Summary: Patches 202051 and 208013 added calls to LTO's PassManager which unconditionally add LoopVectorizePass and SLPVectorizerPass instead of following the logic in PassManagerBuilder::populateModulePassManager and honoring the -vectorize-loops -run-slp-after-loop-vectorization flags.
Reviewers: nadav, aschwaighofer, yijiang
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5884
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220345 91177308-0d34-0410-b5e6-96231b3b80d8
These are named following the IEEE-754 names for these
functions, rather than the libm fmin / fmax to avoid
possible ambiguities. Some languages may implement something
resembling fmin / fmax which return NaN if either operand is
to propagate errors. These implement the IEEE-754 semantics
of returning the other operand if either is a NaN representing
missing data.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220341 91177308-0d34-0410-b5e6-96231b3b80d8
combineMetadata is used when merging two instructions into one. This change teaches it how to merge 'nonnull' - i.e. only preserve it on the new instruction if it's set on both sources. This isn't actually used yet since I haven't adjusted any of the call sites to pass in nonnull as a 'known metadata'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220325 91177308-0d34-0410-b5e6-96231b3b80d8
When changing the type of a load in Chandler's recent InstCombine changes, we can preserve the new 'nonnull' metadata.
I considered adding an assert since 'nonnull' is only valid on pointer types, but casting a pointer to a non-pointer would involve more than a bitcast anyways. If someone extends this transform to handle more than bitcasts, the verifier will report the malformed IR, so a separate assertion isn't needed. Also, the fpmath flags would have the same problem.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220324 91177308-0d34-0410-b5e6-96231b3b80d8
This function was complicated by the fact that it tried to perform
canonicalizations that were already preformed by InstSimplify. Remove
this extra code and move the tests over to InstSimplify. Add asserts to
make sure our preconditions hold before we make any assumptions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220314 91177308-0d34-0410-b5e6-96231b3b80d8
inttoptr or ptrtoint cast provided there is datalayout available.
Eventually, the datalayout can just be required but in practice it will
always be there today.
To go with the ability to expose available values requiring a ptrtoint
or inttoptr cast, helpers are added to perform one of these three casts.
These smarts are necessary to finish canonicalizing loads and stores to
the operational type requirements without regressing fundamental
combines.
I've added some test cases. These should actually improve as the load
combining and store combining improves, but they may fundamentally be
highlighting some missing combines for select in addition to exercising
the specific added logic to load analysis.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220277 91177308-0d34-0410-b5e6-96231b3b80d8
When functions are inlined, instructions without debug information are
attributed to the call site's DebugLoc. After inlining, inlined static
allocas are moved to the caller's entry block, adjacent to the caller's
original static alloca instructions. By retaining the call site's
DebugLoc, these instructions could cause instructions that were
subsequently inserted at the entry block to pick up the same DebugLoc.
Patch by Wolfgang Pieb!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220255 91177308-0d34-0410-b5e6-96231b3b80d8
Our metadata scheme lazily assigns IDs to string metadata, but we have a mechanism to preassign them as well. Using a preassigned ID is helpful since we get compile time type checking, and avoid some (minimal) string construction and comparison. This change adds enum value for three existing metadata types:
+ MD_nontemporal = 9, // "nontemporal"
+ MD_mem_parallel_loop_access = 10, // "llvm.mem.parallel_loop_access"
+ MD_nonnull = 11 // "nonnull"
I went through an updated various uses as well. I made no attempt to get all uses; I focused on the ones which were easily grepable and easily to translate. For example, there were several items in LoopInfo.cpp I chose not to update.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220248 91177308-0d34-0410-b5e6-96231b3b80d8
r220178. First, the creation routine doesn't insert prior to the
terminator of the basic block provided, but really at the end of the
basic block. Instead, get the terminator and insert before that. The
next issue was that we need to ensure multiple PHI node entries for
a single predecessor re-use the same cast instruction rather than
creating new ones.
All of the logic here was without tests previously. I've reduced and
added a test case from the test suite that crashed without both of these
fixes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220186 91177308-0d34-0410-b5e6-96231b3b80d8
logic to look through pointer casts, making them trivially stronger in
the face of loads and stores with intervening pointer casts.
I've included a few test cases that demonstrate the kind of folding
instcombine can do without pointer casts and then variations which
obfuscate the logic through bitcasts. Without this patch, the variations
all fail to optimize fully.
This is more important now than it has been in the past as I've started
moving the load canonicialization to more closely follow the value type
requirements rather than the pointer type requirements and thus this
needs to be prepared for more pointer casts. When I made the same change
to stores several test cases regressed without logic along these lines
so I wanted to systematically improve matters first.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220178 91177308-0d34-0410-b5e6-96231b3b80d8
loads.
This handles many more cases than just the AA metadata, some of them
suggested by Hal in his review of the AA metadata handling patch. I've
tried to test this behavior where tractable to do so.
I'll point out that I have specifically *not* included a test for
debuginfo because it was going to require 2 or 3 times as much work to
craft some input which would survive the "helpful" stripping of debug
info metadata that doesn't match the desired schema. This is another
good example of why the current state of write-ability for our debug
info metadata is unacceptable. I spent over 30 minutes trying to conjure
some test case that would survive, even copying from other debug info
tests, but it always failed to survive with no explanation of why or how
I might fix it. =[
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220165 91177308-0d34-0410-b5e6-96231b3b80d8
The following implements the optimization for sequences of the form:
icmp eq/ne (shl Const2, A), Const1
Such sequences can be transformed to:
icmp eq/ne A, (TrailingZeros(Const1) - TrailingZeros(Const2))
This handles only the equality operators for now. Other operators need
to be handled.
Patch by Ankur Garg!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220162 91177308-0d34-0410-b5e6-96231b3b80d8
by my refactoring of this code.
The method isSafeToLoadUnconditionally assumes that the load will
proceed with the preferred type alignment. Given that, it has to ensure
that the alloca or global is at least that aligned. It has always done
this historically when a datalayout is present, but has never checked it
when the datalayout is absent. When I refactored the code in r220156,
I exposed this path when datalayout was present and that turned the
latent bug into a patent bug.
This fixes the issue by just removing the special case which allows
folding things without datalayout. This isn't worth the complexity of
trying to tease apart when it is or isn't safe without actually knowing
the preferred alignment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220161 91177308-0d34-0410-b5e6-96231b3b80d8
...)) and (load (cast ...)): canonicalize toward the former.
Historically, we've tried to load using the type of the *pointer*, and
tried to match that type as closely as possible removing as many pointer
casts as we could and trading them for bitcasts of the loaded value.
This is deeply and fundamentally wrong.
Repeat after me: memory does not have a type! This was a hard lesson for
me to learn working on SROA.
There is only one thing that should actually drive the type used for
a pointer, and that is the type which we need to use to load from that
pointer. Matching up pointer types to the loaded value types is very
useful because it minimizes the physical size of the IR required for
no-op casts. Similarly, the only thing that should drive the type used
for a loaded value is *how that value is used*! Again, this minimizes
casts. And in fact, the *only* thing motivating types in any part of
LLVM's IR are the types used by the operations in the IR. We should
match them as closely as possible.
I've ended up removing some tests here as they were testing bugs or
behavior that is no longer present. Mostly though, this is just cleanup
to let the tests continue to function as intended.
The only fallout I've found so far from this change was SROA and I have
fixed it to not be impeded by the different type of load. If you find
more places where this change causes optimizations not to fire, those
too are likely bugs where we are assuming that the type of pointers is
"significant" for optimization purposes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220138 91177308-0d34-0410-b5e6-96231b3b80d8
cases where the alloca type, the load types, and the store types used
all disagree.
Previously, the only way that vector-based promotion occured was if the
alloca type was a vector type. This was one of the *very* few remaining
uses of the alloca's type to guide SROA/mem2reg left in LLVM. It turns
out it was a bad idea.
The alloca type can change very easily based on the mixture of types
loaded and stored to that alloca. We shouldn't be relying on it as
a signal for very much. Instead, the source of truth should be loads and
stores. We should canonicalize the loads and stores as much as possible
and then rely on them exclusively in SROA.
When looking and loads and stores, we may find many different candidate
vector types. This change will let SROA try all of them to find a vector
type which is a viable way to promote the entire alloca to a vector
register.
With this change, it becomes possible to do better canonicalization and
optimization of loads and stores without breaking SROA in random ways,
and that should allow fixing a core source of performance loss in hot
numerical loops such as those in Eigen.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220116 91177308-0d34-0410-b5e6-96231b3b80d8
This reverts commit r219899.
This also updates byval-tail-call.ll to make it clear what was breaking.
Adding r219899 again will cause the load/store to disappear.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220093 91177308-0d34-0410-b5e6-96231b3b80d8
DSE's overlap checking contained special logic, used only when no DataLayout
was available, which inferred a complete overwrite when the pointee types were
equal. This logic seems fine for regular loads/stores, but does not work for
memcpy and friends. Instead of fixing this, I'm just removing it.
Philosophically, transformations should not contain enhanced behavior used only
when data layout is lacking (data layout should be strictly additive), and
maintaining these rarely-tested code paths seems not worthwhile at this stage.
Credit to Aliaksei Zasenka for the bug report and the diagnosis. The test case
(slightly reduced from that provided by Aliaksei) replaces the original
contents of test/Transforms/DeadStoreElimination/no-targetdata.ll -- a few
other tests have been updated to have a data layout.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220035 91177308-0d34-0410-b5e6-96231b3b80d8
'AS'.
Using 'S' as this was a terrible idea. Arguably, 'AS' is not much
better, but it at least follows the idea of using initialisms and
removes active confusion about the AllocaSlices variable and a Slice
variable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219963 91177308-0d34-0410-b5e6-96231b3b80d8
clang-modernize.
I did have to clean up the variable types and whitespace a bit because
the use of auto made the code much less readable here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219962 91177308-0d34-0410-b5e6-96231b3b80d8
ArrayRef accessors.
I think this even came up in review that this was over-engineered, and
indeed it was. Time to un-build it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219958 91177308-0d34-0410-b5e6-96231b3b80d8
iterators.
There are a ton of places where it essentially wants ranges
rather than just iterators. This is just the first step that adds the
core slice range typedefs and uses them in a couple of places. I still
have to explicitly construct them because they've not been punched
throughout the entire set of code. More range-based cleanups incoming.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219955 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
Currently, call slot optimization requires that if the destination is an
argument, the argument has the sret attribute. This is to ensure that
the memory access won't trap. In addition to sret, we can also allow the
optimization to happen for arguments that have the new dereferenceable
attribute, which gives the same guarantee.
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5832
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219950 91177308-0d34-0410-b5e6-96231b3b80d8
If a square root call has an FP multiplication argument that can be reassociated,
then we can hoist a repeated factor out of the square root call and into a fabs().
In the simplest case, this:
y = sqrt(x * x);
becomes this:
y = fabs(x);
This patch relies on an earlier optimization in instcombine or reassociate to put the
multiplication tree into a canonical form, so we don't have to search over
every permutation of the multiplication tree.
Because there are no IR-level FastMathFlags for intrinsics (PR21290), we have to
use function-level attributes to do this optimization. This needs to be fixed
for both the intrinsics and in the backend.
Differential Revision: http://reviews.llvm.org/D5787
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219944 91177308-0d34-0410-b5e6-96231b3b80d8
The code committed in r219832 asserted when it attempted to shrink a switch
statement whose type was larger than 64-bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219902 91177308-0d34-0410-b5e6-96231b3b80d8
Make tail recursion elimination a bit more aggressive. This allows us to get
tail recursion on functions that are just branches to a different function. The
fact that the function takes a byval argument does not restrict it from being
optimised into just a tail call.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219899 91177308-0d34-0410-b5e6-96231b3b80d8
For pointer-typed function arguments, enhanced alignment can be asserted using
the 'align' attribute. When inlining, if this enhanced alignment information is
not otherwise available, preserve it using @llvm.assume-based alignment
assumptions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219876 91177308-0d34-0410-b5e6-96231b3b80d8
Truncate the operands of a switch instruction to a narrower type if the upper
bits are known to be all ones or zeros.
rdar://problem/17720004
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219832 91177308-0d34-0410-b5e6-96231b3b80d8
The SLP vectorizer should not vectorize ephemeral values. These are used to
express information to the optimizer, and vectorizing them does not lead to
faster code (because the ephemeral values are dropped prior to code generation,
vectorized or not), and obscures the information the instructions are
attempting to communicate (the logic that interprets the arguments to
@llvm.assume generically does not understand vectorized conditions).
Also, uses by ephemeral values are free (because they, and the necessary
extractelement instructions, will be dropped prior to code generation).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219816 91177308-0d34-0410-b5e6-96231b3b80d8
A few minor changes to prevent @llvm.assume from interfering with loop
vectorization. First, treat @llvm.assume like the lifetime intrinsics, which
are scalarized (but don't otherwise interfere with the legality checking).
Second, ignore the cost of ephemeral instructions in the loop (these will go
away anyway during CodeGen).
Alignment assumptions and other uses of @llvm.assume can often end up inside of
loops that should be vectorized (this is not uncommon for assumptions generated
by __attribute__((align_value(n))), for example).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219741 91177308-0d34-0410-b5e6-96231b3b80d8
Eliminate library calls and intrinsic calls to fabs when the input
is a squared value.
Note that no unsafe-math / fast-math assumptions are needed for
this optimization.
Differential Revision: http://reviews.llvm.org/D5777
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219717 91177308-0d34-0410-b5e6-96231b3b80d8
We assumed that A must be greater than B because the right hand side of
a remainder operator must be nonzero.
However, it is possible for A to be less than B if Pow2 is a power of
two greater than 1.
Take for example:
i32 %A = 0
i32 %B = 31
i32 Pow2 = 2147483648
((Pow2 << 0) >>u 31) is non-zero but A is less than B.
This fixes PR21274.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219713 91177308-0d34-0410-b5e6-96231b3b80d8
This is the same optimization of r219233 with modifications to support PHIs with multiple incoming edges from the same block
and a test to check that this condition is handled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219656 91177308-0d34-0410-b5e6-96231b3b80d8
the IR going into it and to clean up the IR produced by the vectorizers.
Note that these are *off by default* right now while folks collect data
on whether the performance tradeoff is reasonable.
In a build of the 'opt' binary, I see about 2% compile time regression
due to this change on average. This is in my mind essentially the worst
expected case: very little of the opt binary is going to *benefit* from
these extra passes.
I've seen several benchmarks improve in performance my small amounts due
to running these passes, and there are certain (rare) cases where these
passes make a huge difference by either enabling the vectorizer at all
or by hoisting runtime checks out of the outer loop. My primary
motivation is to prevent people from seeing runtime check overhead in
benchmarks where the existing passes and optimizers would be able to
eliminate that.
I've chosen the sequence of passes based on the kinds of things that
seem likely to be relevant for the code at each stage: rotaing loops for
the vectorizer, finding correlated values, loop invariants, and
unswitching opportunities from any runtime checks, and cleaning up
commonalities exposed by the SLP vectorizer.
I'll be pinging existing threads where some of these issues have come up
and will start new threads to get folks to benchmark and collect data on
whether this is the right tradeoff or we should do something else.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219644 91177308-0d34-0410-b5e6-96231b3b80d8
We assumed that negation operations of the form (0 - %Z) resulted in a
negative number. This isn't true if %Z was originally negative.
Substituting the negative number into the remainder operation may result
in undefined behavior because the dividend might be INT_MIN.
This fixes PR21256.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219639 91177308-0d34-0410-b5e6-96231b3b80d8
We have a transform that changes:
(x lshr C1) udiv C2
into:
x udiv (C2 << C1)
However, it is unsafe to do so if C2 << C1 discards any of C2's bits.
This fixes PR21255.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219634 91177308-0d34-0410-b5e6-96231b3b80d8
A helper routine, MultiplyOverflows, was a less efficient
reimplementation of APInt's smul_ov and umul_ov. While we are here,
clean up the code so it's more uniform.
No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219583 91177308-0d34-0410-b5e6-96231b3b80d8
Consider the case where X is 2. (2 <<s 31)/s-2147483648 is zero but we
would fold to X. Note that this is valid when we are in the unsigned
domain because we require NUW: 2 <<u 31 results in poison.
This fixes PR21245.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219568 91177308-0d34-0410-b5e6-96231b3b80d8
consider:
C1 = INT_MIN
C2 = -1
C1 * C2 overflows without a doubt but consider the following:
%x = i32 INT_MIN
This means that (%X /s C1) is 1 and (%X /s C1) /s C2 is -1.
N. B. Move the unsigned version of this transform to InstSimplify, it
doesn't create any new instructions.
This fixes PR21243.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219567 91177308-0d34-0410-b5e6-96231b3b80d8
consider:
mul i32 nsw %x, -2147483648
this instruction will not result in poison if %x is 1
however, if we transform this into:
shl i32 nsw %x, 31
then we will be generating poison because we just shifted into the sign
bit.
This fixes PR21242.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@219566 91177308-0d34-0410-b5e6-96231b3b80d8