is 22
1 Number of iterations is 29
17266 Number of iterations is 3
2598 Number of iterations is 4
798 Number of iterations is 5
273 Number of iterations is 6
186 Number of iterations is 7
80 Number of iterations is 8
42 Number of iterations is 9
Reviewers: chandlerc, davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28129
llvm-svn: 290699
emplace_back is not faster if it is equivalent to push_back. In this cases emplaced value had the
same type that the one stored in container. It is ugly and it might be even slower (see
Scott Meyers presentation about emplacement).
llvm-svn: 290685
when they are call edges at the leaf but may (transitively) be reached
via ref edges.
It turns out there is a simple rule: insert everything as a ref edge
which is a safe conservative default. Then we let the existing update
logic handle promoting some of those to call edges.
Note that it would be fairly cheap to make these call edges right away
if that is desirable by testing whether there is some existing call path
from the source to the target. It just seemed like slightly more
complexity in this code path that isn't strictly necessary. If anyone
feels strongly about handling this differently I'm happy to change it.
llvm-svn: 290649
This adds a combine that canonicalizes a chain of inserts which broadcasts
a value into a single insert + a splat shufflevector.
This fixes PR31286.
Differential Revision: https://reviews.llvm.org/D27992
llvm-svn: 290641
skipping indirectly recursive inline chains.
To do this, we implicitly build an inline stack for each callsite and
check prior to inlining that doing so would not form a cycle. This uses
the exact same technique and even shares some code with the legacy PM
inliner.
This solution remains deeply unsatisfying to me because it means we
cannot actually iterate the inliner externally. Doing so would not be
able to easily detect and avoid such cycles. Some day I would very much
like to have a solution that works without this internal state to detect
cycles, but this is not that day.
llvm-svn: 290590
PMULDQ/PMULUDQ vXi64 instructions only use the even numbered v2Xi32 input elements which SimplifyDemandedVectorElts should try and use.
This builds on r290554 which added supported for 128 and 256-bit.
llvm-svn: 290582
An earlier commit added support for unmasked scalar operations. At that time isel wouldn't generate an optimal sequence for masked operations, but that has now been fixed.
llvm-svn: 290566
inside of `InlineFunction`. Prior to this, call instructions are
specifically being rewritten and replaced within the inlined region,
invalidating some of the call sites.
Several of these regions are using the same technique to walk the
inlined region so this seems clearly safe up to this point.
I've also added a short circuit to the scan for call sites based on what
other code is doing.
With this, the most common crash I've found in the new inliner code is
fixed. I've turned it on for another test case that covers this
scenario.
I'll make my way through most of the other inliner test cases
just to get some easy coverage next.
llvm-svn: 290562
removing fully-dead comdats without removing dead entries in comdats
with live members.
This factors the core logic out of the current inliner's internals to
a reusable utility and leverages that in both places. The factored out
code should also be (minorly) more efficient in cases where we have very
few dead functions or dead comdats to consider.
I've added a test case to cover this behavior of the always inliner.
This is the last significant bug in the new PM's always inliner I've
found (so far).
llvm-svn: 290557
PMULDQ/PMULUDQ vXi64 instructions only use the even numbered v2Xi32 input elements which SimplifyDemandedVectorElts should try and use.
Differential Revision: https://reviews.llvm.org/D28119
llvm-svn: 290554
Mostly use a bit more idiomatic C++ where we can,
so we can combine some things later.
Reviewers: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28111
llvm-svn: 290550
Summary:
I only do this for unmasked cases for now because isel is failing to fold the mask. I'll try to fix that soon.
I'll do the same thing for packed add/sub/mul/div in a future patch.
Reviewers: delena, RKSimon, zvi, craig.topper
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D27879
llvm-svn: 290535
Summary:
This patch adds support for converting the masked vpermv intrinsics into shufflevector instructions if the indices are constants.
We also need to wrap a select instruction around the shuffle to take care of the masking part. InstCombine will take care of optimizing the select if the mask is constant so I didn't bother checking for that.
Reviewers: zvi, delena, spatel, RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D27825
llvm-svn: 290530
whether functions are removed, and fix the new PM's always inliner to
actually pass this test.
Without this, the new PM's always inliner leaves all the functions
kicking around which won't work out very well given the semantics of
always inline.
Doing this really highlights how frustrating the current alwaysinline
semantic contract is though -- why can we put it on *external*
functions, etc?
Also I've added a number of tricky and interesting test cases for
removing functions with the always inliner. There is one remaining case
not handled -- fully removing comdats -- and I've left a FIXME about
this.
llvm-svn: 290457
The pass creates some state which expects to be cleaned up by
a later instance of the same pass. opt-bisect happens to expose
this not ideal design because calling skipLoop() will result in
this state not being cleaned up at times and an assertion firing
in `doFinalization()`. Chandler tells me the new pass manager will
give us options to avoid these design traps, but until it's not ready,
we need a workaround for the current pass infrastructure. Fix provided
by Andy Kaylor, see the review for a complete discussion.
Differential Revision: https://reviews.llvm.org/D25848
llvm-svn: 290427
Use a dummy private function with inline asm calls instead of module
level asm blocks for CFI jumptables.
The main advantage is that now jumptable codegen can be affected by
the function attributes (like target_cpu on ARM). Module level asm
gets the default subtarget based on the target triple, which is often
not good enough.
This change also uses asm constraints/arguments to reference
jumptable targets and aliases directly. We no longer do asm name
mangling in an IR pass.
Differential Revision: https://reviews.llvm.org/D28012
llvm-svn: 290384
The code have been developed by Daniel Berlin over the years, and
the new implementation goal is that of addressing shortcomings of
the current GVN infrastructure, i.e. long compile time for large
testcases, lack of phi predication, no load/store value numbering
etc...
The current code just implements the "core" GVN algorithm, although
other pieces (load coercion, phi handling, predicate system) are
already implemented in a branch out of tree. Once the core is stable,
we'll start adding pieces on top of the base framework.
The test currently living in test/Transform/NewGVN are a copy
of the ones in GVN, with proper `XFAIL` (missing features in NewGVN).
A flag will be added in a future commit to enable NewGVN, so that
interested parties can exercise this code easily.
Differential Revision: https://reviews.llvm.org/D26224
llvm-svn: 290346
from the old pass manager in the new one.
I'm not trying to support (initially) the numerous options that are
currently available to customize the pass pipeline. If we end up really
wanting them, we can add them later, but I suspect many are no longer
interesting. The simplicity of omitting them will help a lot as we sort
out what the pipeline should look like in the new PM.
I've also documented to the best of my ability *why* each pass or group
of passes is used so that reading the pipeline is more helpful. In many
cases I think we have some questionable choices of ordering and I've
left FIXME comments in place so we know what to come back and revisit
going forward. But for now, I've left it as similar to the current
pipeline as I could.
Lastly, I've had to comment out several places where passes are not
ported to the new pass manager or where the loop pass infrastructure is
not yet ready. I did at least fix a few bugs in the loop pass
infrastructure uncovered by running the full pipeline, but I didn't want
to go too far in this patch -- I'll come back and re-enable these as the
infrastructure comes online. But I'd like to keep the comments in place
because I don't want to lose track of which passes need to be enabled
and where they go.
One thing that seemed like a significant API improvement was to require
that we don't build pipelines for O0. It seems to have no real benefit.
I've also switched back to returning pass managers by value as at this
API layer it feels much more natural to me for composition. But if
others disagree, I'm happy to go back to an output parameter.
I'm not 100% happy with the testing strategy currently, but it seems at
least OK. I may come back and try to refactor or otherwise improve this
in subsequent patches but I wanted to at least get a good starting point
in place.
Differential Revision: https://reviews.llvm.org/D28042
llvm-svn: 290325
In r267672, where the loop distribution pragma was introduced, I tried
it hard to keep the old behavior for opt: when opt is invoked
with -loop-distribute, it should distribute the loop (it's off by
default when ran via the optimization pipeline).
As MichaelZ has discovered this has the unintended consequence of
breaking a very common developer work-flow to reproduce compilations
using opt: First you print the pass pipeline of clang
with -debug-pass=Arguments and then invoking opt with the returned
arguments.
clang -debug-pass will include -loop-distribute but the pass is invoked
with default=off so nothing happens unless the loop carries the pragma.
While through opt (default=on) we will try to distribute all loops.
This changes opt's default to off as well to match clang. The tests are
modified to explicitly enable the transformation.
llvm-svn: 290235
No existing client is passing a non-null value here. This will come back
in a slightly different form as part of the type identifier summary work.
Differential Revision: https://reviews.llvm.org/D28006
llvm-svn: 290222
We're currently doing nearly the same thing for @llvm.objectsize in
three different places: two of them are missing checks for overflow,
and one of them could subtly break if InstCombine gets much smarter
about removing alloc sites. Seems like a good idea to not do that.
llvm-svn: 290214
Make it clear that TripCount is the upper bound of the iteration on which
control exits LatchBlock.
Differential Revision: https://reviews.llvm.org/D26675
llvm-svn: 290199
This doesn't implement *every* feature of the existing inliner, but
tries to implement the most important ones for building a functional
optimization pipeline and beginning to sort out bugs, regressions, and
other problems.
Notable, but intentional omissions:
- No alloca merging support. Why? Because it isn't clear we want to do
this at all. Active discussion and investigation is going on to remove
it, so for simplicity I omitted it.
- No support for trying to iterate on "internally" devirtualized calls.
Why? Because it adds what I suspect is inappropriate coupling for
little or no benefit. We will have an outer iteration system that
tracks devirtualization including that from function passes and
iterates already. We should improve that rather than approximate it
here.
- Optimization remarks. Why? Purely to make the patch smaller, no other
reason at all.
The last one I'll probably work on almost immediately. But I wanted to
skip it in the initial patch to try to focus the change as much as
possible as there is already a lot of code moving around and both of
these *could* be skipped without really disrupting the core logic.
A summary of the different things happening here:
1) Adding the usual new PM class and rigging.
2) Fixing minor underlying assumptions in the inline cost analysis or
inline logic that don't generally hold in the new PM world.
3) Adding the core pass logic which is in essence a loop over the calls
in the nodes in the call graph. This is a bit duplicated from the old
inliner, but only a handful of lines could realistically be shared.
(I tried at first, and it really didn't help anything.) All told,
this is only about 100 lines of code, and most of that is the
mechanics of wiring up analyses from the new PM world.
4) Updating the LazyCallGraph (in the new PM) based on the *newly
inlined* calls and references. This is very minimal because we cannot
form cycles.
5) When inlining removes the last use of a function, eagerly nuking the
body of the function so that any "one use remaining" inline cost
heuristics are immediately refined, and queuing these functions to be
completely deleted once inlining is complete and the call graph
updated to reflect that they have become dead.
6) After all the inlining for a particular function, updating the
LazyCallGraph and the CGSCC pass manager to reflect the
function-local simplifications that are done immediately and
internally by the inline utilties. These are the exact same
fundamental set of CG updates done by arbitrary function passes.
7) Adding a bunch of test cases to specifically target CGSCC and other
subtle aspects in the new PM world.
Many thanks to the careful review from Easwaran and Sanjoy and others!
Differential Revision: https://reviews.llvm.org/D24226
llvm-svn: 290161
This patch implements PR31013 by introducing a
DIGlobalVariableExpression that holds a pair of DIGlobalVariable and
DIExpression.
Currently, DIGlobalVariables holds a DIExpression. This is not the
best way to model this:
(1) The DIGlobalVariable should describe the source level variable,
not how to get to its location.
(2) It makes it unsafe/hard to update the expressions when we call
replaceExpression on the DIGLobalVariable.
(3) It makes it impossible to represent a global variable that is in
more than one location (e.g., a variable with multiple
DW_OP_LLVM_fragment-s). We also moved away from attaching the
DIExpression to DILocalVariable for the same reasons.
This reapplies r289902 with additional testcase upgrades and a change
to the Bitcode record for DIGlobalVariable, that makes upgrading the
old format unambiguous also for variables without DIExpressions.
<rdar://problem/29250149>
https://llvm.org/bugs/show_bug.cgi?id=31013
Differential Revision: https://reviews.llvm.org/D26769
llvm-svn: 290153
Background/motivation - I was circling back around to:
https://llvm.org/bugs/show_bug.cgi?id=28296
I made a simple patch for that and noticed some regressions, so added test cases for
those with rL281055, and this is hopefully the minimal fix for just those cases.
But as you can see from the surrounding untouched folds, we are missing commuted patterns
all over the place, and of course there are no regression tests to cover any of those cases.
We could sprinkle "m_c_" dust all over this file and catch most of the missing folds, but
then we still wouldn't have test coverage, and we'd still miss some fraction of commuted
patterns because they require adjustments to the match order.
I'm aware of the concern about the potential compile-time performance impact of adding
matches like this (currently being discussed on llvm-dev), but I don't think there's any
evidence yet to suggest that handling commutative pattern matching more thoroughly is not
a worthwhile goal of InstCombine.
Differential Revision: https://reviews.llvm.org/D24419
llvm-svn: 290067
This reverts commit 289920 (again).
I forgot to implement a Bitcode upgrade for the case where a DIGlobalVariable
has not DIExpression. Unfortunately it is not possible to safely upgrade
these variables without adding a flag to the bitcode record indicating which
version they are.
My plan of record is to roll the planned follow-up patch that adds a
unit: field to DIGlobalVariable into this patch before recomitting.
This way we only need one Bitcode upgrade for both changes (with a
version flag in the bitcode record to safely distinguish the record
formats).
Sorry for the churn!
llvm-svn: 289982
This patch reapplies r289863. The original patch was reverted because it
exposed a bug causing the loop vectorizer to crash in the Python runtime on
PPC. The underlying issue was fixed with r289958.
llvm-svn: 289975
After r288909, instructions feeding predicated instructions may be scalarized
if profitable. Since these instructions will remain scalar, we shouldn't
attempt to type-shrink them. We should only truncate vector types to their
minimal bit widths. This bug was exposed by enabling the vectorization of loops
containing conditional stores by default.
llvm-svn: 289958
stores by default
This uncovers a crasher in the loop vectorizer on PPC when building the
Python runtime. I'll send the testcase to the review thread for the
original commit.
llvm-svn: 289934
This patch implements PR31013 by introducing a
DIGlobalVariableExpression that holds a pair of DIGlobalVariable and
DIExpression.
Currently, DIGlobalVariables holds a DIExpression. This is not the
best way to model this:
(1) The DIGlobalVariable should describe the source level variable,
not how to get to its location.
(2) It makes it unsafe/hard to update the expressions when we call
replaceExpression on the DIGLobalVariable.
(3) It makes it impossible to represent a global variable that is in
more than one location (e.g., a variable with multiple
DW_OP_LLVM_fragment-s). We also moved away from attaching the
DIExpression to DILocalVariable for the same reasons.
This reapplies r289902 with additional testcase upgrades.
<rdar://problem/29250149>
https://llvm.org/bugs/show_bug.cgi?id=31013
Differential Revision: https://reviews.llvm.org/D26769
llvm-svn: 289920
Summary:
Instead of checking whether a global referenced by a function being
imported is defined in the same module, speculatively always add the
referenced globals to the module's export list. After all imports are
computed, for each module prune any not in its defined set from its
export list.
For a huge C++ app with aggressive importing thresholds, even with
D27687 we spent a lot of time invoking modulePath() from
exportGlobalInModule (modulePath() was still the 2nd hottest routine in
profile). The reason is that with comdat/linkonce the summary lists for
each GUID can be long. For the app in question, for example, we were
invoking exportGlobalInModule almost 2 million times, and we traversed
an average of 63 entries in the summary list each time.
This patch reduced the thin link time for the app by about 10% (on top
of D27687) when using aggressive importing thresholds, and about 3.5% on
average with default importing thresholds.
Reviewers: mehdi_amini
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D27755
llvm-svn: 289918
This patch implements PR31013 by introducing a
DIGlobalVariableExpression that holds a pair of DIGlobalVariable and
DIExpression.
Currently, DIGlobalVariables holds a DIExpression. This is not the
best way to model this:
(1) The DIGlobalVariable should describe the source level variable,
not how to get to its location.
(2) It makes it unsafe/hard to update the expressions when we call
replaceExpression on the DIGLobalVariable.
(3) It makes it impossible to represent a global variable that is in
more than one location (e.g., a variable with multiple
DW_OP_LLVM_fragment-s). We also moved away from attaching the
DIExpression to DILocalVariable for the same reasons.
<rdar://problem/29250149>
https://llvm.org/bugs/show_bug.cgi?id=31013
Differential Revision: https://reviews.llvm.org/D26769
llvm-svn: 289902
This pass prepares a module containing type metadata for ThinLTO by splitting
it into regular and thin LTO parts if possible, and writing both parts to
a multi-module bitcode file. Modules that do not contain type metadata are
written unmodified as a single module.
All globals with type metadata are added to the regular LTO module, and
the rest are added to the thin LTO module.
Differential Revision: https://reviews.llvm.org/D27324
llvm-svn: 289899
Summary:
We were reinvoking exportGlobalInModule numerous times redundantly.
No need to re-export globals referenced by a global that was already
imported from its module. This resulted in a large speedup in the thin
link for a big application, particularly when importing aggressiveness
was cranked up.
Reviewers: mehdi_amini
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D27687
llvm-svn: 289896
Lowering to llvm.cttz() will result in constant folding anyway
if the argument to ffs is a constant. Pointed out by Eli for
fls() in D14590.
llvm-svn: 289888
The code change for D27687 accidentally got committed along with the
main change in r289843. Revert it temporarily, so that I can recommit it
along with its test as intended.
llvm-svn: 289875
Summary:
Thin link efficiency improvement. After adding an importing candidate to
the worklist we might have later added it again with a higher threshold.
Skip it when popped from the worklist if we recorded a higher threshold
than the current worklist entry, it will get processed again at the
higher threshold when that entry is popped.
This required adding the summary's GUID to the worklist, so that it can
be used to query the recorded highest threshold for it when we pop from the
worklist.
Reviewers: mehdi_amini
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D27696
llvm-svn: 289867
This patch sets the default value of the "-enable-cond-stores-vec" command line
option to "true".
Differential Revision: https://reviews.llvm.org/D27814
llvm-svn: 289863
Now that a new API to merge debug locations has been committed at r289661 (see
review D26256 for more details), we can use it to "improve" the code added by
revision r280995.
Instead of nulling the debugloc of a commoned instruction, we use the 'merged'
debug location. At the moment, this is just a no functional change since
function `DILocation::getMergedLocation()` is just a stub and would always
return a null location.
Differential Revision: https://reviews.llvm.org/D27804
llvm-svn: 289862
Min/max canonicalization (r287585) exposes the fact that we're missing combines for min/max patterns.
This patch won't solve the example that was attached to that thread, so something else still needs fixing.
The line between InstCombine and InstSimplify gets blurry here because sometimes the icmp instruction that
we want to fold to already exists, but sometimes it's the swapped form of what we want.
Corresponding changes for smax/umin/umax to follow.
Differential Revision: https://reviews.llvm.org/D27531
llvm-svn: 289855
This is split out from D27696, since it turned out to be a bug fix and
not part of the NFC efficiency change.
Keep the same adjusted (possibly decayed) threshold in both the worklist
and the ImportList. Otherwise if we encountered it first along a cold
path, the callee would be added to the worklist with a lower decayed
threshold than when it is later encountered along a hot path. But the
logic uses the threshold recorded in the ImportList entry to check if
we should re-add it, and without this patch the threshold recorded there
is the same along both paths so we don't re-add it. Using the
same possibly decayed threshold in the ImportList ensures we re-add it
later with the higher non-decayed hot path threshold.
llvm-svn: 289843
Simplify CFG will try to sink the last instruction in a series of basic blocks,
creating a "common" instruction in the successor block (sinkLastInstruction).
When it does this, the debug location of the single instruction should be the
merged debug locations of the commoned instructions.
Differential Revision: https://reviews.llvm.org/D27590
llvm-svn: 289828
A number of new patterns for simplifying and/xor of icmp:
(icmp ne %x, 0) ^ (icmp ne %y, 0) => icmp ne %x, %y if the following is true:
1- (%x = and %a, %mask) and (%y = and %b, %mask)
2- %mask is a power of 2.
(icmp eq %x, 0) & (icmp ne %y, 0) => icmp ult %x, %y if the following is true:
1- (%x = and %a, %mask1) and (%y = and %b, %mask2)
2- Let %t be the smallest power of 2 where %mask1 & %t != 0. Then for any
%s that is a power of 2 and %s & %mask2 != 0, we must have %s <= %t.
For example if %mask1 = 24 and %mask2 = 16, setting %s = 16 and %t = 8
violates condition (2) above. So this optimization cannot be applied.
llvm-svn: 289813
After r289755, the AssumptionCache is no longer needed. Variables affected by
assumptions are now found by using the new operand-bundle-based scheme. This
new scheme is more computationally efficient, and also we need much less
code...
llvm-svn: 289756
There was an efficiency problem with how we processed @llvm.assume in
ValueTracking (and other places). The AssumptionCache tracked all of the
assumptions in a given function. In order to find assumptions relevant to
computing known bits, etc. we searched every assumption in the function. For
ValueTracking, that means that we did O(#assumes * #values) work in InstCombine
and other passes (with a constant factor that can be quite large because we'd
repeat this search at every level of recursion of the analysis).
Several of us discussed this situation at the last developers' meeting, and
this implements the discussed solution: Make the values that an assume might
affect operands of the assume itself. To avoid exposing this detail to
frontends and passes that need not worry about it, I've used the new
operand-bundle feature to add these extra call "operands" in a way that does
not affect the intrinsic's signature. I think this solution is relatively
clean. InstCombine adds these extra operands based on what ValueTracking, LVI,
etc. will need and then those passes need only search the users of the values
under consideration. This should fix the computational-complexity problem.
At this point, no passes depend on the AssumptionCache, and so I'll remove
that as a follow-up change.
Differential Revision: https://reviews.llvm.org/D27259
llvm-svn: 289755
Summary: SampleProfileLoader pass may be invoked twice by LTO. The 2nd pass should not append more summary info as it is already preset by the 1st pass.
Reviewers: eraman, davidxl
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D27733
llvm-svn: 289725
Summary: We used to create SampleProfileLoader pass in clang. This makes LTO/ThinLTO unable to add this pass in the linker plugin. This patch moves the SampleProfileLoader pass creation from clang to llvm pass manager builder.
Reviewers: tejohnson, davidxl, dnovillo
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D27743
llvm-svn: 289714
If all the operands to a phi node are compares that have a RHS constant,
instcombine will try to pull them through the phi node, combining them into
a single operation. When it does this, the debug location of the new op
should be the merged debug locations of the phi node arguments.
Patch 8 of 8 for D26256. Folding of a compare that has a RHS constant.
Differential Revision: https://reviews.llvm.org/D26256
llvm-svn: 289704
If all the operands to a phi node are a binop with a RHS constant, instcombine
will try to pull them through the phi node, combining them into a single
operation. When it does this, the debug location of the new op should be the
merged debug locations of the phi node arguments.
Patch 7 of 8 for D26256. Folding of a binop with RHS constant.
Differential Revision: https://reviews.llvm.org/D26256
llvm-svn: 289699
Summary:
Move GVNHoist to later in the optimization pipeline, specifically, to
the function simplification part of the pipeline. The new pipeline
location allows GVNHoist to run on a function after its callees have
been inlined but before the function has been considered for inlining
into its callers, exposing more opportunities for hoisting.
Performance results on AArch64 kryo:
Improvements:
Benchmarks/CoyoteBench/fftbench -24.952%
spec2006/bzip2 -4.071%
internal bmark -3.177%
Benchmarks/PAQ8p/paq8p -1.754%
spec2000/perlbmk -1.328%
spec2006/h264ref -1.140%
Regressions:
internal bmark +1.818%
Benchmarks/mafft/pairlocalalign +1.084%
Reviewers: sebpop, dberlin, hiraditya
Subscribers: aemerson, mehdi_amini, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D27722
llvm-svn: 289696
If all the operands to a phi node are a cast, instcombine will try to pull
them through the phi node, combining them into a single cast. When it does
this, the debug location of the new cast should be the merged debug locations
of the phi node arguments.
Patch 6 of 8 for D26256. Folding of a cast operation.
Differential Revision: https://reviews.llvm.org/D26256
llvm-svn: 289693
If all the operands to a phi node are a load, instcombine will try to pull
them through the phi node, combining them into a single load. When it does
this, the debug location of the new load should be the merged debug locations
of the phi node arguments.
Patch 5 of 8 for D26256. Folding of a load operation.
Differential Revision: https://reviews.llvm.org/D26256
llvm-svn: 289688
If all the operands to a phi node are getelementptr, instcombine
will try to pull them through the phi node, combining them into a single
operation. When it does this, the debug location of the new getelementptr
should be the merged debug locations of the phi node arguments.
Patch 4 of 8 for D26256. Folding of a getelementptr operation.
Differential Revision: https://reviews.llvm.org/D26256
llvm-svn: 289684
If all the operands to a phi node are of the same operation, instcombine
will try to pull them through the phi node, combining them into a single
operation. When it does this, the debug location of the operation should
be the merged debug locations of the phi node arguments.
Patch 3 of 8 for D26256. Folding of a compare operation.
Differential Revision: https://reviews.llvm.org/D26256
llvm-svn: 289681
If all the operands to a phi node are of the same operation, instcombine
will try to pull them through the phi node, combining them into a single
operation. When it does this, the debug location of the operation should
be the merged debug locations of the phi node arguments.
Patch 2 of 8 for D26256. Folding of a binary operation.
Differential Revision: https://reviews.llvm.org/D26256
llvm-svn: 289679
Summary: We used to create SampleProfileLoader pass in clang. This makes LTO/ThinLTO unable to add this pass in the linker plugin. This patch moves the SampleProfileLoader pass creation from clang to llvm pass manager builder.
Reviewers: tejohnson, davidxl, dnovillo
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D27743
llvm-svn: 289669
At least the plugin used by the LibreOffice build
(<https://wiki.documentfoundation.org/Development/Clang_plugins>) indirectly
uses those members (through inline functions in LLVM/Clang include files in turn
using them), but they are not exported by utils/extract_symbols.py on Windows,
and accessing data across DLL/EXE boundaries on Windows is generally
problematic.
Differential Revision: https://reviews.llvm.org/D26671
llvm-svn: 289647
Now we only pass bit 0 of the DemandedElts to optimize operand 1 as we recurse since the upper bits are unused. Similarly we clear bit 0 for optimizing operand 0.
Also calculate UndefElts correctly.
Simplify InstCombineCalls for these instrinics to just call SimplifyDemandedVectorElts for the call instrution to reuse this support.
llvm-svn: 289629
Now we only pass bit 0 of the DemandedElts to optimize operand 1 as we recurse since the upper bits are unused.
Also calculate UndefElts correctly.
Simplify InstCombineCalls for these instrinics to just call SimplifyDemandedVectorElts for the call instrution to reuse this support.
llvm-svn: 289628
Summary:
This patch will add loop metadata on the pre and post loops generated by IRCE.
Currently, we have metadata for disabling optimizations such as vectorization,
unrolling, loop distribution and LICM versioning (and confirmed that these
optimizations check for the metadata before proceeding with the transformation).
The pre and post loops generated by IRCE need not go through loop opts (since
these are slow paths).
Added two test cases as well.
Reviewers: sanjoy, reames
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D26806
llvm-svn: 289588
We currently check if the exact trip count is known and is smaller than the
"tiny loop" bound. We should be checking the maximum bound on the trip count
instead.
Differential Revision: https://reviews.llvm.org/D27690
llvm-svn: 289583
Summary:
This is last in of a series of patches to evolve ADCE.cpp to support
removing of unnecessary control flow.
This patch adds the code to update the control and data flow graphs
to remove the dead control flow.
Also update unit tests to test the capability to remove dead,
may-be-infinite loop which is enabled by the switch
-adce-remove-loops.
Previous patches:
D23824 [ADCE] Add handling of PHI nodes when removing control flow
D23559 [ADCE] Add control dependence computation
D23225 [ADCE] Modify data structures to support removing control flow
D23065 [ADCE] Refactor anticipating new functionality (NFC)
D23102 [ADCE] Refactoring for new functionality (NFC)
Reviewers: dberlin, majnemer, nadav, mehdi_amini
Subscribers: llvm-commits, david2050, freik, twoh
Differential Revision: https://reviews.llvm.org/D24918
llvm-svn: 289548
Only the lower bits of the input element are used. And only the lower element can be undef since the upper bits are zeroed.
Have InstCombineCalls call SimplifyDemandedVectorElts for these intrinsics to reuse this support.
llvm-svn: 289523
Summary:
Since we don't break BBs for function calls. We might get some insane counts
(wrap of unsigned) in the presence of noreturn calls.
This patch sets these counts to zero instead of the wrapped number.
Reviewers: davidxl
Subscribers: xur, eraman, llvm-commits
Differential Revision: https://reviews.llvm.org/D27602
llvm-svn: 289521
This patch ensures the correct minimum bit width during type-shrinking.
Previously when type-shrinking, we always sign-extended values back to their
original width. However, if we are going to sign-extend, and the sign bit is
unknown, we have to increase the minimum bit width by one bit so the
sign-extend will fill the upper bits correctly. If the sign bit is known to be
zero, we can perform a zero-extend instead. This should fix PR31243.
Reference: https://llvm.org/bugs/show_bug.cgi?id=31243
Differential Revision: https://reviews.llvm.org/D27466
llvm-svn: 289470
We could truncate the condition and then try to fold the add into the
original condition value causing wrong case constants to be used.
Move the offset transform ahead of the truncate transform and return
after each transform, so there's no chance of getting confused values.
Fix for:
https://llvm.org/bugs/show_bug.cgi?id=31260
llvm-svn: 289442
This teaches SimplifyDemandedElts that the FMA can be removed if the lower element isn't used. It also teaches it that if upper elements of the first operand aren't used then we can simplify them.
llvm-svn: 289377
These intrinsics don't read the upper elements of their first and second input. These are slightly different the the SSE version which does use the upper bits of its first element as passthru bits since the result goes to an XMM register. For AVX-512 the result goes to a mask register instead.
llvm-svn: 289371
These intrinsics don't read the upper bits of their second input. And the third input is the passthru for masking and that only uses the lower element as well.
llvm-svn: 289370
The motivating example is:
extern int patatino;
int goo() {
int x = 0;
for (int i = 0; i < 1000000; ++i) {
x *= patatino;
}
return x;
}
Currently SCCP will not realize that this function returns always zero,
therefore will try to unroll and vectorize the loop at -O3 producing an
awful lot of (useless) code. With this change, it will just produce:
0000000000000000 <g>:
xor %eax,%eax
retq
llvm-svn: 289175
When trying to vectorize trees that start at insertelement instructions
function tryToVectorizeList() uses vectorization factor calculated as
MinVecRegSize/ScalarTypeSize. But sometimes it does not work as tree
cost for this fixed vectorization factor is too high.
Patch tries to improve the situation. It tries different vectorization
factors from max(PowerOf2Floor(NumberOfVectorizedValues),
MinVecRegSize/ScalarTypeSize) to MinVecRegSize/ScalarTypeSize and tries
to choose the best one.
Differential Revision: https://reviews.llvm.org/D27215
llvm-svn: 289043
Replace @progbits in the section directive with %progbits, because "@" starts a comment on arm/thumb.
Use b.w branch instruction.
Use .thumb_function and .thumb_set for proper arm/thumb interwork. This way jumptable entry addresses on thumb have bit 0 set (correctly). This does not affect CFI check math, because the address of the jumptable start also has that bit set.
This does not work on thumbv5, because it does not support b.w, and the linker would not insert a veneer (trampoline?) to extend the range of b.n. We may need to do full-range plt-style jumptables on thumbv54, which are 12 bytes per entry. Another option is "push lr; bl; pop pc" (4 bytes) but that needs unwinding instructions, etc.
Differential Revision: https://reviews.llvm.org/D27499
llvm-svn: 289008
The fix committed in r288851 doesn't cover all the cases.
In particular, if we have an instruction with side effects
which has a no non-dbg use not depending on the bits, we still
perform RAUW destroying the dbg.value's first argument.
Prevent metadata from being replaced here to avoid the issue.
Differential Revision: https://reviews.llvm.org/D27534
llvm-svn: 288987
This patch attempts to scalarize the operand expressions of predicated
instructions if they were conditionally executed in the original loop. After
scalarization, the expressions will be sunk inside the blocks created for the
predicated instructions. The transformation essentially performs
un-if-conversion on the operands.
The cost model has been updated to determine if scalarization is profitable. It
compares the cost of a vectorized instruction, assuming it will be
if-converted, to the cost of the scalarized instruction, assuming that the
instructions corresponding to each vector lane will be sunk inside a predicated
block, possibly avoiding execution. If it's more profitable to scalarize the
entire expression tree feeding the predicated instruction, the expression will
be scalarized; otherwise, it will be vectorized. We only consider the cost of
the entire expression to accurately estimate the cost of the required
insertelement and extractelement instructions.
Differential Revision: https://reviews.llvm.org/D26083
llvm-svn: 288909
In the case of a fully redundant load LI dominated by an equivalent load V, GVN
should always preserve the original debug location of V. Otherwise, we risk to
introduce an incorrect stepping.
If V has debug info, then clearly it should not be modified. If V has a null
debugloc, then it is still potentially incorrect to propagate LI's debugloc
because LI may not post-dominate V.
Differential Revision: https://reviews.llvm.org/D27468
llvm-svn: 288903
When a function F is inlined, InlineFunction extends the debug location of every
instruction inlined from F by adding an InlinedAt.
However, if an instruction has a 'null' debug location, InlineFunction would
propagate the callsite debug location to it. This behavior existed since
revision 210459.
Revision 210459 was originally committed specifically to workaround the lack of
debug information for instructions inlined from intrinsic functions (which are
usually declared with attributes `__always_inline__, __nodebug__`).
The problem with revision 210459 is that it doesn't make any sort of distinction
between instructions inlined from a 'nodebug' function and instructions which
are inlined from a function built with debug info. This issue may lead to
incorrect stepping in the debugger.
This patch works under the assumption that a nodebug function does not have a
DISubprogram. When a function F is inlined into another function G,
InlineFunction checks if F has debug info associated with it.
For nodebug functions, the InlineFunction logic is unchanged (i.e. it would
still propagate the callsite debugloc to the inlined instructions). Otherwise,
InlineFunction no longer propagates the callsite debug location.
Differential Revision: https://reviews.llvm.org/D27462
llvm-svn: 288895
Requesting metadata for a global is a relatively expensive operation as it
involves a map lookup, but it's one that we need to do relatively frequently in
this pass to collect the list of type metadata nodes associated with a global.
This change improves the performance of type metadata queries by prebuilding
data structures that keep the global together with its list of type metadata,
and changing the pass to use that data structure wherever we were previously
passing global references around.
This change also eliminates some O(N^2) behavior by collecting the list of
globals associated with each type identifier during the first pass over the
list of globals rather than visiting each global to compute that list every
time we add a new type identifier.
Reduces pass runtime on a module containing Chrome's vtables from over 60s
to 0.9s.
Differential Revision: https://reviews.llvm.org/D27484
llvm-svn: 288859
BDCE has two phases:
1. It asks SimplifyDemandedBits if all the bits of an instruction are dead, and if so,
replaces all its uses with the constant zero.
2. Then, it asks SimplifyDemandedBits again if the instruction is really dead
(no side effects etc..) and if so, eliminates it.
Now, in 1) if all the bits of an instruction are dead, we may end up replacing a dbg use:
%call = tail call i32 (...) @g() #4, !dbg !15
tail call void @llvm.dbg.value(metadata i32 %call, i64 0, metadata !8, metadata !16), !dbg !17
->
%call = tail call i32 (...) @g() #4, !dbg !15
tail call void @llvm.dbg.value(metadata i32 0, i64 0, metadata !8, metadata !16), !dbg !17
but not eliminating the call because it may have arbitrary side effects.
In other words, we lose some debug informations.
This patch fixes the problem making sure that BDCE does nothing with the instruction if
it has side effects and no non-dbg uses.
Differential Revision: https://reviews.llvm.org/D27471
llvm-svn: 288851
There are two cases handled here:
1) a branch on undef
2) a switch with an undef condition.
Both cases are currently handled by ResolvedUndefsIn. If we have
a branch on undef, we force its value to false (which is trivially
foldable). If we have a switch on undef, we force to the first
constant (which is also foldable).
llvm-svn: 288725
so we can stop using DW_OP_bit_piece with the wrong semantics.
The entire back story can be found here:
http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20161114/405934.html
The gist is that in LLVM we've been misinterpreting DW_OP_bit_piece's
offset field to mean the offset into the source variable rather than
the offset into the location at the top the DWARF expression stack. In
order to be able to fix this in a subsequent patch, this patch
introduces a dedicated DW_OP_LLVM_fragment operation with the
semantics that we used to apply to DW_OP_bit_piece, which is what we
actually need while inside of LLVM. This patch is complete with a
bitcode upgrade for expressions using the old format. It does not yet
fix the DWARF backend to use DW_OP_bit_piece correctly.
Implementation note: We discussed several options for implementing
this, including reserving a dedicated field in DIExpression for the
fragment size and offset, but using an custom operator at the end of
the expression works just fine and is more efficient because we then
only pay for it when we need it.
Differential Revision: https://reviews.llvm.org/D27361
rdar://problem/29335809
llvm-svn: 288683
This solves a secondary problem seen in PR6137:
https://llvm.org/bugs/show_bug.cgi?id=6137#c6
This is similar to the bitwise logic op fold added with:
https://reviews.llvm.org/rL287707
And like that patch, I'm artificially restricting the
transform from vector <-> scalar types until we're sure
that the backend can handle that.
llvm-svn: 288584
For -O0 there might be unreachable BBs, which breaks the assumption that all the
BBs have an auxiliary data structure. In this patch, we add another interface
called findBBInfo() so that a nullptr can be returned for the unreachable BBs
(and the callers can ignore those BBs).
This fixes the bug reported
https://llvm.org/bugs/show_bug.cgi?id=31209
Differential Revision: https://reviews.llvm.org/D27280
llvm-svn: 288528
This reverts commit r288497, as it broke the AArch64 build of Compiler-RT's
builtins (twice: once in r288412 and once in r288497). We should investigate
this offline.
llvm-svn: 288508
When trying to vectorize trees that start at insertelement instructions
function tryToVectorizeList() uses vectorization factor calculated as
MinVecRegSize/ScalarTypeSize. But sometimes it does not work as tree
cost for this fixed vectorization factor is too high.
Patch tries to improve the situation. It tries different vectorization
factors from max(PowerOf2Floor(NumberOfVectorizedValues),
MinVecRegSize/ScalarTypeSize) to MinVecRegSize/ScalarTypeSize and tries
to choose the best one.
Differential Revision: https://reviews.llvm.org/D27215
llvm-svn: 288497
Now that PointerType is no longer a SequentialType, all SequentialTypes
have an associated number of elements, so we can move that information to
the base class, allowing for a number of simplifications.
Differential Revision: https://reviews.llvm.org/D27122
llvm-svn: 288464
As proposed on llvm-dev:
http://lists.llvm.org/pipermail/llvm-dev/2016-October/106640.html
This is for a couple of reasons:
- Values of type PointerType are unlike the other SequentialTypes (arrays
and vectors) in that they do not hold values of the element type. By moving
PointerType we can unify certain aspects of how the other SequentialTypes
are handled.
- PointerType will have no place in the SequentialType hierarchy once
pointee types are removed, so this is a necessary step towards removing
pointee types.
Differential Revision: https://reviews.llvm.org/D26595
llvm-svn: 288462
Instead, expose whether the current type is an array or a struct, if an array
what the upper bound is, and if a struct the struct type itself. This is
in preparation for a later change which will make PointerType derive from
Type rather than SequentialType.
Differential Revision: https://reviews.llvm.org/D26594
llvm-svn: 288458
Summary:
We were doing an optimization in the ThinLTO backends of importing
constant unnamed_addr globals unconditionally as a local copy (regardless
of whether the thin link decided to import them). This should be done in
the thin link instead, so that resulting exported references are marked
and promoted appropriately, but will need a summary enhancement to mark
these variables as constant unnamed_addr.
The function import logic during the thin link was trying to handle
this proactively, by conservatively marking all values referenced in
the initializer lists of exported global variables as also exported.
However, this only handled values referenced directly from the
initializer list of an exported global variable. If the value is itself
a constant unnamed_addr variable, we could end up exporting its
references as well. This caused multiple issues. The first is that the
transitively exported references weren't promoted. Secondly, some could
not be promoted/renamed (e.g. they had a section or other constraint).
recursively, instead of just adding the first level of initializer list
references to the ExportList directly.
Remove this optimization and the associated handling in the function
import backend. SPEC measurements indicate we weren't getting much
from it in any case.
Fixes PR31052.
Reviewers: mehdi_amini
Subscribers: krasin, llvm-commits
Differential Revision: https://reviews.llvm.org/D26880
llvm-svn: 288446
Sanjay Patel