Fix for InlineSpiller accessing not updated dom tree base information.
Summary:
The InlineSpiller was accessing the DominatorTreeBase directly
through the public data member DT in the MachineDominatorTree.
This is not a good idea as the "cached" information in
SplitCriticalEdges is not applied before the access.
The DominatorTreeBase must be accessed through the member
function getBase() in MachineDominatorTree.
The fault was introduced in r266162.
I think the public data member DT in the MachineDominatorTree
should have been made private in the original code (r215576)
that introduced the concept of lazily updating the
MachineDominatorTree information from
MachineBasicBlock::SplitCriticalEdge().
Patch by Karl-Johan Karlsson <karl-johan.karlsson@ericsson.com>
Simon Pilgrim [Wed, 4 Jan 2017 08:05:42 +0000 (08:05 +0000)]
[X86] Attempt to pre-truncate arithmetic operations if useful
In some cases its more efficient to combine TRUNC( BINOP( X, Y ) ) --> BINOP( TRUNC( X ), TRUNC( Y ) ) if the binop is legal for the truncated types.
This is true for vector integer multiplication (especially vXi64), as well as ADD/AND/XOR/OR in cases where we only need to truncate one of the inputs at runtime (e.g. a duplicated input or an one use constant we can fold).
Further work could be done here - scalar cases (especially i64) could often benefit (if we avoid partial registers etc.), other opcodes, and better analysis of when truncating the inputs reduces costs.
I have considered implementing this for all targets within the DAGCombiner but wasn't sure we could devise a suitable cost model system that would give us the range we need.
On a 64-bit system, the DWARFDebugLine::Row struct is 32 bytes. Each field has the following byte offsets:
0-7: Address
8-11: Line
12-13: Column
14-15: File
16-19: Isa
20-23: Discriminator
24+: bit fields
The packing is fine until the "Isa" field, which is an 8-bit int that occupies 4 bytes. We can instead move Discriminator into the 16-19 slot, and pack Isa into the 20-23 range along with the bit fields:
0-7: Address
8-11: Line
12-13: Column
14-15: File
16-19: Discriminator
20-23: Isa + bit fields
This layout is only 24 bytes. This 25% reduction in size may seem small but a large binary can have line tables with thousands of rows stored in a vector.
Greg Clayton [Wed, 4 Jan 2017 00:10:50 +0000 (00:10 +0000)]
Correct the parent testing to avoid the special case where a DIE has a depth of 1
This test was testing that we could correctly find the parent of a DIE, but it was actually just testing the special case where a DIE's depth was 1. This corrects that error by adding an extra level into the the DWARF to ensure that we correctly get the parent by looking for the parent with a depth that is 1 less than the current depth.
Teresa Johnson [Tue, 3 Jan 2017 23:19:29 +0000 (23:19 +0000)]
[ThinLTO] Import type as decl only when non-null Identifier
As per post-commit review for r289993 (D27775), we can only safely
import a type as a decl if it has an Identifier, as the Name alone
is not enough to be unique across modules.
Zachary Turner [Tue, 3 Jan 2017 23:12:36 +0000 (23:12 +0000)]
Fix the MSVC version check.
I'm not sure what determines the minor version, but it appears
that it's possible for a fully updated, release version of
VS2015 with Update 3 can go (at least) as low as 19.00.24213.1.
Updating the compiler version check to account for this so we
don't generate superfluous warnings.
Ahmed Bougacha [Tue, 3 Jan 2017 20:33:22 +0000 (20:33 +0000)]
[CodeGen] Simplify logic that looks for returned call operands. NFC-ish.
Use getReturnedArgOperand() instead of rolling our own. Note that it's
equivalent because there can only be one 'returned' operand.
The existing code was also incorrect: there already was awkward logic to
ignore callee/EH blocks, but operands can now also be operand bundles,
in which case we'll look for non-existent parameter attributes.
Unfortunately, this isn't observable in-tree, as it only crashes when
exercising the regular call lowering logic with operand bundles.
Still, this is a nice small cleanup anyway.
Michal Gorny [Tue, 3 Jan 2017 16:33:50 +0000 (16:33 +0000)]
[ADT] APFloatBase: Prevent collapsing semPPCDoubleDouble and semBogus
Provide a distinct contents for semBogus and semPPCDoubleDouble in order
to prevent compilers from collapsing them to a single memory address,
while we heavily rely on every semantic having distinct address.
This happens if insecure optimization collapsing identical values is
enabled. As a result, APFloats of semBogus are indistinguishable from
semPPCDoubleDouble -- and whenever the move constructor is used, the old
value beings being incorrectly recognized as a semPPCDoubleDouble.
Since the values in semPPCDoubleDouble are not used anywhere,
we can easily solve this issue via altering the value of one of the
fields and therefore ensuring that the collapse can not occur.
Craig Topper [Tue, 3 Jan 2017 07:36:41 +0000 (07:36 +0000)]
[X86] Move 128-bit shuffle mask widening check into lowerV2X128VectorShuffle to reduce code duplication. Use the now available widened mask to simplify some code inside lowerV2X128VectorShuffle.
Craig Topper [Tue, 3 Jan 2017 05:45:57 +0000 (05:45 +0000)]
[AVX-512] Remove vinsert intrinsics and autoupgrade to native shufflevectors. There are some codegen problems here that I'll try to fix in future commits.
Craig Topper [Tue, 3 Jan 2017 05:45:46 +0000 (05:45 +0000)]
[AVX-512] Remove vextract intrinsics and autoupgrade to native shufflevectors. This unfortunately generates some really terrible code without VLX support due to v2i1 and v4i1 not being legal.
[XRay] Merge instrumentation point table emission code into AsmPrinter.
Summary:
No need to have this per-architecture. While there, unify 32-bit ARM's
behaviour with what changed elsewhere and start function names lowercase
as per the coding standards. Individual entry emission code goes to the
entry's own class.
Fully tested on amd64, cross-builds on both ARMs and PowerPC.
Xin Tong [Mon, 2 Jan 2017 20:27:23 +0000 (20:27 +0000)]
Make sure total loop body weight is preserved in loop peeling
Summary:
Regardless how the loop body weight is distributed, we should preserve
total loop body weight. i.e. we should have same weight reaching the body of the loop
or its duplicates in peeled and unpeeled case.
Davide Italiano [Mon, 2 Jan 2017 18:41:34 +0000 (18:41 +0000)]
[NewGVN] Restore old code to placate buildbots.
Apparently my suggestion of using ternary doesn't really work
as clang complains about incompatible types on LHS and RHS. Some
GCC versions happen to accept the code but clang behaviour is
correct here.
Michal Gorny [Mon, 2 Jan 2017 18:19:35 +0000 (18:19 +0000)]
[cmake] Add LLVM_ENABLE_DIA_SDK option, and expose it in LLVMConfig
Add an explicit LLVM_ENABLE_DIA_SDK option to control building support
for DIA SDK-based debugging. Control its value to match whether DIA SDK
support was found and expose it in LLVMConfig (alike LLVM_ENABLE_ZLIB).
Its value is needed for LLDB to determine whether to run tests requiring
DIA support. Currently it is obtained from llvm/Config/config.h;
however, this file is not available for standalone builds. Following
this change, LLDB will be modified to use the value from LLVMConfig.
Emit .cfi_sections before the first .cfi_startproc
GNU as rejects input where .cfi_sections is used after .cfi_startproc,
if the new section differs from the old. Adjust our output to always
emit .cfi_sections before the first .cfi_startproc to minimize necessary
code.
Daniel Berlin [Mon, 2 Jan 2017 18:00:53 +0000 (18:00 +0000)]
NewGVN: Add UnknownExpression and create them for things we can't symbolize. Kill fragile machinery for handling null expressions.
Summary:
This avoids the very fragile code for null expressions. We could also use a denseset that tracks which things have null expressions instead, but that seems pretty fragile and premature optimization.
This resolves a number of infinite loop cases, test reductions coming.
Daniel Berlin [Mon, 2 Jan 2017 18:00:46 +0000 (18:00 +0000)]
NewGVN: Fix PR31480, PR31483, PR31499, by rewriting how memory congruence handling works.
Summary: Previously, we tried to fix up the equivalences during symbolic evaluation. This does not work. Now, we change the equivalences during congruence finding, where it belongs. We also initialize the equivalence table to give a maximal answer.
AVX-512 Loop Vectorizer: Cost calculation for interleave load/store patterns.
X86 target does not provide any target specific cost calculation for interleave patterns.It uses the common target-independent calculation, which gives very high numbers. As a result, the scalar version is chosen in many cases. The situation on AVX-512 is even worse, since we have 3-src shuffles that significantly reduce the cost.
In this patch I calculate the cost on AVX-512. It will allow to compare interleave pattern with gather/scatter and choose a better solution (PR31426).
* Shiffle-broadcast cost will be changed in Simon's upcoming patch.
Florian Hahn [Sun, 1 Jan 2017 13:58:27 +0000 (13:58 +0000)]
[selectiondag] Check PromotedFloats map during expansive checks.
Summary:
`PromotedFloats` needs to be checked in
`DAGTypeLegalizer::PerformExpensiveChecks`. This patch fixes a few type
legalization failures with expansive checks for ARM fp16 tests.
Sanjoy Das [Sat, 31 Dec 2016 22:12:34 +0000 (22:12 +0000)]
Fix an issue with isGuaranteedToTransferExecutionToSuccessor
I'm not sure if this was intentional, but today
isGuaranteedToTransferExecutionToSuccessor returns true for readonly and
argmemonly calls that may throw. This commit changes the function to
not implicitly infer nounwind this way.
Even if we eventually specify readonly calls as not throwing,
isGuaranteedToTransferExecutionToSuccessor is not the best place to
infer that. We should instead teach FunctionAttrs or some other such
pass to tag readonly functions / calls as nounwind instead.
Sanjay Patel [Sat, 31 Dec 2016 17:37:01 +0000 (17:37 +0000)]
[ValueTracking] make dominator tree requirement explicit for isKnownNonNullFromDominatingCondition(); NFCI
I don't think this hole is currently exposed, but I crashed regression tests for
jump-threading and loop-vectorize after I added calls to isKnownNonNullAt() in
InstSimplify as part of trying to solve PR28430:
https://llvm.org/bugs/show_bug.cgi?id=28430
That's because they call into value tracking with a context instruction, but no
other parts of the query structure filled in.
For more background, see the discussion in:
https://reviews.llvm.org/D27855
Philip Reames [Sat, 31 Dec 2016 02:33:22 +0000 (02:33 +0000)]
[SmallPtrSet] Introduce a find primitive and rewrite count/erase in terms of it
This was originally motivated by a compile time problem I've since figured out how to solve differently, but the cleanup seemed useful. We had the same logic - which essentially implemented find - in several places. By commoning them out, I can implement find and allow erase to be inlined at the call sites if profitable.
Philip Reames [Fri, 30 Dec 2016 18:00:55 +0000 (18:00 +0000)]
[CVP] Adjust iteration order to reduce the amount of work required
CVP doesn't care about the order of blocks visited, but by using a pre-order traversal over the graph we can a) not visit unreachable blocks and b) optimize as we go so that analysis of later blocks produce slightly more precise results.
I noticed this via inspection and don't have a concrete example which points to the issue.
Mehdi Amini [Fri, 30 Dec 2016 01:15:50 +0000 (01:15 +0000)]
Fix test change in r290736: restore index generation
I remove one extra line, but because annoyingly llvm-lit does not
clean the output directory before running the test, it didn't fail
locally (the file was present from a previous run).
Dehao Chen [Fri, 30 Dec 2016 00:50:28 +0000 (00:50 +0000)]
Use continuous boosting factor for complete unroll.
Summary:
The current loop complete unroll algorithm checks if unrolling complete will reduce the runtime by a certain percentage. If yes, it will apply a fixed boosting factor to the threshold (by discounting cost). The problem for this approach is that the threshold abruptly. This patch makes the boosting factor a function of runtime reduction percentage, capped by a fixed threshold. In this way, the threshold changes continuously.
The patch also simplified the code by reducing one parameter in UP.
The patch only affects code-gen of two speccpu2006 benchmark:
445.gobmk binary size decreases 0.08%, no performance change.
464.h264ref binary size increases 0.24%, no performance change.
[LICM] Remove unneeded tracking of whether changes were made. NFC.
"Changed" doesn't actually change within the loop, so there's
no reason to keep track of it - we always return false during
analysis and true after the transformation is made.
David Majnemer [Fri, 30 Dec 2016 00:28:58 +0000 (00:28 +0000)]
[InstCombine] More thoroughly canonicalize the position of zexts
We correctly canonicalized (add (sext x), (sext y)) to (sext (add x, y))
where possible. However, we didn't perform the same canonicalization
for zexts or for muls.
projects / llvm / log