This is a follow-up from r163302, which added a transformation to
SimplifyCFG that turns some switches into loads from lookup tables.
It was pointed out that some targets, such as GPUs and deeply embedded
targets, might not find this appropriate, but SimplifyCFG doesn't have
enough information about the target to decide this.
This patch adds the reverse transformation to CodeGenPrep: it turns
loads from lookup tables back into switches for targets where we do not
build jump tables (assuming these are also the targets where lookup
tables are inappropriate).
Hopefully we will eventually get to have target information in
SimplifyCFG, and then this CodeGenPrep transformation can be removed.
llvm-svn: 164206
from the dragonegg build bots when we turned on the full version of the
pass. Included a much reduced test case for this pesky bug, despite
bugpoint's uncooperative behavior.
Also, I audited all the similar code I could find and didn't spot any
other cases where this mistake cropped up.
llvm-svn: 164178
working on FCA splitting. Instead of refusing to form a common type when
there are uses of a subsection of the alloca as well as a use of the
entire alloca, just skip the subsection uses and continue looking for
a whole-alloca use with a type that we can use.
This produces slightly prettier IR I think, and also fixes the other
failure in the test.
llvm-svn: 164146
FCAs. This is essential in order to promote allocas that are used in
struct returns by frontends like Clang. The FCA load would block the
rest of the pass from firing, resulting is significant regressions with
the bullet benchmark in the nightly test suite.
Thanks to Duncan for repeated discussions about how best to do this, and
to both him and Benjamin for review.
This appears to have blocked many places where the pass tries to fire,
and so I'm expect somewhat different results with this fix added.
As with the last big patch, I'm including a change to enable the SROA by
default *temporarily*. Ben is going to remove this as soon as the LNT
bots pick up the patch. I'm just trying to get a round of LNT numbers
from the stable machines in the lab.
NOTE: Four clang tests are expected to fail in the brief window where
this is enabled. Sorry for the noise!
llvm-svn: 164119
new one, and add support for running the new pass in that mode and in
that slot of the pass manager. With this the new pass can completely
replace the old one within the pipeline.
The strategy for enabling or disabling the SSAUpdater logic is to do it
by making the requirement of the domtree analysis optional. By default,
it is required and we get the standard mem2reg approach. This is usually
the desired strategy when run in stand-alone situations. Within the
CGSCC pass manager, we disable requiring of the domtree analysis and
consequentially trigger fallback to the SSAUpdater promotion.
In theory this would allow the pass to re-use a domtree if one happened
to be available even when run in a mode that doesn't require it. In
practice, it lets us have a single pass rather than two which was
simpler for me to wrap my head around.
There is a hidden flag to force the use of the SSAUpdater code path for
the purpose of testing. The primary testing strategy is just to run the
existing tests through that path. One notable difference is that it has
custom code to handle lifetime markers, and one of the tests has been
enhanced to exercise that code.
This has survived a bootstrap and the test suite without serious
correctness issues, however my run of the test suite produced *very*
alarming performance numbers. I don't entirely understand or trust them
though, so more investigation is on-going.
To aid my understanding of the performance impact of the new SROA now
that it runs throughout the optimization pipeline, I'm enabling it by
default in this commit, and will disable it again once the LNT bots have
picked up one iteration with it. I want to get those bots (which are
much more stable) to evaluate the impact of the change before I jump to
any conclusions.
NOTE: Several Clang tests will fail because they run -O3 and check the
result's order of output. They'll go back to passing once I disable it
again.
llvm-svn: 163965
destination.
Updated previous implementation to fix a case not covered:
// PBI: br i1 %x, TrueDest, BB
// BI: br i1 %y, TrueDest, FalseDest
The other case was handled correctly.
// PBI: br i1 %x, BB, FalseDest
// BI: br i1 %y, TrueDest, FalseDest
Also tried to use 64-bit arithmetic instead of APInt with scale to simplify the
computation. Let me know if you have other opinions about this.
llvm-svn: 163954
This is essentially a ground up re-think of the SROA pass in LLVM. It
was initially inspired by a few problems with the existing pass:
- It is subject to the bane of my existence in optimizations: arbitrary
thresholds.
- It is overly conservative about which constructs can be split and
promoted.
- The vector value replacement aspect is separated from the splitting
logic, missing many opportunities where splitting and vector value
formation can work together.
- The splitting is entirely based around the underlying type of the
alloca, despite this type often having little to do with the reality
of how that memory is used. This is especially prevelant with unions
and base classes where we tail-pack derived members.
- When splitting fails (often due to the thresholds), the vector value
replacement (again because it is separate) can kick in for
preposterous cases where we simply should have split the value. This
results in forming i1024 and i2048 integer "bit vectors" that
tremendously slow down subsequnet IR optimizations (due to large
APInts) and impede the backend's lowering.
The new design takes an approach that fundamentally is not susceptible
to many of these problems. It is the result of a discusison between
myself and Duncan Sands over IRC about how to premptively avoid these
types of problems and how to do SROA in a more principled way. Since
then, it has evolved and grown, but this remains an important aspect: it
fixes real world problems with the SROA process today.
First, the transform of SROA actually has little to do with replacement.
It has more to do with splitting. The goal is to take an aggregate
alloca and form a composition of scalar allocas which can replace it and
will be most suitable to the eventual replacement by scalar SSA values.
The actual replacement is performed by mem2reg (and in the future
SSAUpdater).
The splitting is divided into four phases. The first phase is an
analysis of the uses of the alloca. This phase recursively walks uses,
building up a dense datastructure representing the ranges of the
alloca's memory actually used and checking for uses which inhibit any
aspects of the transform such as the escape of a pointer.
Once we have a mapping of the ranges of the alloca used by individual
operations, we compute a partitioning of the used ranges. Some uses are
inherently splittable (such as memcpy and memset), while scalar uses are
not splittable. The goal is to build a partitioning that has the minimum
number of splits while placing each unsplittable use in its own
partition. Overlapping unsplittable uses belong to the same partition.
This is the target split of the aggregate alloca, and it maximizes the
number of scalar accesses which become accesses to their own alloca and
candidates for promotion.
Third, we re-walk the uses of the alloca and assign each specific memory
access to all the partitions touched so that we have dense use-lists for
each partition.
Finally, we build a new, smaller alloca for each partition and rewrite
each use of that partition to use the new alloca. During this phase the
pass will also work very hard to transform uses of an alloca into a form
suitable for promotion, including forming vector operations, speculating
loads throguh PHI nodes and selects, etc.
After splitting is complete, each newly refined alloca that is
a candidate for promotion to a scalar SSA value is run through mem2reg.
There are lots of reasonably detailed comments in the source code about
the design and algorithms, and I'm going to be trying to improve them in
subsequent commits to ensure this is well documented, as the new pass is
in many ways more complex than the old one.
Some of this is still a WIP, but the current state is reasonbly stable.
It has passed bootstrap, the nightly test suite, and Duncan has run it
successfully through the ACATS and DragonEgg test suites. That said, it
remains behind a default-off flag until the last few pieces are in
place, and full testing can be done.
Specific areas I'm looking at next:
- Improved comments and some code cleanup from reviews.
- SSAUpdater and enabling this pass inside the CGSCC pass manager.
- Some datastructure tuning and compile-time measurements.
- More aggressive FCA splitting and vector formation.
Many thanks to Duncan Sands for the thorough final review, as well as
Benjamin Kramer for lots of review during the process of writing this
pass, and Daniel Berlin for reviewing the data structures and algorithms
and general theory of the pass. Also, several other people on IRC, over
lunch tables, etc for lots of feedback and advice.
llvm-svn: 163883
a pair of switch/branch where both depend on the value of the same variable and
the default case of the first switch/branch goes to the second switch/branch.
Code clean up and fixed a few issues:
1> handling the case where some cases of the 2nd switch are invalidated
2> correctly calculate the weight for the 2nd switch when it is a conditional eq
Testing case is modified from Alastair's original patch.
llvm-svn: 163635
The lookup tables did not get built in a deterministic order.
This makes them get built in the order that the corresponding phi nodes
were found.
llvm-svn: 163305
This adds a transformation to SimplifyCFG that attemps to turn switch
instructions into loads from lookup tables. It works on switches that
are only used to initialize one or more phi nodes in a common successor
basic block, for example:
int f(int x) {
switch (x) {
case 0: return 5;
case 1: return 4;
case 2: return -2;
case 5: return 7;
case 6: return 9;
default: return 42;
}
This speeds up the code by removing the hard-to-predict jump, and
reduces code size by removing the code for the jump targets.
llvm-svn: 163302
switch, make sure we include the value for the cases when calculating edge
value from switch to the default destination.
rdar://12241132
llvm-svn: 163270
pointers-to-strong-pointers may be in play. These can lead to retains and
releases happening in unstructured ways, foiling the optimizer. This fixes
rdar://12150909.
llvm-svn: 163180
Scan the body of the loop and find instructions that may trap.
Use this information when deciding if it is safe to hoist or sink instructions.
Notice that we can optimize the search of instructions that may throw in the case of nested loops.
rdar://11518836
llvm-svn: 163132
This code used to only handle malloc-like calls, which do not read memory.
r158919 changed it to check isNoAliasFn(), which includes strdup-like and
realloc-like calls, but it was not checking for dependencies on the memory
read by those calls.
llvm-svn: 163106
We update until we hit a fixpoint. This is probably slow but also
slightly simplifies the code. It should also fix the occasional
invalid domtrees observed when building with expensive checking.
I couldn't find a case where this had a measurable slowdown, but
if someone finds a pathological case where it does we may have
to find a cleverer way of updating dominators here.
Thanks to Duncan for the test case.
llvm-svn: 163091
The old PHI updating code in loop-rotate was replaced with SSAUpdater a while
ago, it has no problems with comples PHIs. What had to be fixed is detecting
whether a loop was already rotated and updating dominators when multiple exits
were present.
This change increases overall code size a bit, mostly due to additional loop
unrolling opportunities. Passes test-suite and selfhost with -verify-dom-info.
Fixes PR7447.
Thanks to Andy for the input on the domtree updating code.
llvm-svn: 162912
This disables malloc-specific optimization when -fno-builtin (or -ffreestanding)
is specified. This has been a problem for a long time but became more severe
with the recent memory builtin improvements.
Since the memory builtin functions are used everywhere, this required passing
TLI in many places. This means that functions that now have an optional TLI
argument, like RecursivelyDeleteTriviallyDeadFunctions, won't remove dead
mallocs anymore if the TLI argument is missing. I've updated most passes to do
the right thing.
Fixes PR13694 and probably others.
llvm-svn: 162841
This optimization is really just replacing allocas wholesale with
globals, there is no scalarization.
The underlying motivation for this patch is to simplify the SROA pass
and focus it on splitting and promoting allocas.
llvm-svn: 162271
The previous fix only checked for simple cycles, use a set to catch longer
cycles too.
Drop the broken check from the ObjectSizeOffsetEvaluator. The BoundsChecking
pass doesn't have to deal with invalid IR like InstCombine does.
llvm-svn: 162120
I really need to find a way to automate this, but I can't come up with a regex
that has no false positives while handling tricky cases like custom check
prefixes.
llvm-svn: 162097
where some fact lake a=b dominates a use in a phi, but doesn't dominate the
basic block itself.
This feature could also be implemented by splitting critical edges, but at least
with the current algorithm reasoning about the dominance directly is faster.
The time for running "opt -O2" in the testcase in pr10584 is 1.003 times slower
and on gcc as a single file it is 1.0007 times faster.
llvm-svn: 162023
- memcpy size is wrongly truncated into 32-bit and treat 8GB memcpy is
0-sized memcpy
- as 0-sized memcpy/memset is already removed before SimplifyMemTransfer
and SimplifyMemSet in visitCallInst, replace 0 checking with
assertions.
- replace getZExtValue() with getLimitedValue() according to
Eli Friedman
llvm-svn: 161923
and allow some optimizations to turn conditional branches into unconditional.
This commit adds a simple control-flow optimization which merges two consecutive
basic blocks which are connected by a single edge. This allows the codegen to
operate on larger basic blocks.
rdar://11973998
llvm-svn: 161852
multiple scalar promotions on a single loop. This also has the effect of
preserving the order of stores sunk out of loops, which is aesthetically
pleasing, and it happens to fix the testcase in PR13542, though it doesn't
fix the underlying problem.
llvm-svn: 161459
An unsigned value converted to floating-point will always be greater than
a negative constant. Unfortunately InstCombine reversed the check so that
unsigned values were being optimized to always be greater than all positive
floating-point constants. <rdar://problem/12029145>
llvm-svn: 161452
We give a bonus for every argument because the argument setup is not needed
anymore when the function is inlined. With this patch we interpret byval
arguments as a compact representation of many arguments. The byval argument
setup is implemented in the backend as an inline memcpy, so to model the
cost as accurately as possible we take the number of pointer-sized elements
in the byval argument and give a bonus of 2 instructions for every one of
those. The bonus is capped at 8 elements, which is the number of stores
at which the x86 backend switches from an expanded inline memcpy to a real
memcpy. It would be better to use the real memcpy threshold from the backend,
but it's not available via TargetData.
This change brings the performance of c-ray in line with gcc 4.7. The included
test case tries to reproduce the c-ray problem to catch regressions for this
benchmark early, its performance is dominated by the inline decision of a
specific call.
This only has a small impact on most code, more on x86 and arm than on x86_64
due to the way the ABI works. When building LLVM for x86 it gives a small
inline cost boost to virtually any function using StringRef or STL allocators,
but only a 0.01% increase in overall binary size. The size of gcc compiled by
clang actually shrunk by a couple bytes with this patch applied, but not
significantly.
llvm-svn: 161413
instsimplify+inline strategy.
The crux of the problem is that instsimplify was reasonably relying on
an invariant that is true within any single function, but is no longer
true mid-inline the way we use it. This invariant is that an argument
pointer != a local (alloca) pointer.
The fix is really light weight though, and allows instsimplify to be
resiliant to these situations: when checking the relation ships to
function arguments, ensure that the argumets come from the same
function. If they come from different functions, then none of these
assumptions hold. All credit to Benjamin Kramer for coming up with this
clever solution to the problem.
llvm-svn: 161410
Previously, MBP essentially aligned every branch target it could. This
bloats code quite a bit, especially non-looping code which has no real
reason to prefer aligned branch targets so heavily.
As Andy said in review, it's still a bit odd to do this without a real
cost model, but this at least has much more plausible heuristics.
Fixes PR13265.
llvm-svn: 161409
This can happen as long as the instruction is not reachable. Instcombine does generate these unreachable malformed selects when doing RAUW
llvm-svn: 160874
of an array element (rather than at the beginning of the element) and extended
into the next element, then the load from the second element was being handled
wrong due to incorrect updating of the notion of which byte to load next. This
fixes PR13442. Thanks to Chris Smowton for reporting the problem, analyzing it
and providing a fix.
llvm-svn: 160711
might be deliberate "one time" leaks, so that leak checkers can find them.
This is a reapply of r160602 with the fix that this time I'm committing the
code I thought I was committing last time; the I->eraseFromParent() goes
*after* the break out of the loop.
llvm-svn: 160664
r160529 that was subsequently reverted. The fix was to not call
GV->eraseFromParent() right before the caller does the same. The existing
testcases already caught this bug if run under valgrind.
llvm-svn: 160602
GetBestDestForJumpOnUndef() assumes there is at least 1 successor, which isn't
true if the block ends in an indirect branch with no successors. Fix this by
bailing out earlier in this case.
llvm-svn: 160546
Fixes PR13371: indvars pass incorrectly substitutes 'undef' values.
I do not like this fix. It's needed until/unless the meaning of undef
changes. It attempts to be complete according to the IR spec, but I
don't have much confidence in the implementation given the difficulty
testing undefined behavior. Worse, this invalidates some of my
hard-fought work on indvars and LSR to optimize pointer induction
variables. It results benchmark regressions, which I'll track
internally. On x86_64 no LTO I see:
-3% huffbench
-3% 400.perlbench
-8% fhourstones
My only suggestion for recovering is to change the meaning of
undef. If we could trust an arbitrary instruction to produce a some
real value that can be manipulated (e.g. incremented) according to
non-undef rules, then this case could be easily handled with SCEV.
llvm-svn: 160421
All SCEV expressions used by LSR formulae must be safe to
expand. i.e. they may not contain UDiv unless we can prove nonzero
denominator.
Fixes PR11356: LSR hoists UDiv.
llvm-svn: 160205
%shr = lshr i64 %key, 3
%0 = load i64* %val, align 8
%sub = add i64 %0, -1
%and = and i64 %sub, %shr
ret i64 %and
to:
%shr = lshr i64 %key, 3
%0 = load i64* %val, align 8
%sub = add i64 %0, 2305843009213693951
%and = and i64 %sub, %shr
ret i64 %and
The demanded bit optimization is actually a pessimization because add -1 would
be codegen'ed as a sub 1. Teach the demanded constant shrinking optimization
to check for negated constant to make sure it is actually reducing the width
of the constant.
rdar://11793464
llvm-svn: 160101
This patch removes ~70 lines in InstCombineLoadStoreAlloca.cpp and makes both functions a bit more aggressive than before :)
In theory, we can be more aggressive when removing an alloca than a malloc, because an alloca pointer should never escape, but we are not taking advantage of this anyway
llvm-svn: 159952
This means we can do cheap DSE for heap memory.
Nothing is done if the pointer excapes or has a load.
The churn in the tests is mostly due to objectsize, since we want to make sure we
don't delete the malloc call before evaluating the objectsize (otherwise it becomes -1/0)
llvm-svn: 159876
another mechanical change accomplished though the power of terrible Perl
scripts.
I have manually switched some "s to 's to make escaping simpler.
While I started this to fix tests that aren't run in all configurations,
the massive number of tests is due to a really frustrating fragility of
our testing infrastructure: things like 'grep -v', 'not grep', and
'expected failures' can mask broken tests all too easily.
Essentially, I'm deeply disturbed that I can change the testsuite so
radically without causing any change in results for most platforms. =/
llvm-svn: 159547
versions of Bash. In addition, I can back out the change to the lit
built-in shell test runner to support this.
This should fix the majority of fallout on Darwin, but I suspect there
will be a few straggling issues.
llvm-svn: 159544
This was done through the aid of a terrible Perl creation. I will not
paste any of the horrors here. Suffice to say, it require multiple
staged rounds of replacements, state carried between, and a few
nested-construct-parsing hacks that I'm not proud of. It happens, by
luck, to be able to deal with all the TCL-quoting patterns in evidence
in the LLVM test suite.
If anyone is maintaining large out-of-tree test trees, feel free to poke
me and I'll send you the steps I used to convert things, as well as
answer any painful questions etc. IRC works best for this type of thing
I find.
Once converted, switch the LLVM lit config to use ShTests the same as
Clang. In addition to being able to delete large amounts of Python code
from 'lit', this will also simplify the entire test suite and some of
lit's architecture.
Finally, the test suite runs 33% faster on Linux now. ;]
For my 16-hardware-thread (2x 4-core xeon e5520): 36s -> 24s
llvm-svn: 159525
The original algorithm only used recursive pair fusion of equal-length
types. This is now extended to allow pairing of any types that share
the same underlying scalar type. Because we would still generally
prefer the 2^n-length types, those are formed first. Then a second
set of iterations form the non-2^n-length types.
Also, a call to SimplifyInstructionsInBlock has been added after each
pairing iteration. This takes care of DCE (and a few other things)
that make the following iterations execute somewhat faster. For the
same reason, some of the simple shuffle-combination cases are now
handled internally.
There is some additional refactoring work to be done, but I've had
many requests for this feature, so additional refactoring will come
soon in future commits (as will additional test cases).
llvm-svn: 159330
Original commit message:
If a constant or a function has linkonce_odr linkage and unnamed_addr, mark it
hidden. Being linkonce_odr guarantees that it is available in every dso that
needs it. Being a constant/function with unnamed_addr guarantees that the
copies don't have to be merged.
llvm-svn: 159272
before the expression root. Any existing operators that are changed to use one
of them needs to be moved between it and the expression root, and recursively
for the operators using that one. When I rewrote RewriteExprTree I accidentally
inverted the logic, resulting in the compacting going down from operators to
operands rather than up from operands to the operators using them, oops. Fix
this, resolving PR12963.
llvm-svn: 159265
// C - zext(bool) -> bool ? C - 1 : C
if (ZExtInst *ZI = dyn_cast<ZExtInst>(Op1))
if (ZI->getSrcTy()->isIntegerTy(1))
return SelectInst::Create(ZI->getOperand(0), SubOne(C), C);
This ends up forming sext i1 instructions that codegen to terrible code. e.g.
int blah(_Bool x, _Bool y) {
return (x - y) + 1;
}
=>
movzbl %dil, %eax
movzbl %sil, %ecx
shll $31, %ecx
sarl $31, %ecx
leal 1(%rax,%rcx), %eax
ret
Without the rule, llvm now generates:
movzbl %sil, %ecx
movzbl %dil, %eax
incl %eax
subl %ecx, %eax
ret
It also helps with ARM (and pretty much any target that doesn't have a sext i1 :-).
The transformation was done as part of Eli's r75531. He has given the ok to
remove it.
rdar://11748024
llvm-svn: 159230
merge all zero-sized alloca's into one, fixing c43204g from the Ada ACATS
conformance testsuite. What happened there was that a variable sized object
was being allocated on the stack, "alloca i8, i32 %size". It was then being
passed to another function, which tested that the address was not null (raising
an exception if it was) then manipulated %size bytes in it (load and/or store).
The optimizers cleverly managed to deduce that %size was zero (congratulations
to them, as it isn't at all obvious), which made the alloca zero size, causing
the optimizers to replace it with null, which then caused the check mentioned
above to fail, and the exception to be raised, wrongly. Note that no loads
and stores were actually being done to the alloca (the loop that does them is
executed %size times, i.e. is not executed), only the not-null address check.
llvm-svn: 159202
The primary advantage is that loop optimizations will be applied in a
stable order. This helps debugging and unit test creation. It is also
a better overall implementation without pathologically bad performance
on deep functions.
On large functions (llvm-stress --size=200000 | opt -loops)
Before: 0.1263s
After: 0.0225s
On deep functions (after tweaking llvm-stress, thanks Nadav):
Before: 0.2281s
After: 0.0227s
See r158790 for more comments.
The loop tree is now consistently generated in forward order, but loop
passes are applied in reverse order over the program. If we have a
loop optimization that prefers forward order, that can easily be
achieved by adding a different type of LoopPassManager.
llvm-svn: 159183
- simplifycfg: invoke undef/null -> unreachable
- instcombine: invoke new -> invoke expect(0, 0) (an arbitrary NOOP intrinsic; only done if the allocated memory is unused, of course)
- verifier: allow invoke of intrinsics (to make the previous step work)
llvm-svn: 159146
hidden. Being linkonce_odr guarantees that it is available in every dso that
needs it. Being a constant/function with unnamed_addr guarantees that the
copies don't have to be merged.
llvm-svn: 159136
This fixes PR5997.
These transforms were disabled because codegen couldn't deal with other
uses of trunc(x). This is now handled by the peephole pass.
This causes no regressions on x86-64.
llvm-svn: 159003
- provide more extensive set of functions to detect library allocation functions (e.g., malloc, calloc, strdup, etc)
- provide an API to compute the size and offset of an object pointed by
Move a few clients (GVN, AA, instcombine, ...) to the new API.
This implementation is a lot more aggressive than each of the custom implementations being replaced.
Patch reviewed by Nick Lewycky and Chandler Carruth, thanks.
llvm-svn: 158919
The present implementation handles only TBAA and FP metadata, discarding everything else.
For debug metadata, the current behavior is maintained (the debug metadata associated with
one of the instructions will be kept, discarding that attached to the other).
This should address PR 13040.
llvm-svn: 158606
Dynamic GEPs created by SROA needed to insert extra "i32 0"
operands to index through structs and arrays to get to the
vector being indexed.
llvm-svn: 158590
example degenerate phi nodes and binops that use themselves in unreachable code.
Thanks to Charles Davis for the testcase that uncovered this can of worms.
llvm-svn: 158508
This patch extends FoldBranchToCommonDest to fold unconditional branches.
For unconditional branches, we fold them if it is easy to update the phi nodes
in the common successors.
rdar://10554090
llvm-svn: 158392
POD type, causing memory corruption when mapping to APInts with bitwidth > 64.
Merge another crash testcase into crash.ll while there.
llvm-svn: 158369
topologies, it is quite possible for a leaf node to have huge multiplicity, for
example: x0 = x*x, x1 = x0*x0, x2 = x1*x1, ... rapidly gives a value which is x
raised to a vast power (the multiplicity, or weight, of x). This patch fixes
the computation of weights by correctly computing them no matter how big they
are, rather than just overflowing and getting a wrong value. It turns out that
the weight for a value never needs more bits to represent than the value itself,
so it is enough to represent weights as APInts of the same bitwidth and do the
right overflow-avoiding dance steps when computing weights. As a side-effect it
reduces the number of multiplies needed in some cases of large powers. While
there, in view of external uses (eg by the vectorizer) I made LinearizeExprTree
static, pushing the rank computation out into users. This is progress towards
fixing PR13021.
llvm-svn: 158358
This saves a cast, and zext is more expensive on platforms with subreg support
than trunc is. This occurs in the BSD implementation of memchr(3), see PR12750.
On the synthetic benchmark from that bug stupid_memchr and bsd_memchr have the
same performance now when not inlining either function.
stupid_memchr: 323.0us
bsd_memchr: 321.0us
memchr: 479.0us
where memchr is the llvm-gcc compiled bsd_memchr from osx lion's libc. When
inlining is enabled bsd_memchr still regresses down to llvm-gcc memchr time,
I haven't fully understood the issue yet, something is grossly mangling the
loop after inlining.
llvm-svn: 158297
-%a + 42
into
42 - %a
previously we were emitting:
-(%a + 42)
This fixes the infinite loop in PR12338. The generated code is still not perfect, though.
Will work on that next
llvm-svn: 158237
can move instructions within the instruction list. If the instruction just
happens to be the one the basic block iterator is pointing to, and it is
moved to a different basic block, then we get into an infinite loop due to
the iterator running off the end of the basic block (for some reason this
doesn't fire any assertions). Original commit message:
Grab-bag of reassociate tweaks. Unify handling of dead instructions and
instructions to reoptimize. Exploit this to more systematically eliminate
dead instructions (this isn't very useful in practice but is convenient for
analysing some testcase I am working on). No need for WeakVH any more: use
an AssertingVH instead.
llvm-svn: 158199
instructions to reoptimize. Exploit this to more systematically eliminate
dead instructions (this isn't very useful in practice but is convenient for
analysing some testcase I am working on). No need for WeakVH any more: use
an AssertingVH instead.
llvm-svn: 158073
replacement to make it at least as generic as the instruction being replaced.
This includes:
* dropping nsw/nuw flags
* getting the least restrictive tbaa and fpmath metadata
* merging ranges
Fixes PR12979.
llvm-svn: 157958
- compute size & offset at the same time. The side-effects of this are that we now support negative GEPs. It's now approaching a phase that it can be reused by other passes (e.g., lowering of the objectsize intrinsic)
- use APInt throughout to handle wrap-arounds
- add support for PHI instrumentation
- add a cache (required for recursive PHIs anyway)
- remove hoisting support for now, since it was wrong in a few cases
sorry for the churn here.. tests will follow soon.
llvm-svn: 157775
This also required making recursive simplifications until
nothing changes or a hard limit (currently 3) is hit.
With the simplification in place indvars can canonicalize
loops of the form
for (unsigned i = 0; i < a-b; ++i)
into
for (unsigned i = 0; i != a-b; ++i)
which used to fail because SCEV created a weird umax expr
for the backedge taken count.
llvm-svn: 157701
The test case feeds the following into InstCombine's visitSelect:
%tobool8 = icmp ne i32 0, 0
%phitmp = select i1 %tobool8, i32 3, i32 0
Then instcombine replaces the right side of the switch with 0, doesn't notice
that nothing changes and tries again indefinitely.
This fixes PR12897.
llvm-svn: 157587
then it doesn't alter the instructions composing it, however it would continue
to move the instructions to just before the expression root. Ensure it doesn't
move them either, so now it really does nothing if there is nothing to do. That
commit also ensured that nsw etc flags weren't cleared if the expression was not
being changed. Tweak this a bit so that it doesn't clear flags on the initial
part of a computation either if that part didn't change but later bits did.
llvm-svn: 157518
with arbitrary topologies (previously it would give up when hitting a diamond
in the use graph for example). The testcase from PR12764 is now reduced from
a pile of additions to the optimal 1617*%x0+208. In doing this I changed the
previous strategy of dropping all uses for expression leaves to one of dropping
all but one use. This works out more neatly (but required a bunch of tweaks)
and is also safer: some recently fixed bugs during recursive linearization were
because the linearization code thinks it completely owns a node if it has no uses
outside the expression it is linearizing. But if the node was also in another
expression that had been linearized (and thus all uses of the node from that
expression dropped) then the conclusion that it is completely owned by the
expression currently being linearized is wrong. Keeping one use from within each
linearized expression avoids this kind of mistake.
llvm-svn: 157467
LowerSwitch::Clusterify : main functinality was replaced with CRSBuilder::optimize, so big part of Clusterify's code was reduced.
test/Transform/LowerSwitch/feature.ll - this test was refactored: grep + count was replaced with FileCheck usage.
llvm-svn: 157384
CHECK. The latter error was hidden by the former, and the test harness
used by e.g. "make check" silently ignored that opt was printing an
error message about an unknown flag instead of running on the test file.
llvm-svn: 157341
leader table. That's because it wasn't expecting instructions to turn up as
leader for a value number that is not its own, but equality propagation could
create this situation. One solution is to have the leader table use a WeakVH
but this slows down GVN by about 5%. Instead just have equality propagation not
add instructions to the leader table, only constants and arguments. In theory
this might cause GVN to run more (each time it changes something it runs again)
but it doesn't seem to occur enough to cause a slow down.
llvm-svn: 157251
so that it can be reused in MemCpyOptimizer. This analysis is needed to remove
an unnecessary memcpy when returning a struct into a local variable.
rdar://11341081
PR12686
llvm-svn: 156776
Instruction::IsIdenticalToWhenDefined.
This manifested itself when inlining two calls to the same function. The
inlined function had a switch statement that returned one of a set of
global variables. Without this modification, the two phi instructions that
chose values from the branches of the switch instruction inlined from the
callee were considered equivalent and jump-threading replaced a load for the
first switch value with a phi selecting from the second switch, thereby
producing incorrect code.
This patch has been tested with "make check-all", "lnt runteste nt", and
llvm self-hosted, and on the original program that had this problem,
wireshark.
<rdar://problem/11025519>
llvm-svn: 156548
refactor code a bit to enable future changes to support run-time information
add support to compute allocation sizes at run-time if penalty > 1 (e.g., malloc(x), calloc(x, y), and VLAs)
llvm-svn: 156515
replace the operands of expressions with only one use with undef and generate
a new expression for the original without using RAUW to update the original.
Thus any copies of the original expression held in a vector may end up
referring to some bogus value - and using a ValueHandle won't help since there
is no RAUW. There is already a mechanism for getting the effect of recursion
non-recursively: adding the value to be recursed on to RedoInsts. But it wasn't
being used systematically. Have various places where recursion had snuck in at
some point use the RedoInsts mechanism instead. Fixes PR12169.
llvm-svn: 156379
<rdar://problem/11291436>.
This is a second attempt at a fix for this, the first was r155468. Thanks
to Chandler, Bob and others for the feedback that helped me improve this.
llvm-svn: 155866
Allow the "SplitCriticalEdge" function to split the edge to a landing pad. If
the pass is *sure* that it thinks it knows what it's doing, then it may go ahead
and specify that the landing pad can have its critical edge split. The loop
unswitch pass is one of these passes. It will split the critical edges of all
edges coming from a loop to a landing pad not within the loop. Doing so will
retain important loop analysis information, such as loop simplify.
llvm-svn: 155817
Target specific types should not be vectorized. As a practical matter,
these types are already register matched (at least in the x86 case),
and codegen does not always work correctly (at least in the ppc case,
and this is not worth fixing because ppc_fp128 is currently broken and
will probably go away soon).
llvm-svn: 155729
instead of getAggregateElement. This has the advantage of being
more consistent and allowing higher-level constant folding to
procede even if an inner extract element cannot be folded.
Make ConstantFoldInstruction call ConstantFoldConstantExpression
on the instruction's operands, making it more consistent with
ConstantFoldConstantExpression itself. This makes sure that
ConstantExprs get TargetData-aware folding before being handed
off as operands for further folding.
This causes more expressions to be folded, but due to a known
shortcoming in constant folding, this currently has the side effect
of stripping a few more nuw and inbounds flags in the non-targetdata
side of constant-fold-gep.ll. This is mostly harmless.
This fixes rdar://11324230.
llvm-svn: 155682
elements to minimize the number of multiplies required to compute the
final result. This uses a heuristic to attempt to form near-optimal
binary exponentiation-style multiply chains. While there are some cases
it misses, it seems to at least a decent job on a very diverse range of
inputs.
Initial benchmarks show no interesting regressions, and an 8%
improvement on SPASS. Let me know if any other interesting results (in
either direction) crop up!
Credit to Richard Smith for the core algorithm, and helping code the
patch itself.
llvm-svn: 155616
constants in C++11 mode. I have no idea why it required such particular
circumstances to get here, the code seems clearly to rely upon unchecked
assumptions.
Specifically, when we decide to form an index into a struct type, we may
have gone through (at least one) zero-length array indexing round, which
would have left the offset un-adjusted, and thus not necessarily valid
for use when indexing the struct type.
This is just an canonicalization step, so the correct thing is to refuse
to canonicalize nonsensical GEPs of this form. Implemented, and test
case added.
Fixes PR12642. Pair debugged and coded with Richard Smith. =] I credit
him with most of the debugging, and preventing me from writing the wrong
code.
llvm-svn: 155466
Original commit message:
Defer some shl transforms to DAGCombine.
The shl instruction is used to represent multiplication by a constant
power of two as well as bitwise left shifts. Some InstCombine
transformations would turn an shl instruction into a bit mask operation,
making it difficult for later analysis passes to recognize the
constsnt multiplication.
Disable those shl transformations, deferring them to DAGCombine time.
An 'shl X, C' instruction is now treated mostly the same was as 'mul X, C'.
These transformations are deferred:
(X >>? C) << C --> X & (-1 << C) (When X >> C has multiple uses)
(X >>? C1) << C2 --> X << (C2-C1) & (-1 << C2) (When C2 > C1)
(X >>? C1) << C2 --> X >>? (C1-C2) & (-1 << C2) (When C1 > C2)
The corresponding exact transformations are preserved, just like
div-exact + mul:
(X >>?,exact C) << C --> X
(X >>?,exact C1) << C2 --> X << (C2-C1)
(X >>?,exact C1) << C2 --> X >>?,exact (C1-C2)
The disabled transformations could also prevent the instruction selector
from recognizing rotate patterns in hash functions and cryptographic
primitives. I have a test case for that, but it is too fragile.
llvm-svn: 155362
1) Make the checked assertions a bit more precise. We really want the
canonical forms coming out of reassociate to be exactly what is
expected.
2) Remove other passes, and switch the test to actually directly check
that reassociate makes the important transforms and
canonicalizations.
3) Fold in a related test case now that we're using FileCheck. Make the
same tidying changes to it.
llvm-svn: 155311
While the patch was perfect and defect free, it exposed a really nasty
bug in X86 SelectionDAG that caused an llc crash when compiling lencod.
I'll put the patch back in after fixing the SelectionDAG problem.
llvm-svn: 155181
The shl instruction is used to represent multiplication by a constant
power of two as well as bitwise left shifts. Some InstCombine
transformations would turn an shl instruction into a bit mask operation,
making it difficult for later analysis passes to recognize the
constsnt multiplication.
Disable those shl transformations, deferring them to DAGCombine time.
An 'shl X, C' instruction is now treated mostly the same was as 'mul X, C'.
These transformations are deferred:
(X >>? C) << C --> X & (-1 << C) (When X >> C has multiple uses)
(X >>? C1) << C2 --> X << (C2-C1) & (-1 << C2) (When C2 > C1)
(X >>? C1) << C2 --> X >>? (C1-C2) & (-1 << C2) (When C1 > C2)
The corresponding exact transformations are preserved, just like
div-exact + mul:
(X >>?,exact C) << C --> X
(X >>?,exact C1) << C2 --> X << (C2-C1)
(X >>?,exact C1) << C2 --> X >>?,exact (C1-C2)
The disabled transformations could also prevent the instruction selector
from recognizing rotate patterns in hash functions and cryptographic
primitives. I have a test case for that, but it is too fragile.
llvm-svn: 155136
This is mostly to test the waters. I'd like to get results from FNT
build bots and other bots running on non-x86 platforms.
This feature has been pretty heavily tested over the last few months by
me, and it fixes several of the execution time regressions caused by the
inlining work by preventing inlining decisions from radically impacting
block layout.
I've seen very large improvements in yacr2 and ackermann benchmarks,
along with the expected noise across all of the benchmark suite whenever
code layout changes. I've analyzed all of the regressions and fixed
them, or found them to be impossible to fix. See my email to llvmdev for
more details.
I'd like for this to be in 3.1 as it complements the inliner changes,
but if any failures are showing up or anyone has concerns, it is just
a flag flip and so can be easily turned off.
I'm switching it on tonight to try and get at least one run through
various folks' performance suites in case SPEC or something else has
serious issues with it. I'll watch bots and revert if anything shows up.
llvm-svn: 154816
When vectorizing pointer types it is important to realize that potential
pairs cannot be connected via the address pointer argument of a load or store.
This is because even after vectorization, the address is still a scalar because
the address of the higher half of the pair is implicit from the address of the
lower half (it need not be, and should not be, explicitly computed).
llvm-svn: 154735
Take this opportunity to generalize the indirectbr bailout logic for
loop transformations. CFG transformations will never get indirectbr
right, and there's no point trying.
llvm-svn: 154386
GEPs, bit casts, and stores reaching it but no other instructions. These
often show up during the iterative processing of the inliner, SROA, and
DCE. Once we hit this point, we can completely remove the alloca. These
were actually showing up in the final, fully optimized code in a bunch
of inliner tests I've been working on, and notably they show up after
LLVM finishes optimizing away all function calls involved in
hash_combine(a, b).
llvm-svn: 154285
speculate. Without this, loop rotate (among many other places) would
suddenly stop working in the presence of debug info. I found this
looking at loop rotate, and have augmented its tests with a reduction
out of a very hot loop in yacr2 where failing to do this rotation costs
sometimes more than 10% in runtime performance, perturbing numerous
downstream optimizations.
This should have no impact on performance without debug info, but the
change in performance when debug info is enabled can be extreme. As
a consequence (and this how I got to this yak) any profiling of
performance problems should be treated with deep suspicion -- they may
have been wildly innacurate of debug info was enabled for profiling. =/
Just a heads up.
llvm-svn: 154263
simplification has been performed. This is a bit less efficient
(requires another ilist walk of the basic blocks) but shouldn't matter
in practice. More importantly, it's just too much work to keep track of
all the various ways the return instructions can be mutated while
simplifying them. This fixes yet another crasher, reported by Daniel
Dunbar.
llvm-svn: 154179
dead code, including dead return instructions in some cases. Otherwise,
we end up having a bogus poniter to a return instruction that blows up
much further down the road.
It turns out that this pattern is both simpler to code, easier to update
in the face of enhancements to the inliner cleanup, and likely cheaper
given that it won't add dead instructions to the list.
Thanks to John Regehr's numerous test cases for teasing this out.
llvm-svn: 154157
This allows us to keep passing reduced masks to SimplifyDemandedBits, but
know about all the bits if SimplifyDemandedBits fails. This allows instcombine
to simplify cases like the one in the included testcase.
llvm-svn: 154011
http://llvm.org/bugs/show_bug.cgi?id=12343
We have not trivial way for splitting edges that are goes from indirect branch. We can do it with some tricks, but it should be additionally discussed. And it is still dangerous due to difficulty of indirect branches controlling.
Fix forbids this case for unswitching.
llvm-svn: 153879
a single missing character. Somehow, this had gone untested. I've added
tests for returns-twice logic specifically with the always-inliner that
would have caught this, and fixed the bug.
Thanks to Matt for the careful review and spotting this!!! =D
llvm-svn: 153832
on a per-callsite walk of the called function's instructions, in
breadth-first order over the potentially reachable set of basic blocks.
This is a major shift in how inline cost analysis works to improve the
accuracy and rationality of inlining decisions. A brief outline of the
algorithm this moves to:
- Build a simplification mapping based on the callsite arguments to the
function arguments.
- Push the entry block onto a worklist of potentially-live basic blocks.
- Pop the first block off of the *front* of the worklist (for
breadth-first ordering) and walk its instructions using a custom
InstVisitor.
- For each instruction's operands, re-map them based on the
simplification mappings available for the given callsite.
- Compute any simplification possible of the instruction after
re-mapping, and store that back int othe simplification mapping.
- Compute any bonuses, costs, or other impacts of the instruction on the
cost metric.
- When the terminator is reached, replace any conditional value in the
terminator with any simplifications from the mapping we have, and add
any successors which are not proven to be dead from these
simplifications to the worklist.
- Pop the next block off of the front of the worklist, and repeat.
- As soon as the cost of inlining exceeds the threshold for the
callsite, stop analyzing the function in order to bound cost.
The primary goal of this algorithm is to perfectly handle dead code
paths. We do not want any code in trivially dead code paths to impact
inlining decisions. The previous metric was *extremely* flawed here, and
would always subtract the average cost of two successors of
a conditional branch when it was proven to become an unconditional
branch at the callsite. There was no handling of wildly different costs
between the two successors, which would cause inlining when the path
actually taken was too large, and no inlining when the path actually
taken was trivially simple. There was also no handling of the code
*path*, only the immediate successors. These problems vanish completely
now. See the added regression tests for the shiny new features -- we
skip recursive function calls, SROA-killing instructions, and high cost
complex CFG structures when dead at the callsite being analyzed.
Switching to this algorithm required refactoring the inline cost
interface to accept the actual threshold rather than simply returning
a single cost. The resulting interface is pretty bad, and I'm planning
to do lots of interface cleanup after this patch.
Several other refactorings fell out of this, but I've tried to minimize
them for this patch. =/ There is still more cleanup that can be done
here. Please point out anything that you see in review.
I've worked really hard to try to mirror at least the spirit of all of
the previous heuristics in the new model. It's not clear that they are
all correct any more, but I wanted to minimize the change in this single
patch, it's already a bit ridiculous. One heuristic that is *not* yet
mirrored is to allow inlining of functions with a dynamic alloca *if*
the caller has a dynamic alloca. I will add this back, but I think the
most reasonable way requires changes to the inliner itself rather than
just the cost metric, and so I've deferred this for a subsequent patch.
The test case is XFAIL-ed until then.
As mentioned in the review mail, this seems to make Clang run about 1%
to 2% faster in -O0, but makes its binary size grow by just under 4%.
I've looked into the 4% growth, and it can be fixed, but requires
changes to other parts of the inliner.
llvm-svn: 153812
The powi intrinsic requires special handling because it always takes a single
integer power regardless of the result type. As a result, we can vectorize
only if the powers are equal. Fixes PR12364.
llvm-svn: 153797
CodeGenPrepare sinks compare instructions down to their uses to prevent
live flags and predicate registers across basic blocks.
PRE of a compare instruction prevents that, forcing the i1 compare
result into a general purpose register. That is usually more expensive
than the redundant compare PRE was trying to eliminate in the first
place.
llvm-svn: 153657
Original commit message for r153521 (aka r153423):
Use the new range metadata in computeMaskedBits and add a new optimization to
instruction simplify that lets us remove an and when loding a boolean value.
llvm-svn: 153587
blocks in the function cloner. This removes the last case of trivially
dead code that I've been seeing in the wild getting inlined, analyzed,
re-inlined, optimized, only to be deleted. Nukes a FIXME from the
cleanup tests.
llvm-svn: 153572
undefined behavior, which Rafael was kind enough to fix.
Original commit message for r153423:
Use the new range metadata in computeMaskedBits and add a new optimization to
instruction simplify that lets us remove an and when loding a boolean value.
llvm-svn: 153521
Original commit message:
Use the new range metadata in computeMaskedBits and add a new optimization to
instruction simplify that lets us remove an and when loading a boolean value.
llvm-svn: 153452
constant-offsets of a common base using the generic GEP-walking logic
I added for computing pointer differences in the same situation.
llvm-svn: 153419
inbounds GEPs. This isn't really necessary for simplifying pointer
differences, but I'm planning to re-use the same code to simplify
pointer comparisons where it is necessary. Since real code almost
exclusively uses inbounds GEPs, it doesn't seem worth it to support the
extra complexity of turning it on and off. If anyone would like that
back, feel free to shout. Note that instcombine will still catch any of
these patterns.
llvm-svn: 153418
aggressively. There are lots of dire warnings about this being expensive
that seem to predate switching to the TrackingVH-based value remapper
that is automatically updated on RAUW. This makes it easy to not just
prune single-entry PHIs, but to fully simplify PHIs, and to recursively
simplify the newly inlined code to propagate PHINode simplifications.
This introduces a bit of a thorny problem though. We may end up
simplifying a branch condition to a constant when we fold PHINodes, and
we would like to nuke any dead blocks resulting from this so that time
isn't wasted continually analyzing them, but this isn't easy. Deleting
basic blocks *after* they are fully cloned and mapped into the new
function currently requires manually updating the value map. The last
piece of the simplification-during-inlining puzzle will require either
switching to WeakVH mappings or some other piece of refactoring. I've
left a FIXME in the testcase about this.
llvm-svn: 153410
* Removed test/lib/llvm.exp - it is no longer needed
* Deleted the dg.exp reading code from test/lit.cfg. There are no dg.exp files
left in the test suite so this code is no longer required. test/lit.cfg is
now much shorter and clearer
* Removed a lot of duplicate code in lit.local.cfg files that need access to
the root configuration, by adding a "root" attribute to the TestingConfig
object. This attribute is dynamically computed to provide the same
information as was previously provided by the custom getRoot functions.
* Documented the config.root attribute in docs/CommandGuide/lit.pod
llvm-svn: 153408
to instead rely on much more generic and powerful instruction
simplification in the function cloner (and thus inliner).
This teaches the pruning function cloner to use instsimplify rather than
just the constant folder to fold values during cloning. This can
simplify a large number of things that constant folding alone cannot
begin to touch. For example, it will realize that 'or' and 'and'
instructions with certain constant operands actually become constants
regardless of what their other operand is. It also can thread back
through the caller to perform simplifications that are only possible by
looking up a few levels. In particular, GEPs and pointer testing tend to
fold much more heavily with this change.
This should (in some cases) have a positive impact on compile times with
optimizations on because the inliner itself will simply avoid cloning
a great deal of code. It already attempted to prune proven-dead code,
but now it will be use the stronger simplifications to prove more code
dead.
llvm-svn: 153403
regressed seriously here, we are no longer removing allocas during
inline cleanup. This appears to be because of lifetime markers "using"
them. =/ I'll look into this shortly.
llvm-svn: 153394
same basic block, and it's not safe to insert code in the successor
blocks if the edges are critical edges. Splitting those edges is
possible, but undesirable, especially on the unwind side. Instead,
make the bottom-up code motion to consider invokes to be part of
their successor blocks, rather than part of their parent blocks, so
that it doesn't push code past them and onto the edges. This fixes
PR12307.
llvm-svn: 153343
Do not call SplitBlockPredecessors on a loop preheader when one of the
predecessors is an indirectbr. Otherwise, you will hit this assert:
!isa<IndirectBrInst>(Preds[i]->getTerminator()) && "Cannot split an edge from an IndirectBrInst"
llvm-svn: 153134
instead of skipping the current loop.
My prior fix was incomplete because of an overzealous compile-time optimization:
Better fix for: <rdar://problem/11049788> Segmentation fault: 11 in LoopStrengthReduce
llvm-svn: 153131
alignment. If that's the case, then we want to make sure that we don't increase
the alignment of the store instruction. Because if we increase it to be "more
aligned" than the pointer, code-gen may use instructions which require a greater
alignment than the pointer guarantees.
<rdar://problem/11043589>
llvm-svn: 152907
It was added in 2007 as the first cut at supporting no-inline
attributes, but we didn't have function attributes of any form at the
time. However, it was added without any mention in the LangRef or other
documentation.
Later on, in 2008, Devang added function notes for 'inline=never' and
then turned them into proper function attributes. From that point
onward, as far as I can tell, the world moved on, and no one has touched
'llvm.noinline' in any meaningful way since.
It's time has now come. We have had better mechanisms for doing this for
a long time, all the frontends I'm aware of use them, and this is just
holding back progress. Given that it was never a documented feature of
the IR, I've provided no auto-upgrade support. If people know of real,
in-the-wild bitcode that relies on this, yell at me and I'll add it, but
I *seriously* doubt anyone cares.
llvm-svn: 152904
Only record IVUsers that are dominated by simplified loop
headers. Otherwise SCEVExpander will crash while looking for a
preheader.
I previously tried to work around this in LSR itself, but that was
insufficient. This way, LSR can continue to run if some uses are not
in simple loops, as long as we don't attempt to analyze those users.
Fixes <rdar://problem/11049788> Segmentation fault: 11 in LoopStrengthReduce
llvm-svn: 152892
correlated pairs of pointer arguments at the callsite. This is designed
to recognize the common C++ idiom of begin/end pointer pairs when the
end pointer is a constant offset from the begin pointer. With the
C-based idiom of a pointer and size, the inline cost saw the constant
size calculation, and this provides the same level of information for
begin/end pairs.
In order to propagate this information we have to search for candidate
operations on a pair of pointer function arguments (or derived from
them) which would be simplified if the pointers had a known constant
offset. Then the callsite analysis looks for such pointer pairs in the
argument list, and applies the appropriate bonus.
This helps LLVM detect that half of bounds-checked STL algorithms
(such as hash_combine_range, and some hybrid sort implementations)
disappear when inlined with a constant size input. However, it's not
a complete fix due the inaccuracy of our cost metric for constants in
general. I'm looking into that next.
Benchmarks showed no significant code size change, and very minor
performance changes. However, specific code such as hashing is showing
significantly cleaner inlining decisions.
llvm-svn: 152752
candidate set for subsequent inlining, try to simplify the arguments to
the inner call site now that inlining has been performed.
The goal here is to propagate and fold constants through deeply nested
call chains. Without doing this, we loose the inliner bonus that should
be applied because the arguments don't match the exact pattern the cost
estimator uses.
Reviewed on IRC by Benjamin Kramer.
llvm-svn: 152556
Typically instcombine has handled this, but pointer differences show up
in several contexts where we would like to get constant folding, and
cannot afford to run instcombine. Specifically, I'm working on improving
the constant folding of arguments used in inline cost analysis with
instsimplify.
Doing this in instsimplify implies some algorithm changes. We have to
handle multiple layers of all-constant GEPs because instsimplify cannot
fold them into a single GEP the way instcombine can. Also, we're only
interested in all-constant GEPs. The result is that this doesn't really
replace the instcombine logic, it's just complimentary and focused on
constant folding.
Reviewed on IRC by Benjamin Kramer.
llvm-svn: 152555
The 'CmpInst::isFalseWhenEqual' function returns 'false' for values other than
simply equality. For instance, it returns 'false' for <= or >=. This isn't the
correct behavior for this transformation, which is checking for strict equality
and non-equality. It was causing the gcc.c-torture/execute/frame-address.c test
to fail because it would completely (and incorrectly) optimize a whole function
into a 'ret i32 0'.
llvm-svn: 152497
traversal, consider nodes for which the only successors are backedges
which the traversal is ignoring to be exit nodes. This fixes a problem
where the bottom-up traversal was failing to visit split blocks along
split loop backedges. This fixes rdar://10989035.
llvm-svn: 152421
