HSA runtime fails to find the symbols for Init and Fini kernels as
they mark with internal linkage, changing the linkage to external
to fix those errors.
Differential Revision: https://reviews.llvm.org/D110054
New mode option that allows for either running the default fusion kind that happens today or doing either of producer-consumer or sibling fusion. This will also be helpful to minimize the compile-time of the fusion tests.
Reviewed By: bondhugula, dcaballe
Differential Revision: https://reviews.llvm.org/D110102
This patch fixes the pattern for the P10 instructions Vector Shift Left
Double by Bit Immediate VN-form and Vector Shift Right Double by Bit
Immediate VN-form. The third argument should be a target constant (`timm`)
instead of an `i32` because an immediate is expected.
Reviewed By: lei
Differential Revision: https://reviews.llvm.org/D109920
HIP currently uses -mlink-builtin-bitcode to link all bitcode libraries, which
changes the linkage of functions to be internal once they are linked in. This
works for common bitcode libraries since these functions are not intended
to be exposed for external callers.
However, the functions in the sanitizer bitcode library is intended to be
called by instructions generated by the sanitizer pass. If their linkage is
changed to internal, their parameters may be altered by optimizations before
the sanitizer pass, which renders them unusable by the sanitizer pass.
To fix this issue, HIP toolchain links the sanitizer bitcode library with
-mlink-bitcode-file, which does not change the linkage.
A struct BitCodeLibraryInfo is introduced in ToolChain as a generic
approach to pass the bitcode library information between ToolChain and Tool.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D110304
A defined assignment subroutine invoked in the context of a WHERE
statement or construct must necessarily be elemental (C1032).
Differential Revision: https://reviews.llvm.org/D109932
This can avoid a loss of decoupling with the scalar unit on cores
with decoupled scalar and vector units.
We should support FP too, but those use extract_element and not a
custom ISD node so it is a little different. I also left a FIXME
in the test for i64 extract and store on RV32.
Reviewed By: frasercrmck
Differential Revision: https://reviews.llvm.org/D109482
The sparse constant provides a constant tensor in coordinate format. We first split the sparse constant into a constant tensor for indices and a constant tensor for values. We then generate a loop to fill a sparse tensor in coordinate format using the tensors for the indices and the values. Finally, we convert the sparse tensor in coordinate format to the destination sparse tensor format.
Add tests.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D110373
The minor code refactorization introduces the TASK_TIED constant inside
kmp_gsupprot.cpp as a replacement for the literal value 1.
The mentioned constant is now used in both kmp_tasking.cpp and
kmp_gsupport.cpp files.
Differential Revision: https://reviews.llvm.org/D110441
If one input of a fixed vector multiply is a sign/zero extend and
the other operand is a splat of a scalar, we can use a widening
multiply if the scalar value has sufficient sign/zero bits.
Reviewed By: frasercrmck
Differential Revision: https://reviews.llvm.org/D110028
This is no-functional-change-intended, but it hopefully makes things
slightly clearer and more efficient to have transforms that require
'shl' be called only from visitShl(). Further cleanup is possible.
Note that getTheLanaiTarget is declared in
TargetInfo/LanaiTargetInfo.h, which LanaiDisassembler.cpp includes.
Identified with readability-redundant-declaration.
This path defines the newly added `__kmpc_disitrute_static_init`
functions in the device runtime library. These functions are currently
exact copies of the current worksharing method but can be tuned later.
Depends on D110429
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D110430
This patch adds a new RTL function for worksharing. Currently we use
`__kmpc_for_static_init` for both the `distribute` and `parallel`
portion of the loop clause. This patch replaces the `distribute` portion
with a new runtime call `__kmpc_distribute_static_init`. Currently this
will be used exactly the same way, but will make it easier in the future
to fine-tune the distribute and parallel portion of the loop.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D110429
Apparently macOS is padding the name result with several padding zeroes at
the end. Just strip them all to pretend it's a C-string.
Thanks to Pavel for suggesting this fix.
The tests only specify -march, so when the tests are run on AIX the target OS defaults to AIX, which causes the tests to misbehave.
This patch constrains the tests by specifying -mtriple instead of -march.
Reviewed By: daltenty, jsji, MaskRay
Differential Revision: https://reviews.llvm.org/D110186
This is another step towards trying to re-apply D110170
by eliminating conflicting transforms that cause infinite loops.
a47c8e40c7 was a previous patch in this direction.
The diffs here are mostly cosmetic, but intentional:
1. The existing code that would handle this pattern in FoldShiftByConstant()
is limited to 'shl' only now. The formatting change to IsLeftShift shows
that we could move several transforms into visitShl() directly for
efficiency because they are not common shift transforms.
2. The tests are regenerated to show new instruction names to prove that
we are getting (almost) identical logic results.
3. The one case where we differ ("trunc_sandwich_small_shift1") shows that
we now use a narrow 'and' instruction. Previously, we relied on another
transform to do that, but it is limited to legal types. That seems to
be a legacy constraint from when IR analysis and codegen were less robust.
https://alive2.llvm.org/ce/z/JxyGA4
declare void @llvm.assume(i1)
define i8 @src(i32 %x, i32 %c0, i8 %c1) {
; The sum of the shifts must not overflow the source width.
%z1 = zext i8 %c1 to i32
%sum = add i32 %c0, %z1
%ov = icmp ult i32 %sum, 32
call void @llvm.assume(i1 %ov)
%sh1 = lshr i32 %x, %c0
%tr = trunc i32 %sh1 to i8
%sh2 = lshr i8 %tr, %c1
ret i8 %sh2
}
define i8 @tgt(i32 %x, i32 %c0, i8 %c1) {
%z1 = zext i8 %c1 to i32
%sum = add i32 %c0, %z1
%maskc = lshr i8 -1, %c1
%s = lshr i32 %x, %sum
%t = trunc i32 %s to i8
%a = and i8 %t, %maskc
ret i8 %a
}
It appears that this test assumes that the toolchain utilizes the integrated
assembler by default, since the expected output in the CHECKs are
compilation_database.o.
However, this test fails on AIX as AIX does not utilize the integrated assembler.
On AIX, the output instead is of the form /tmp/compilation_database-*.s.
Thus, this patch explicitly adds the -fintegrated-as option to match the
assumption that the integrated assembler is used by default.
Differential Revision: https://reviews.llvm.org/D110431
The trace tests crashed on darwin because of some thread
initialization issues (thread initialization is somewhat
different on darwin).
Instead of starting real threads, create a new ThreadState
in the main thread. This makes the tests more unit-testy
and hopefully won't crash on darwin (there is almost no
platform-specific code involved now).
This will also help with future trace tests that will need
more than 1 thread. Creating more than 1 real thread and
dispatching test actions across multiple threads in the
required deterministic order is painful.
Depends on D110539.
Reviewed By: melver
Differential Revision: https://reviews.llvm.org/D110546
Currently detection of races with TLS/stack initialization
is broken because we imitate the write before thread initialization,
so it's modelled with a wrong thread/epoch.
Fix that and add a test.
Reviewed By: melver
Differential Revision: https://reviews.llvm.org/D110538
KILL instructions are sometimes present and prevented hard
clauses from being formed.
Fix this by ignoring all meta instructions in clauses.
Differential Revision: https://reviews.llvm.org/D106042
We used to put the canonical spelling of flags after alias processing
on that line. For clang-cl in particular, that meant that we put flags
on that line that the clang-cl driver doesn't even accept, and the
"Driver args:" line wasn't usable.
Differential Revision: https://reviews.llvm.org/D110458
Add a convenience method to add supplementary registers that takes care
of adding invalidate_regs to all (potentially) overlapping registers.
Differential Revision: https://reviews.llvm.org/D110023
Function specialization was crashing on poison values and constexpr values.
The problem is that these values are not added to the solver, so it crashes
when a lookup is performed for these values. This fixes that by not
specialising on these values. For poison that is obvious, but for constexpr
this is a change in behaviour. Thus, in one way this is a bit of a stopgap, but
specialising on constexpr values wasn't done very intentionally, and need some
more work and tests if we wanted to support this.
As a follow up, we need to look if the solver should exit more gracefully and
return a "don't know", or that it should really support these constexprs.
This should fix PR51600 (https://bugs.llvm.org/show_bug.cgi?id=51600).
Differential Revision: https://reviews.llvm.org/D110529
This change is to add some missing details to the help text and command
guide:
- Added a note to the command guide that --debug-macro also dumps
.debug_macinfo.
- Added a note to the command guide that --debug-frame and --eh_frame
are aliases, and in cases where both sections are present one command
outputs both.
- Changed the wording in the help output for --ignore-case and --regex to
closer match the command guide.
The StringConvert API is no longer used anywhere but in debugserver.
Since debugserver does not use LLVM API, we cannot replace it with
llvm::to_integer() and llvm::to_float() there. Let's just move
the sources into debugserver.
Differential Revision: https://reviews.llvm.org/D110478
Refactor the XML converting attribute and text getters to use LLVM API.
While at it, remove some redundant error and missing XML support
handling, as the called base functions do that anyway. Add tests
for these methods.
Note that this patch changes the getter behavior to be IMHO more
correct. In particular:
- negative and overflowing integers are now reported as failures to
convert, rather than being wrapped over or capped
- digits followed by text are now reported as failures to convert
to double, rather than their numeric part being converted
Differential Revision: https://reviews.llvm.org/D110410
Use the in-project clang, llvm-link and opt if available and unless
CMake cache variables specify to use a different compiler. This applies
D101265 to the new DeviceRTL's CMakeLists.txt which was copied before
D101265 was applied.
Fixes the openmp-offloading-cuda-runtime builder which was failing
since D110006.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D110251
Let the calling pass or pattern replace the uses of the original root operation. Internally, the tileAndFuse still replaces uses and updates operands but only of newly created operations.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D110169
When rebase_exec=true in DidAttach(), all modules are loaded
before the rendezvous breakpoint is set, which means the
LoadInterpreterModule() method is not called and m_interpreter_module
is not initialized.
This causes the very first rendezvous breakpoint hit with
m_initial_modules_added=false to accidentally unload the
module_sp that corresponds to the dynamic loader.
This bug (introduced in D92187) was causing the rendezvous
mechanism to not work in Android 28. The mechanism works
fine on older/newer versions of Android.
Test: Verified rendezvous on Android 28 and 29
Test: Added dlopen test
Reviewed By: labath
Differential Revision: https://reviews.llvm.org/D109797