If amdgpu-unsafe-fp-atomics is specified, allow {flat|global}_atomic_add_f32 even if atomic modes don't match.
Differential Revision: https://reviews.llvm.org/D95391
It was discussed a few years ago and agreed that it makes sense to
remove this assertion as other targets do not perform similar register
size checking in inline assembly constraint logic, so the check just
adds a needless barrier on AVR.
This patch removes the assertion and removes 'XFAIL' from two Generic
CodeGen tests for AVR as a result.
This makes sure that AddClang.cmake is installed alongside other Clang
CMake modules. This mirrors LLVM and MLIR in this respect and is
required when building the new Flang driver out of tree (as it depends
on Clang and includes AddClang.cmake).
Reviewed By: bogner
Differential Revision: https://reviews.llvm.org/D94533
Add support for the following layout options:
* -ffree-form
* -ffixed-form
- -ffixed-line-length=n (alias -ffixed-line-length-n)
Additionally remove options `-fno-free-form` and `-fno-fixed-form` as they were initially added to forward to gfortran but gfortran does not support these flags.
This patch adds the flag FlangOnlyOption to the existing options `-ffixed-form`, `-ffree-form` and `-ffree-line-length-` in Options.td. As of commit 6a75496836, these flags are not currently forwarded to gfortran anyway.
The default fixed line length in FrontendOptions is 72, based off the current default in Fortran::parser::Options. The line length cannot be set to a negative integer, or a positive integer less than 7 excluding 0, consistent with the behaviour of gfortran.
This patch does not add `-ffree-line-length-n` as Fortran::parser::Options does not have a variable for free form columns.
Whilst the `fixedFormColumns` variable is used in f18 for `-ffree-line-length-n`, f18 only allows `-ffree-line-length-none`/`-ffree-line-length-0` and not a user-specified value. `fixedFormcolumns` cannot be used in the new driver as it is ignored in the frontend when dealing with free form files.
Summary of changes:
- Remove -fno-fixed-form and -fno-free-form from Options.td
- Make -ffixed-form, -ffree-form and -ffree-line-length-n FlangOnlyOption in Options.td
- Create AddFortranDialectOptions method in Flang.cpp
- Create FortranForm enum in FrontendOptions.h
- Add fortranForm_ and fixedFormColumns_ to Fortran::frontend::FrontendOptions
- Update fixed-form-test.f so that it guarantees that it fails when forced as a free form file to better facilitate testing.
Differential Revision: https://reviews.llvm.org/D95460
These class methods simply return a new UnivariateLinearPolyBase
(e.g. ElementCount), and do not modify the object in any way or form,
so qualify for being 'const'.
This modified patch avoids redirecting the unit in which a subprogram is
created if type units are enabled -- DIEs were getting children allocated
from different units memory pools. Original commit message:
[DWARF] Create subprogram's DIE in DISubprogram's unit
This is a fix for PR48790. Over in D70350, subprogram DIEs were permitted
to be shared between CUs. However, the creation of a subprogram DIE can be
triggered early, from other CUs. The subprogram definition is then created
in one CU, and when the function is actually emitted children are attached
to the subprogram that expect to be in another CU. This breaks internal CU
references in the children.
Fix this by redirecting the creation of subprogram DIEs in
getOrCreateContextDIE to the CU specified by it's DISubprogram definition.
This ensures that the subprogram DIE is always created in the correct CU.
Differential Revision: https://reviews.llvm.org/D94976
This new f16 shuffle under Neon would hit an assert in
GeneratePerfectShuffle as it would try to treat a f16 vector as an i8.
Add f16 handling, treating them like an i16.
Differential Revision: https://reviews.llvm.org/D95446
Historically, the Vector to LLVM dialect conversion subsumed the Standard to
LLVM dialect conversion patterns. This was necessary because the conversion
infrastructure did not have sufficient support for reconciling type
conversions. This support is now available. Only keep the patterns related to
the Vector dialect in the Vector to LLVM conversion and require type casts
operations to be inserted if necessary. These casts will be removed by
following conversions if possible. Update integration tests to also run the
Standard to LLVM conversion.
There is a significant amount of test churn, which is due to (a) unnecessarily
strict tests in VectorToLLVM and (b) many patterns actually targeting Standard
dialect ops instead of LLVM dialect ops leading to tests actually exercising a
Vector->Standard->LLVM conversion. This churn is a good illustration of the
reason to make the conversion partial: now the tests only check the code in the
Vector to LLVM conversion and will not be randomly broken by changes in
Standard to LLVM conversion.
Arguably, it may be possible to extract Vector to Standard patterns into a
separate pass, but given the ongoing splitting of the Standard dialect, such
pass will be short-lived and will require further refactoring.
Depends On D95626
Reviewed By: nicolasvasilache, aartbik
Differential Revision: https://reviews.llvm.org/D95685
In dialect conversion infrastructure, source materialization applies as part of
the finalization procedure to results of the newly produced operations that
replace previously existing values with values having a different type.
However, such operations may be created to replace operations created in other
patterns. At this point, it is possible that the results of the _original_
operation are still in use and have mismatching types, but the results of the
_intermediate_ operation that performed the type change are not in use leading
to the absence of source materialization. For example,
%0 = dialect.produce : !dialect.A
dialect.use %0 : !dialect.A
can be replaced with
%0 = dialect.other : !dialect.A
%1 = dialect.produce : !dialect.A // replaced, scheduled for removal
dialect.use %1 : !dialect.A
and then with
%0 = dialect.final : !dialect.B
%1 = dialect.other : !dialect.A // replaced, scheduled for removal
%2 = dialect.produce : !dialect.A // replaced, scheduled for removal
dialect.use %2 : !dialect.A
in the same rewriting, but only the %1->%0 replacement is currently considered.
Change the logic in dialect conversion to look up all values that were replaced
by the given value and performing source materialization if any of those values
is still in use with mismatching types. This is performed by computing the
inverse value replacement mapping. This arguably expensive manipulation is
performed only if there were some type-changing replacements. An alternative
could be to consider all replaced operations and not only those that resulted
in type changes, but it would harm pattern-level composability: the pattern
that performed the non-type-changing replacement would have to be made aware of
the type converter in order to call the materialization hook.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D95626
This patch implements generation of remaining header search arguments.
It's done manually in C++ as opposed to TableGen, because we need the flexibility and don't anticipate reuse.
This patch also tests the generation of header search options via a round-trip. This way, the code gets exercised whenever Clang is built and tested in asserts mode. All `check-clang` tests pass.
Reviewed By: dexonsmith
Differential Revision: https://reviews.llvm.org/D94472
As noted in https://reviews.llvm.org/D93459, the formatting of
multi-line descriptions of clEnumValN and the likes is unfavorable.
Thus this patch adds support for correctly indenting these.
Reviewed By: serge-sans-paille
Differential Revision: https://reviews.llvm.org/D93494
When SGPRs are spilled to VGPRs, they can overwrite any lane. We need
to preserve the value of inactive lanes in function calls, so we save
the register even if it is marked as caller saved.
Also, teach buildPrologSpill to work when no registers are free like in
CodeGen/AMDGPU/pei-scavenge-vgpr-spill.mir and update the comment on
findScratchNonCalleeSaveRegister as it is not used anymore to realign
the stack pointer since D95865.
Differential Revision: https://reviews.llvm.org/D95946
This patch allows detecting conflicts with variables defined in the current
CompoundStmt or If/While/For variable init statements.
Reviewed By: hokein
Differential Revision: https://reviews.llvm.org/D95925
This revision defines a Linalg contraction in general terms:
1. Has 2 input and 1 output shapes.
2. Has at least one reduction dimension.
3. Has only projected permutation indexing maps.
4. its body computes `u5(u1(c) + u2(u3(a) * u4(b)))` on some field
(AddOpType, MulOpType), where u1, u2, u3, u4 and u5 represent scalar unary
operations that may change the type (e.g. for mixed-precision).
As a consequence, when vectorization of such an op occurs, the only special
behavior is that the (unique) MulOpType is vectorized into a
`vector.contract`. All other ops are handled in a generic fashion.
In the future, we may wish to allow more input arguments and elementwise and
constant operations that do not involve the reduction dimension(s).
A test is added to demonstrate the proper vectorization of matmul_i8_i8_i32.
Differential revision: https://reviews.llvm.org/D95939
doubly-nested implicit CXXConstructExprs.
Ensure that we transform the parameter initializer using
TransformInitializer rather than TransformExpr so that we properly strip
down and rebuild the initialization, including any necessary
CXXBindTemporaryExprs. Otherwise we can end up forgetting to destroy
temporary objects used to construct a constructor parameter.
This separation improves the layering and paves the way for more interfaces coming up in the future.
Differential revision: https://reviews.llvm.org/D95941
- On Windows, extended lambda has extra issues due to the numbering
schemes are different between the host compilation (Microsoft C++ ABI)
and the device compilation (Itanium C++ ABI. Additional device side
lambda number is required per lambda for the host compilation to
correctly mangle the device-side lambda name.
- A hybrid numbering context `MSHIPNumberingContext` is introduced to
number a lambda for both host- and device-compilations.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D69322
This change compresses the context string by removing cycles due to recursive function for CS profile generation. Removing recursion cycles is a way to normalize the calling context which will be better for the sample aggregation and also make the context promoting deterministic.
Specifically for implementation, we recognize adjacent repeated frames as cycles and deduplicated them through multiple round of iteration.
For example:
Considering a input context string stack:
[“a”, “a”, “b”, “c”, “a”, “b”, “c”, “b”, “c”, “d”]
For first iteration,, it removed all adjacent repeated frames of size 1:
[“a”, “b”, “c”, “a”, “b”, “c”, “b”, “c”, “d”]
For second iteration, it removed all adjacent repeated frames of size 2:
[“a”, “b”, “c”, “a”, “b”, “c”, “d”]
So in the end, we get compressed output:
[“a”, “b”, “c”, “d”]
Compression will be called in two place: one for sample's context key right after unwinding, one is for the eventual context string id in the ProfileGenerator.
Added a switch `compress-recursion` to control the size of duplicated frames, default -1 means no size limit.
Added unit tests and regression test for this.
Differential Revision: https://reviews.llvm.org/D93556
A module in the cache with an error should just be a cache miss. If
allowing errors (with -fallow-pcm-with-compiler-errors), a rebuild is
needed so that the appropriate diagnostics are output and in case search
paths have changed. If not allowing errors, the module was built
*allowing* errors and thus should be rebuilt regardless.
Reviewed By: akyrtzi
Differential Revision: https://reviews.llvm.org/D95989
For CS profile generation, the process of call stack unwinding is time-consuming since for each LBR entry we need linear time to generate the context( hash, compression, string concatenation). This change speeds up this by grouping all the call frame within one LBR sample into a trie and aggregating the result(sample counter) on it, deferring the context compression and string generation to the end of unwinding.
Specifically, it uses `StackLeaf` as the top frame on the stack and manipulates(pop or push a trie node) it dynamically during virtual unwinding so that the raw sample can just be recoded on the leaf node, the path(root to leaf) will represent its calling context. In the end, it traverses the trie and generates the context on the fly.
Results:
Our internal branch shows about 5X speed-up on some large workloads in SPEC06 benchmark.
Differential Revision: https://reviews.llvm.org/D94110
This change compresses the context string by removing cycles due to recursive function for CS profile generation. Removing recursion cycles is a way to normalize the calling context which will be better for the sample aggregation and also make the context promoting deterministic.
Specifically for implementation, we recognize adjacent repeated frames as cycles and deduplicated them through multiple round of iteration.
For example:
Considering a input context string stack:
[“a”, “a”, “b”, “c”, “a”, “b”, “c”, “b”, “c”, “d”]
For first iteration,, it removed all adjacent repeated frames of size 1:
[“a”, “b”, “c”, “a”, “b”, “c”, “b”, “c”, “d”]
For second iteration, it removed all adjacent repeated frames of size 2:
[“a”, “b”, “c”, “a”, “b”, “c”, “d”]
So in the end, we get compressed output:
[“a”, “b”, “c”, “d”]
Compression will be called in two place: one for sample's context key right after unwinding, one is for the eventual context string id in the ProfileGenerator.
Added a switch `compress-recursion` to control the size of duplicated frames, default -1 means no size limit.
Added unit tests and regression test for this.
Differential Revision: https://reviews.llvm.org/D93556
We could extend this with an interface to allow dialect to perform a type
conversion, but that would make the folder creating operation which isn't
the case at the moment, and isn't necessarily always desirable.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D95991
The collapseLoops method implements a transformations facilitating the implementation of the collapse-clause. It takes a list of loops from a loop nest and reduces it to a single loop that can be used by other methods that are implemented on just a single loop, such as createStaticWorkshareLoop.
This patch shares some changes with D92974 (such as adding some getters to CanonicalLoopNest), used by both patches.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D93268
I switched the watch simulator test from i386 to using x86_64, but
apparently that's not supported on the bots. Rollback to using i386 and
solve the original issue by passing the target, similar to what I did
in TestSimulatorPlatform.py.
This change implements profile generation infra for pseudo probe in llvm-profgen. During virtual unwinding, the raw profile is extracted into range counter and branch counter and aggregated to sample counter map indexed by the call stack context. This change introduces the last step and produces the eventual profile. Specifically, the body of function sample is recorded by going through each probe among the range and callsite target sample is recorded by extracting the callsite probe from branch's source.
Please refer https://groups.google.com/g/llvm-dev/c/1p1rdYbL93s and https://reviews.llvm.org/D89707 for more context about CSSPGO and llvm-profgen.
**Implementation**
- Extended `PseudoProbeProfileGenerator` for pseudo probe based profile generation.
- `populateBodySamplesWithProbes` reading range counter is responsible for recording function body samples and inferring caller's body samples.
- `populateBoundarySamplesWithProbes` reading branch counter is responsible for recording call site target samples.
- Each sample is recorded with its calling context(named `ContextId`). Remind that the probe based context key doesn't include the leaf frame probe info, so the `ContextId` string is created from two part: one from the probe stack strings' concatenation and other one from the leaf frame probe.
- Added regression test
Test Plan:
ninja & ninja check-llvm
Differential Revision: https://reviews.llvm.org/D92998
Similar to the G_PTR_ADD + G_LOAD twiddling we do in `preISelLower`.
The imported patterns expect scalars only, so they can't handle things like
```
G_STORE %ptr1, %ptr2
```
To get around this, use s64 instead.
(This probably makes a good portion of the manual selection code for G_STORE
dead.)
This is a 0.2% geomean code size improvement on CTMark at -Os.
(Best is consumer-typeset @ -0.7%)
Differential Revision: https://reviews.llvm.org/D95908
When we have a zeroext parameter coming in on the stack, build
```
%x = G_LOAD ...
%x_assert_zext = G_ASSERT_ZEXT %x, narrow_size
%trunc = G_TRUNC %x_assert_zext
```
Rather than just loading into the truncated type.
This allows us to optimize cases like this: https://godbolt.org/z/vfjhW8
Differential Revision: https://reviews.llvm.org/D95805
Konstantin Zhuravlyov