Adam Nemet [Fri, 27 Jan 2017 06:38:31 +0000 (06:38 +0000)]
[opt-viewer] Put critical items in parallel
Summary:
Put opt-viewer critical items in parallel
Patch by Brian Cain!
Requires features from Python 2.7
**Performance**
Below are performance results across various configurations. These were taken on an i5-5200U (dual core + HT). They were taken with a small subset of the YAML output of building Python 3.6.0b3 with LTO+PGO. 60 YAML files.
"ImportError" vs "class<...CLoader>" below are just confirming the expected configuration (with/without CLoader).
The below was measured on AMD A8-5500B (4 cores) with 224 input YAML files, showing a ~1.75x speed increase over the baseline with libYAML. I suspect it would scale well on high-end servers.
The Windows on ARM target uses custom division for normal division as
the backend needs to insert division-by-zero checks. However, it is
designed to only handle non-vectorized division. ARM has custom
lowering for vectorized division as that can avoid loading registers
with the values and invoke a division routine for each one, preferring
to lower using NEON instructions. Fall back to the custom lowering for
the NEON instructions if we encounter a vectorized division.
Daniel Berlin [Fri, 27 Jan 2017 02:37:11 +0000 (02:37 +0000)]
NewGVN: Add basic dead and redundant store elimination
Summary:
This adds basic dead and redundant store elimination to
NewGVN. Unlike our current DSE, it will happily do cross-block DSE if
it meets our requirements.
We get a bunch of DSE's simple.ll cases, and some stuff it doesn't.
Unlike DSE, however, we only try to eliminate stores of the same value
to the same memory location, not just general stores to the same
memory location.
Tim Shen [Fri, 27 Jan 2017 02:11:07 +0000 (02:11 +0000)]
[APFloat] Reduce some dispatch boilerplates. NFC.
Summary: This is an attempt to reduce the verbose manual dispatching code in APFloat. This doesn't handle multiple dispatch on single discriminator (e.g. APFloat::add(const APFloat&)), nor handles multiple dispatch on multiple discriminators (e.g. APFloat::convert()).
Justin Lebar [Fri, 27 Jan 2017 00:58:58 +0000 (00:58 +0000)]
[NVPTX] Upgrade NVVM intrinsics in InstCombineCalls.
Summary:
There are many NVVM intrinsics that we can't entirely get rid of, but
that nonetheless often correspond to target-generic LLVM intrinsics.
For example, if flush denormals to zero (ftz) is enabled, we can convert
@llvm.nvvm.ceil.ftz.f to @llvm.ceil.f32. On the other hand, if ftz is
disabled, we can't do this, because @llvm.ceil.f32 will be lowered to a
non-ftz PTX instruction. In this case, we can, however, simplify the
non-ftz nvvm ceil intrinsic, @llvm.nvvm.ceil.f, to @llvm.ceil.f32.
These transformations are particularly useful because they let us
constant fold instructions that appear in libdevice, the bitcode library
that ships with CUDA and essentially functions as its libm.
Justin Lebar [Fri, 27 Jan 2017 00:58:34 +0000 (00:58 +0000)]
[ValueTracking] Add comment that CannotBeOrderedLessThanZero does the wrong thing for powi.
Summary:
CannotBeOrderedLessThanZero(powi(x, exp)) returns true if
CannotBeOrderedLessThanZero(x). But powi(-0, exp) is negative if exp is
odd, so we actually want to return SignBitMustBeZero(x).
Except that also isn't right, because we want to return true if x is
NaN, even if x has a negative sign bit.
What we really need in order to fix this is a consistent approach in
this function to handling the sign bit of NaNs. Without this it's very
difficult to say what the correct behavior here is.
This is technically UB as the LangRef is written today if %a is ever less than
-0. But emitting code that's compliant with the current definition of sqrt
would require a branch, which would then prevent us from matching this idiom in
SelectionDAG (which we do today -- ISD::FSQRT has defined behavior on negative
inputs), because SelectionDAG looks at one BB at a time.
Nothing in LLVM takes advantage of this undefined behavior, as far as we can
tell, and the fact that llvm.sqrt has UB dates from its initial addition to the
LangRef.
Chandler Carruth [Fri, 27 Jan 2017 00:50:21 +0000 (00:50 +0000)]
[PM] Flesh out almost all of the late loop passes.
With this the per-module pass pipeline is *extremely* close to the
legacy PM. The missing pieces are:
- PruneEH (or some equivalent)
- ArgumentPromotion
- LoopLoadElimination
- LoopUnswitch
I'm going to work through those in essentially that order but this seems
like a worthwhile incremental step toward the end state.
One difference in what I have here from the legacy PM is that I've
consolidated some of the per-function passes at the very end of the
pipeline into the main optimization function pipeline. The intervening
passes are *really* uninteresting and so this seems very likely to have
any effect other than minor improvement to locality.
Note that there are still some failures in the test suite, but the
compiler doesn't crash or assert.
Tim Northover [Thu, 26 Jan 2017 23:39:14 +0000 (23:39 +0000)]
GlobalISel: support debug intrinsics.
The translation scheme is mostly cribbed from FastISel, and it's not entirely
convincing semantically. But it does seem to work in the common cases and allow
variables to be printed so it can't be all wrong.
Andrew Kaylor [Thu, 26 Jan 2017 23:27:59 +0000 (23:27 +0000)]
Add intrinsics for constrained floating point operations
This commit introduces a set of experimental intrinsics intended to prevent
optimizations that make assumptions about the rounding mode and floating point
exception behavior. These intrinsics will later be extended to specify
flush-to-zero behavior. More work is also required to model instruction
dependencies in machine code and to generate these instructions from clang
(when required by pragmas and/or command line options that are not currently
supported).
Chandler Carruth [Thu, 26 Jan 2017 23:21:17 +0000 (23:21 +0000)]
[PM] Enable the main loop pass pipelines with everything but
loop-unswitch in the main pipelines for the new PM.
All of these now work, and Clang built using this pipeline can build the
test suite and SPEC without hitting any asserts of ASan failures.
There are still some bugs hiding though -- 7 tests regress with the new
PM. I'm going to be investigating these, but it seems worthwhile to at
least get the pipelines in place so that others can play with them, and
they aren't completely broken.
Davide Italiano [Thu, 26 Jan 2017 23:12:53 +0000 (23:12 +0000)]
[obj2yaml] Produce correct output for invalid relocations.
R_X86_64_NONE can be emitted without a symbol associated (well,
in theory it should never be emitted in an ABI-compliant relocatable
object). So, if there's no symbol associated to a reloc, emit one
with an empty name, instead of crashing.
Daniel Berlin [Thu, 26 Jan 2017 22:21:48 +0000 (22:21 +0000)]
NewGVN: Fix bug exposed by PR31761
Summary:
This does not actually fix the testcase in PR31761 (discussion is
ongoing on the testcase), but does fix a bug it exposes, where stores
were not properly clobbering loads.
We accomplish this by unifying the memory equivalence infratructure
back into the normal congruence infrastructure, and then properly
destroying congruence classes when memory state leaders disappear.
Sanjay Patel [Thu, 26 Jan 2017 22:08:10 +0000 (22:08 +0000)]
[InstCombine] fold (X >>u C) << C --> X & (-1 << C)
We already have this fold when the lshr has one use, but it doesn't need that
restriction. We may be able to remove some code from foldShiftedShift().
Also, move the similar:
(X << C) >>u C --> X & (-1 >>u C)
...directly into visitLShr to help clean up foldShiftByConstOfShiftByConst().
That whole function seems questionable since it is called by commonShiftTransforms(),
but there's really not much in common if we're checking the shift opcodes for every
fold.
Ahmed Bougacha [Thu, 26 Jan 2017 22:07:37 +0000 (22:07 +0000)]
[GlobalISel] Remove duplicate function using variadic templates. NFC.
I think the initial version of r293172 was trying:
std::forward<Args...>(args)...
which doesn't compile. This seems like the correct way:
std::forward<Args>(args)...
Sean Fertile [Thu, 26 Jan 2017 18:59:15 +0000 (18:59 +0000)]
[PPC] cleanup of mayLoad/mayStore flags and memory operands.
1) Explicitly sets mayLoad/mayStore property in the tablegen files on load/store
instructions.
2) Updated the flags on a number of intrinsics indicating that they write
memory.
3) Added SDNPMemOperand flags for some target dependent SDNodes so that they
propagate their memory operand
Replace addEarlyAsPossiblePasses callback with adjustPassManager
This change introduces adjustPassManager target callback giving a
target an opportunity to tweak PassManagerBuilder before pass
managers are populated.
This generalizes and replaces addEarlyAsPossiblePasses target
callback. In particular that can be used to add custom passes to
extension points other than EP_EarlyAsPossible.
Serge Rogatch [Thu, 26 Jan 2017 16:17:03 +0000 (16:17 +0000)]
[XRay][Arm32] Reduce the portion of the stub and implement more staging for tail calls - in LLVM
Summary:
This patch provides more staging for tail calls in XRay Arm32 . When the logging part of XRay is ready for tail calls, its support in the core part of XRay Arm32 may be as easy as changing the number passed to the handler from 1 to 2.
Coupled patch:
- https://reviews.llvm.org/D28674
Nirav Dave [Thu, 26 Jan 2017 16:02:24 +0000 (16:02 +0000)]
In visitSTORE, always use FindBetterChain, rather than only when UseAA is enabled.
* Simplify Consecutive Merge Store Candidate Search
Now that address aliasing is much less conservative, push through
simplified store merging search and chain alias analysis which only
checks for parallel stores through the chain subgraph. This is cleaner
as the separation of non-interfering loads/stores from the
store-merging logic.
When merging stores search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited.
This improves the quality of the output SelectionDAG and the output
Codegen (save perhaps for some ARM cases where we correctly constructs
wider loads, but then promotes them to float operations which appear
but requires more expensive constant generation).
Some minor peephole optimizations to deal with improved SubDAG shapes (listed below)
Additional Minor Changes:
1. Finishes removing unused AliasLoad code
2. Unifies the chain aggregation in the merged stores across code
paths
3. Re-add the Store node to the worklist after calling
SimplifyDemandedBits.
4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
arbitrary, but seems sufficient to not cause regressions in
tests.
5. Remove Chain dependencies of Memory operations on CopyfromReg
nodes as these are captured by data dependence
6. Forward loads-store values through tokenfactors containing
{CopyToReg,CopyFromReg} Values.
7. Peephole to convert buildvector of extract_vector_elt to
extract_subvector if possible (see
CodeGen/AArch64/store-merge.ll)
8. Store merging for the ARM target is restricted to 32-bit as
some in some contexts invalid 64-bit operations are being
generated. This can be removed once appropriate checks are
added.
This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.
Many tests required some changes as memory operations are now
reorderable, improving load-store forwarding. One test in
particular is worth noting:
CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store
forwarding converts a load-store pair into a parallel store and
a memory-realized bitcast of the same value. However, because we
lose the sharing of the explicit and implicit store values we
must create another local store. A similar transformation
happens before SelectionDAG as well.
Rafael Espindola [Thu, 26 Jan 2017 15:02:31 +0000 (15:02 +0000)]
Use shouldAssumeDSOLocal in classifyGlobalReference.
And teach shouldAssumeDSOLocal that ppc has no copy relocations.
The resulting code handle a few more case than before. For example, it
knows that a weak symbol can be resolved to another .o file, but it
will still be in the main executable.
Daniil Fukalov [Thu, 26 Jan 2017 13:33:17 +0000 (13:33 +0000)]
[SCEV] Introduce add operation inlining limit
Inlining in getAddExpr() can cause abnormal computational time in some cases.
New parameter -scev-addops-inline-threshold is intruduced with default value 500.
Daniel Sanders [Thu, 26 Jan 2017 11:10:14 +0000 (11:10 +0000)]
[globalisel] Re-factor ISel matchers into a hierarchy. NFC
Summary:
This should make it possible to easily add everything needed to import all
the existing SelectionDAG rules. It should also serve the likely
kinds of GlobalISel rules (some of which are not currently representable
in SelectionDAG) once we've nailed down the tablegen definition for that.
The hierarchy is as follows:
MatcherRule - A matching rule. Currently used to emit C++ ISel code but will
| also be used to emit test cases and tablegen definitions in the
| near future.
|- Instruction(s) - Represents the instruction to be matched.
|- Instruction Predicate(s) - Test the opcode, arithmetic flags, etc. of an
| instruction.
\- Operand(s) - Represents a particular operand of the instruction. In the
| future, there may be subclasses to test the same predicates
| on multiple operands (including for variadic instructions).
\ Operand Predicate(s) - Test the type, register bank, etc. of an operand.
This is where the ComplexPattern equivalent
will be represented. It's also
nested-instruction matching will live as a
predicate that follows the DefUse chain to the
Def and tests a MatcherRule from that position.
Support for multiple instruction matchers in a rule has been retained from
the existing code but has been adjusted to assert when it is used.
Previously it would silently drop all but the first instruction matcher.
The tablegen-erated file is not functionally changed but has more
parentheses and no longer attempts to format the if-statements since
keeping track of the indentation is tricky in the presence of the matcher
hierarchy. It would be nice to have CMakes tablegen() run the output
through clang-format (when available) so we don't have to complicate
TableGen with pretty-printing.
It's also worth mentioning that this hierarchy will also be able to emit
TableGen definitions and test cases in the near future. This is the reason
for favouring explicit emit*() calls rather than the << operator.
Reviewers: aditya_nandakumar, rovka, t.p.northover, qcolombet, ab
Chandler Carruth [Thu, 26 Jan 2017 10:41:09 +0000 (10:41 +0000)]
[LV] Fix an issue where forming LCSSA in the place that we did would
change the set of uniform instructions in the loop causing an assert
failure.
The problem is that the legalization checking also builds data
structures mapping various facts about the loop body. The immediate
cause was the set of uniform instructions. If these then change when
LCSSA is formed, the data structures would already have been built and
become stale. The included test case triggered an assert in loop
vectorize that was reduced out of the new PM's pipeline.
The solution is to form LCSSA early enough that no information is cached
across the changes made. The only really obvious position is outside of
the main logic to vectorize the loop. This also has the advantage of
removing one case where forming LCSSA could mutate the loop but we
wouldn't track that as a "Changed" state.
If it is significantly advantageous to do some legalization checking
prior to this, we can do a more careful positioning but it seemed best
to just back off to a safe position first.
Simon Dardis [Thu, 26 Jan 2017 10:19:02 +0000 (10:19 +0000)]
[mips] N64 static relocation model support
This patch makes one change to GOT handling and two changes to N64's
relocation model handling. Furthermore, the jumptable encodings have
been corrected for static N64.
Big GOT handling is now done via a new SDNode MipsGotHi - this node is
unconditionally lowered to an lui instruction.
The first change to N64's relocation handling is the lifting of the
restriction that N64 always uses PIC. Now it is possible to target static
environments.
The second change adds support for 64 bit symbols and enables them by
default. Previously N64 had patterns for sym32 mode only. In this mode all
symbols are assumed to have 32 bit addresses. sym32 mode support
is selectable with attribute 'sym32'. A follow on patch for clang will
add the necessary frontend parameter.
Diana Picus [Thu, 26 Jan 2017 09:20:47 +0000 (09:20 +0000)]
[ARM] GlobalISel: Load i1, i8 and i16 args from stack
Add support for loading i1, i8 and i16 arguments from the stack, with or without
the ABI extension flags.
When the ABI extension flags are present, we load a 4-byte value, otherwise we
preserve the size of the load and let the instruction selector replace it with a
LDRB/LDRH. This generates the same thing as DAGISel.
Chandler Carruth [Thu, 26 Jan 2017 08:31:54 +0000 (08:31 +0000)]
[PM] Use PoisoningVH correctly when merely deleting entries in a map
with it.
This code was dereferencing the PoisoningVH which isn't allowed once it
is poisoned. But the code itself really doesn't need to access the
pointer, it is just doing the safe stuff of clearing out data structures
keyed on the pointer value.
Change the code to use iterators to erase directly from a DenseMap. This
is also substantially more efficient as it avoids lots of hashing and
lookups to do the erasure. DenseMap supports iterating behind the
iteration which is fairly easy to implement.
Sadly, I don't have a test case here. I'm not even close and I don't
know that I ever will be. The issue is that several of the tricky
aspects of fixing this only show up when you cause the stack's
SmallVector to be in *EXACTLY* the right location. I only ever got
a reproduction for those with Clang, and only with *exactly* the right
command line flags. Any adjustment, even to seemingly unrelated flags,
would make partial and half-way solutions magically start to "work". In
good news, all of this was caught with the LLVM test suite. Also, there
is no *specific* code here that is untested, just that the old pattern
of code won't immediately fail on any test case I've managed to
contrive.
Jonas Paulsson [Thu, 26 Jan 2017 07:03:25 +0000 (07:03 +0000)]
[TargetTransformInfo] Refactor and improve getScalarizationOverhead()
Refactoring to remove duplications of this method.
New method getOperandsScalarizationOverhead() that looks at the present unique
operands and add extract costs for them. Old behaviour was to just add extract
costs for one operand of the type always, which still happens in
getArithmeticInstrCost() if no operands are provided by the caller.
This is a good start of improving on this, but there are more places
that can be improved by using getOperandsScalarizationOverhead().
Review: Hal Finkel
https://reviews.llvm.org/D29017
Craig Topper [Thu, 26 Jan 2017 05:17:13 +0000 (05:17 +0000)]
[X86] Add demanded elts support for the inputs to pclmul intrinsic
This intrinsic uses bit 0 and bit 4 of an immediate argument to determine which bits of its inputs to read. This patch uses this information to simplify the demanded elements of the input vectors.
Chandler Carruth [Thu, 26 Jan 2017 02:13:50 +0000 (02:13 +0000)]
[PM] Simplify the new PM interface to the loop unroller and expose two
factory functions for the two modes the loop unroller is actually used
in in-tree: simplified full-unrolling and the entire thing including
partial unrolling.
I've also wired these up to nice names so you can express both of these
being in a pipeline easily. This is a precursor to actually enabling
these parts of the O2 pipeline.
Chandler Carruth [Thu, 26 Jan 2017 02:07:20 +0000 (02:07 +0000)]
[Loops] Restructure the LoopInfo verify function so that it more
directly walks the current loop structure verifying that a matching
structure can be found in a freshly computed version.
Also pull things out of containers when necessary once an issue is found
and print them directly.
This makes it substantially easier to debug verification failures as
the process stops at the exact point in the loop nest where they diverge
and has in easily accessed local variables (or printed to stderr
already) the loops and other information needed to analyze the failure.
[LoopUnroll] Properly update loopinfo for runtime unrolling by 2
Even when we don't create a remainder loop (that is, when we unroll by 2), we
may duplicate nested loops into the remainder. This is complicated by the fact
the remainder may itself be either inserted into an outer loop, or at the top
level. In the latter case, we may need to create new top-level loops.
Kevin Enderby [Wed, 25 Jan 2017 23:57:32 +0000 (23:57 +0000)]
Change the test added in r293099 so it does not have the string "llvm-nm" to fix
the clang-x86-windows-msvc2015 bot as the name is "llvm-nm.EXE" in that case.
Adam Nemet [Wed, 25 Jan 2017 23:20:33 +0000 (23:20 +0000)]
New OptimizationRemarkEmitter pass for MIR
This allows MIR passes to emit optimization remarks with the same level
of functionality that is available to IR passes.
It also hooks up the greedy register allocator to report spills. This
allows for interesting use cases like increasing interleaving on a loop
until spilling of registers is observed.
I still need to experiment whether reporting every spill scales but this
demonstrates for now that the functionality works from llc
using -pass-remarks*=<pass>.
Adam Nemet [Wed, 25 Jan 2017 23:20:25 +0000 (23:20 +0000)]
[OptDiag] Split code region out of DiagnosticInfoOptimizationBase
Code region is the only part of this class that is IR-specific. Code
region is moved down in the inheritance tree to a new derived class,
called DiagnosticInfoIROptimization.
All the existing remarks are derived from this new class now.
This allows the new MIR pass-remark classes to be derived from
DiagnosticInfoOptimizationBase.
Also because we keep the name DiagnosticInfoOptimizationBase, the clang
parts don't need any adjustment.
Kevin Enderby [Wed, 25 Jan 2017 21:33:38 +0000 (21:33 +0000)]
Add a warning when the llvm-nm -print-size flag is used on a Mach-O file as
Mach-O files don’t have size information about the symbols in the object file
format unlike ELF.
Also add the part of the fix to llvm-nm that was missed with r290001 so
-arch armv7m works.
projects / llvm / log