While the IndVars issue (PR50384) has been resolved,
and the compile performance improved, a new blocker emerged,
the codegen machine instruction scheduling is also quadratic.
So we still can't really specify the right value here.
Filed PR50584.
`__profd_*` variables are referenced by code only when value profiling is
enabled. If disabled (e.g. default -fprofile-instr-generate), the symbols just
waste space on ELF/Mach-O. We change the comdat symbol from `__profd_*` to
`__profc_*` because an internal symbol does not provide deduplication features
on COFF. The choice doesn't matter on ELF.
(In -DLLVM_BUILD_INSTRUMENTED_COVERAGE=on build, there is now no `__profd_*` symbols.)
On Windows this enables further optimization. We are no longer affected by the
link.exe limitation: an external symbol in IMAGE_COMDAT_SELECT_ASSOCIATIVE can
cause duplicate definition error.
https://lists.llvm.org/pipermail/llvm-dev/2021-May/150758.html
We can thus use llvm.compiler.used instead of llvm.used like ELF (D97585).
This avoids many `/INCLUDE:` directives in `.drectve`.
Here is rnk's measurement for Chrome:
```
This reduced object file size of base_unittests.exe, compiled with coverage, optimizations, and gmlt debug info by 10%:
#BEFORE
$ find . -iname '*.obj' | xargs du -b | awk '{ sum += $1 } END { print sum}'
1047758867
$ du -cksh base_unittests.exe
82M base_unittests.exe
82M total
# AFTER
$ find . -iname '*.obj' | xargs du -b | awk '{ sum += $1 } END { print sum}'
937886499
$ du -cksh base_unittests.exe
78M base_unittests.exe
78M total
```
The change is NFC for Mach-O.
Reviewed By: davidxl, rnk
Differential Revision: https://reviews.llvm.org/D103372
When SimplifyIndVars infers IR nowrap flags from SCEV, this may
happen in two ways: Either nowrap flags were already present in
SCEV and just get transferred to IR. Or zero/sign extension of
addrecs infers additional nowrap flags, and those get transferred
to IR. In the latter case, calling forgetValue() ensures that the
newly inferred nowrap flags get propagated to any other SCEV
expressions based on the addrec. However, the invalidation can
also have a major compile-time effect in some cases. For
https://bugs.llvm.org/show_bug.cgi?id=50384 with n=512 compile-
time drops from 7.1s to 0.8s without this invalidation. At the
same time, removing the invalidation doesn't affect any codegen
in test-suite.
Differential Revision: https://reviews.llvm.org/D103424
This patch was split from https://reviews.llvm.org/D102246
[SampleFDO] New hierarchical discriminator for Flow Sensitive SampleFDO
This is for llvm-profdata part of change. It sets the bit masks for the
profile reader in llvm-profdata. Also add an internal option
"-fs-discriminator-pass" for show and merge command to process the profile
offline.
This patch also moved setDiscriminatorMaskedBitFrom() to
SampleProfileReader::create() to simplify the interface.
Differential Revision: https://reviews.llvm.org/D103550
If we ended up with two phi instructions in a block, and we needed to fix up
the banks for the first one, we'd end up inserting our COPY before the second
phi.
E.g.
```
%x = G_PHI ...
%fixup = COPY ...
%y = G_PHI ...
```
This is invalid MIR, and breaks assumptions made by the register allocator later
down the line. With the verifier enabled, it also emits a verification error.
This teaches fixupPHIOpBanks to walk past any phi instructions in the block
when emitting the fixup copies.
Here's an example of the crashing code (same as added testcase):
https://godbolt.org/z/h5j1x3o6e
Differential Revision: https://reviews.llvm.org/D103582
This patch sets the isCommutable attribute for several opcodes that have
the "reg = OPCODE reg, reg" format.
Differential Revision: https://reviews.llvm.org/D103653
This patch changes the `isKnownHeapToStack` and `isAssumedHeapToStack`
member functions to return if a function call is going to be altered by
HeapToStack.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D103574
This patch adds an option to `lookupAAFor` that allows it to return a
nullptr if the state of the looked up attribute is invalid. This is so
future passes can use this to query other attributes with the guarantee
that they are valid.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D103556
All that really matters is that the VLMAX of the preceding
instructions is the same as the VLMAX required by the mask
operation.
Also update the vmsge(u) handling to use the SEW/LMUL we use for
other mask register operations. We were matching it to the compare
before. Some cases will be improve if we fix masked compares to
use tail agnostic policy. I think they ignore the tail policy
anyway.
Reviewed By: frasercrmck
Differential Revision: https://reviews.llvm.org/D103299
Need to emit a call for __kmpc_cancel_barrier in the exit block for
__kmpc_cancel function call if cancellation of the parallel block is
requested.
Differential Revision: https://reviews.llvm.org/D103646
setcc (csel 0, 1, cond, X), 1, ne ==> csel 0, 1, !cond, X
Where X is a condition code setting instruction.
Co-authored-by: Paul Walker <paul.walker@arm.com>
Differential Revision: https://reviews.llvm.org/D103256
Due to the dependency on runtime unrolling, UnJ is only
enabled by default on in-order scheduling models,
and if a cpu is specified through -mcpu.
Differential Revision: https://reviews.llvm.org/D103604
When using and ACLE intrinsic for an SVE2 shift, if the predicate passed
has all relevant lanes active, then use a reversed version of the
instruction if beneficial.
This is similar to b865eead7657 ( D103617 ) and fixes:
https://llvm.org/PR50575
41b71f718b94c6f12b did this and more (noted with TODO
comments in the tests), but it didn't handle the case
where the destination is narrower than the source, so
it got reverted.
This is a simple match-and-replace. If there's evidence
that the TODO cases are useful, we can revisit/extend.
Was reverted in 0507fc2ffc9, in phi-coalesce-subreg.mir I'd explicitly named
some passes to run instead of specifying a range. As a result some
two-address-instrs weren't correctly rewritten and the verifier got upset.
Original commit message:
[DebugInstrRef][2/3] Track PHI values through register coalescing
In the instruction referencing variable location model, we store variable
locations that point at PHIs in MachineFunction during register allocation.
Unfortunately, register coalescing can substantially change the locations
of registers, and so that PHI-variable-location side table needs
maintenence during the pass.
This patch builds an index from the side table, and whenever a vreg gets
coalesced into another vreg, update the index to record the new vreg that
the PHI happens in. It also accepts a limited range of subregister
coalescing, for example merging a subregister into a larger class.
Differential Revision: https://reviews.llvm.org/D86813
The CHECK-NOT directives in
tools/llvm-dwarfdump/X86/parent_recurse_depth.s can accidentally match
something in the path of the object file created by yaml2obj, for
example:
llvm-project/llvm/test/tools/llvm-dwarfdump/X86/parent_recurse_depth.s:13:12:
error: ONE-NOT: excluded string found in input
^
<stdin>:1:22: note: found here
builds/llvm-projects/mainline/release/test/tools/llvm-dwarfdump/X86/Output/parent_recurse_depth.s.tmp.o: file format elf64-x86-64
^~~~
This commit alleviate this issue by consuming the file name from the
output, forcing all the CHECK-NOT to match what comes after.
Reviewed By: Higuoxing
Differential Revision: https://reviews.llvm.org/D103676
This patch extends the SelectionDAG's ability to constant-fold vector
arithmetic to include support for SPLAT_VECTOR. This is not only for
scalable-vector types but also for fixed-length vector types, which
helps Hexagon in a couple of cases.
The original RISC-V test case was in fact an infinite DAGCombine loop.
The pattern `and (truncate v1), (truncate v2)` can be combined to
`truncate (and v1, v2)` but the truncate can similarly be combined back
to `truncate (and v1, v2)` (but, crucially, only when one of `v1` or
`v2` is a constant vector).
It wasn't exposed in on fixed-length types because a TRUNCATE of a
constant BUILD_VECTOR was folded into the BUILD_VECTOR itself, whereas
this did not happen for the equivalent (scalable-vector) SPLAT_VECTOR.
Reviewed By: RKSimon, craig.topper
Differential Revision: https://reviews.llvm.org/D103246
Summary: When -strict-dwarf=true is specified, the calling convention info
DW_CC_pass_by_value or DW_CC_pass_by_reference can only be generated at DWARF5.
Reviewed By: shchenz, dblaikie
Differential Revision: https://reviews.llvm.org/D103300
LLDB is currently not selected in LLVM release testing and thus it
doesnt make its way into prebuilt binaries which build with default
configuration. This patch enables LLDB by default in test-release
script.
Assuming LLDB build by default was disabled back in 2016 LLDB support
for various architectures has a long way since then. It has buildbots
for most architectures and supports a case to be included by default.
Also lldb build can easily be disabled in case some release managers
choose to do so.
Reviewed By: tstellar
Differential Revision: https://reviews.llvm.org/D101864
Don't propagate launch bound related attributes to
address taken functions and their callees. The idea
is to do a traversal over the call graph starting at
address taken functions and erase the attributes
set by previous logic i.e. process().
This two phase approach makes sure that we don't
miss out on deep nested callees from address taken
functions as a function might be called directly as
well as indirectly.
This patch is also reattempt to D94585 as latent issues
are fixed in hasAddressTaken function in the recent
past.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D103138
Before packing LDS globals into a sorted structure, make sure that
their alignment is properly updated based on their size. This will make
sure that the members of sorted structure are properly aligned, and
hence it will further reduce the probability of unaligned LDS access.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D103261
Make extended binary the default output format for CSSPGO. This avoids having to pass flag every time when generating profile. It also matches llvm-profdata where binary profile is the default (should we switch to extbinary as default for llvm-profdata?).
We plan to compress name table for context profile, which depends on the built-in compression of extbinary.
Differential Revision: https://reviews.llvm.org/D103650
Some floating point lib calls have ABI attributes that need to be set on
the caller. Found via D103412.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D103415
In a `-DCMAKE_BUILD_TYPE=Release -DLLVM_TARGETS_TO_BUILD=X86 -DLLVM_LINK_LLVM_DYLIB=on -DCLANG_LINK_CLANG_DYLIB=on`
build, libLLVM-13git.so is 2% smaller and libclang-cpp.so is 1% smaller (on top of -Wl,-Bsymbolic-functions).
There may be some small performance improvement as well because GCC
-fPIC suppresses interprocedural optimizations for non-inline
definitions by default.
Note: we cannot add -fno-semantic-interposition for Clang<13. Clang<13's
implementation additionally optimizes global variables, which is incompatible
with unfortunate ELF -fno-pic default: direct access relocations for external
data. If the executable has a -fno-pic object file referencing a global variable
declared in a public header, the direct access relocation will cause a copy
relocation. The executable and libLLVM.so/libclang-cpp.so will disagree on the
address.
Reviewed By: phosek
Differential Revision: https://reviews.llvm.org/D102453
Building LLVM with the LLVM_ENABLE_MODULES cmake option fails when the modules are being compiled due to missing includes. This is a side effect of some transitive includes that changed recently.
Differential Revision: https://reviews.llvm.org/D103645
This builds on D103584. The change eliminates the coupling between unroll heuristic and implementation w.r.t. knowing when the passed in trip count is an exact trip count or a max trip count. In theory the new code is slightly less powerful (since it relies on exact computable trip counts), but in practice, it appears to cover all the same cases. It can also be extended if needed.
The test change shows what appears to be a bug in the existing code around the interaction of peeling and unrolling. The original loop only ran 8 iterations. The previous output had the loop peeled by 2, and then an exact unroll of 8. This meant the loop ran a total of 10 iterations which appears to have been a miscompile.
Differential Revision: https://reviews.llvm.org/D103620
