directives.
If the default datasharing is set to none, the datasharing attributes
for variables in the condition of the if clause for the inner taskloop
directive must be verified.
According to OpenMP 5.0, if clause can be used in for simd directive. If
condition in the if clause if false, the non-vectorized version of the
loop must be executed.
According to OpenMP 5.0, if clause can be used in simd directive. If
condition in the if clause if false, the non-vectorized version of the
loop must be executed.
Summary:
For the extended defaultmap, most of the work is inside sema.
The only difference for codegen is to set different initial
maptype for different implicit-behavior.
Reviewers: jdoerfert, ABataev
Reviewed By: ABataev
Subscribers: dreachem, sandoval, cfe-commits
Tags: #clang, #openmp
Differential Revision: https://reviews.llvm.org/D69204
simd-based directives.
According to OpenMP 5.0 standard, ordered simd, atomic and simd
directives are allowed as nested directives in the simd-based
directives.
If the context selector score was not specified, its value must be set
to 0. Simplify the processing of unspecified scores + save memory in
attribute representation.
Need to analyze inner target regions in case of implicit mapping of the
data members when target region is created in one of the class member
functions.
The expression of the priority clause must be captured in the combined
task-based directives, like 'parallel master taskloop' directive.
llvm-svn: 375026
The expression of the num_tasks clause must be captured in the combined
task-based directives, like 'parallel master taskloop' directive.
llvm-svn: 374819
The expression of the grainsize clause must be captured in the combined
task-based directives, like 'parallel master taskloop' directive.
llvm-svn: 374810
Added parsing/sema/codegen support for 'parallel master taskloop'
constructs. Some of the clauses, like 'grainsize', 'num_tasks', 'final'
and 'priority' are not supported in full, only constant expressions can
be used currently in these clauses.
llvm-svn: 374791
constructs.
If OpenMP construct includes several capturing regions and the variable
is declared as private, the length of the inner variable length array is
not captured in outer captured regions, only in the innermost region.
Patch fixes this bug.
llvm-svn: 374787
CUDA/HIP program may be compiled with -fopenmp. In this case, -fopenmp is only passed to host compilation
to take advantages of multi-threads computation.
CUDA/HIP and OpenMP both use Sema::DeviceCallGraph to store functions to be analyzed and remove them
once they decide the function is sure to be emitted. CUDA/HIP and OpenMP have different functions to determine
if a function is sure to be emitted.
To check host/device correctly for CUDA/HIP when -fopenmp is enabled, there needs a unified logic to determine
whether a function is to be emitted. The logic needs to be aware of both CUDA and OpenMP logic.
Differential Revision: https://reviews.llvm.org/D67837
llvm-svn: 374263
and of vendors, not or.
If several vendors are provided in the same vendor context trait, the
context shall match only if all vendors are matching, not one of them.
This is per OpenMP 5.0, 2.3.3 Matching and Scoring Context Selectors,
all selectors in the construct, device, and implementation sets of the
context selector appear in the corresponding trait set of the OpenMP
context.
llvm-svn: 374107
C linkage.
After some discussion with OpenMP developers, it was decided that the
functions with the different C linkage can be used in declare variant
directive.
llvm-svn: 374057
We previously failed to treat an array with an instantiation-dependent
but not value-dependent bound as being an instantiation-dependent type.
We now track the array bound expression as part of a constant array type
if it's an instantiation-dependent expression.
llvm-svn: 373685
Context selectors may include optional score clause in format
`score(<expr>):`, where `<expr>` must be a constant integer expression.
Added parsing/sema analysis only.
llvm-svn: 373502
defined body.
If the function is currently defined, we should not emit a warning that
it might be emitted already because it was not really emitted.
llvm-svn: 373243
This patches fixes the case when a user-defined mapper is attached to
the elements of an array, and to report error when a mapper is used for
types other than struct, class, and union.
Patch by Lingda Li <lildmh@gmail.com>
Differential Revision: https://reviews.llvm.org/D67978
llvm-svn: 373023
sets.
According to OpenMP 5.0, context selector set might include several
context selectors, separated with commas. Patch fixes this problem.
llvm-svn: 372235
Added attribute for declare variant directive. It will allow to handle
declare variant directive at the codegen and will allow to add extra
checks.
llvm-svn: 372147
Need to return original declaration group with FunctionTemplateDecl, not
the inner FunctionDecl, to correctly handle parsing of directives with
the templates parameters.
llvm-svn: 372011
In order to enable future improvements to our attribute diagnostics,
this moves info from ParsedAttr into CommonAttributeInfo, then makes
this type the base of the *Attr and ParsedAttr types. Quite a bit of
refactoring took place, including removing a bunch of redundant Spelling
Index propogation.
Differential Revision: https://reviews.llvm.org/D67368
llvm-svn: 371875
Chun Chen.
The previous patch (https://reviews.llvm.org/D54441) support the
relational-op != very well for openmp canonical loop form, however,
it didn't update the diagnosis message. So this patch is simply
update the diagnosis message by adding !=, update the test
related to it, and update the section number for canonical loop
form for OpenMP 5.0 in comment.
Differential Revision: https://reviews.llvm.org/D66559
llvm-svn: 371631
construct.
OpenMP 5.0 introduced new clause for declare target directive, device_type clause, which may accept values host, nohost, and any. Host means
that the function must be emitted only for the host, nohost - only for
the device, and any - for both, device and the host.
llvm-svn: 369775
For `map`, the following restriction changed in OpenMP 5.0:
* OpenMP 4.5 [2.15.5.1, Restrictions]: "A list item cannot appear in
both a map clause and a data-sharing attribute clause on the same
construct.
* OpenMP 5.0 [2.19.7.1, Restrictions]: "A list item cannot appear in
both a map clause and a data-sharing attribute clause on the same
construct unless the construct is a combined construct."
This patch removes this restriction in the case of combined constructs
and OpenMP 5.0, and it updates Sema not to capture a scalar by copy in
the target region when `firstprivate` and `map` appear for that scalar
on a combined target construct.
This patch also adds a fixme to a test that now reveals that a
diagnostic about loop iteration variables is dropped in the case of
OpenMP 5.0. That bug exists regardless of this patch's changes.
Reviewed By: ABataev, jdoerfert, hfinkel, kkwli0
Differential Revision: https://reviews.llvm.org/D65835
llvm-svn: 369619
If the function is marked as declare target in a standalone directive,
the delayed diagnostics is not emitted. Patch fixes this problem.
llvm-svn: 369432
target.
According to OpenMP 5.0, if a lambda declaration and definition appears between a declare target directive and the matching end declare target directive, all variables that are captured by the lambda expression must also appear in a to clause.
llvm-svn: 369146
Now that we've moved to C++14, we no longer need the llvm::make_unique
implementation from STLExtras.h. This patch is a mechanical replacement
of (hopefully) all the llvm::make_unique instances across the monorepo.
Differential revision: https://reviews.llvm.org/D66259
llvm-svn: 368942
Added basic support for non-rectangular loops. It requires an additional
analysis of min/max boundaries for non-rectangular loops. Since only
linear dependency is allowed, we can do this analysis.
llvm-svn: 368903
Summary:
Added support for basic analysis of the linear variables and linear step
expression. Linear loop iteration variables must be excluded from this
analysis, only non-loop iteration variables must be analyzed.
Reviewers: NoQ
Subscribers: guansong, cfe-commits, caomhin, kkwli0
Tags: #clang
Differential Revision: https://reviews.llvm.org/D65461
llvm-svn: 368295
If the threadprivate variable is used in the copyin clause on inner
parallel directive with TLS support, we capture this variable in all
outer OpenMP scopes. It leads to the fact that in all scopes we're
working with the original variable, not the threadprivate copies.
llvm-svn: 366483
variables.
Loop control variables are private in loop-based constructs and we shall
take this into account when generate the code for inner constructs.
Currently, those variables are reported as shared in many cases. Moved
the analysis of the data-sharing attributes of the loop control variable
to an early semantic stage to correctly handle their attributes.
llvm-svn: 366474
checkDecl is only valid for VarDecls or FieldDecls, since getCanonicalDecl
expects only these. Prevent other Decl kinds (such as CXXMethodDecls and
EnumConstantDecls) from entering and asserting.
Differential Revision: https://reviews.llvm.org/D64842
llvm-svn: 366336
Fixed the processing of the unsupported VLAs in the reduction clauses.
Used targetDiag if the diagnostics can be delayed and emit it
immediately if the target does not support VLAs and we're parsing target
directive with the reduction clauses.
llvm-svn: 365821
Previously, lambda captures were processed in the function called during
capturing the variables. It leads to the recursive functions calls and
may result in the compiler crash.
llvm-svn: 364820
If the variable is used in the OpenMP region implicitly, we need to
check the data-sharing attributes for such variables and generate
implicit clauses for them. Patch improves analysis of such variables for
better handling of data-sharing rules.
llvm-svn: 364683
Fixed handling of the data-sharing attributes for static members when
requesting top most attribute. Previously, it might return the incorrect
attributes for static members if they were overriden in the outer
constructs.
llvm-svn: 364655
According to the OpenMP 5.0 standard, the loop iteration variable in the associated
for-loop of a simd construct with just one associated for-loop may be
listed in a private, lastprivate, or linear clause with a linear-step
that is the increment of the associated for-loop. Also, the loop
teration variables in the associated for-loops of a simd construct with
multiple associated for-loops may be listed in a private or lastprivate
clause.
llvm-svn: 364650
The errors for incorrectly specified data-sharing attributes for simd
constructs must be emitted only for the explicitly provided clauses, not
the predetermined ones.
llvm-svn: 364647
The device code must use the same long double type as the host.
Otherwise the code cannot be linked and executed properly. Patch adds
only basic support and checks for supporting of the host long double
double on the device.
llvm-svn: 363717
Summary: This patch avoids the emission of maps for target link variables when unified memory is present.
Reviewers: ABataev, caomhin
Reviewed By: ABataev
Subscribers: guansong, jdoerfert, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D60883
llvm-svn: 363435
the captured region scope.
This removes a case where we would build expressions (and mark
declarations odr-used) in the wrong scope.
Remove the now-unused 'capture initializer' field on sema::Capture
(except for 'this' captures, which still need to be cleaned up).
No functionality change intended (except that we now very slightly more
precisely determine whether we need to use a capture or not when another
captured region encloses an OpenMP captured region).
llvm-svn: 362179
Instead of duplicating access to the directive stack throughout
SemaOpenMP.cpp, consolidate it to a few methods and call those
everywhere else. In passing, simplify adjacent code where possible.
No functionality change intended.
llvm-svn: 362172
Summary:
Constant evaluator does not work on value-dependent or type-dependent
expressions.
Also fixed bugs uncovered by these assertions.
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D61522
llvm-svn: 361050
If the combined directive has default(none) clause and has clauses for
inner directive that reference some variables, for which data-sharing
attributes are not specified, the error messages should be emitted for
such variables.
llvm-svn: 360365
If the default(none) was specified for the construct, we might miss
diagnostic for the globals without explicitly specified data-sharing
attributes. Patch fixes this problem.
llvm-svn: 360362
This has introduced (exposed?) a crash in clang sema,
that does not happen without this patch.
I'll followup in the original bugreport and commit with reproducer.
This reverts commit r360061.
llvm-svn: 360327
This implementation isn't sound as per the standard.
It erroneously diagnoses e.g. the following case:
```
$ cat test.cpp
void f(int n) {
#pragma omp parallel default(none) if(n)
;
}
```
```
$ ./bin/clang -fopenmp test.cpp
test.cpp:2:40: error: variable 'n' must have explicitly specified data sharing attributes
#pragma omp parallel default(none) if(n)
^
test.cpp:2:31: note: explicit data sharing attribute requested here
#pragma omp parallel default(none) if(n)
^
1 error generated.
```
As per OpenMP Application Programming Interface Version 5.0 November 2018:
* 2.19.4.1default Clause
The default clause explicitly determines the data-sharing attributes of
variables that are referenced *in a parallel, teams, or task generating
construct and would otherwise be implicitly determined
(see Section 2.19.1.1 on page 270).
* 2.6.1 Determining the Number of Threads for a parallel Region
Using a variable in an if or num_threads clause expression of a parallel
construct causes an implicit reference to the variable in all enclosing
constructs. The if clause expression and the num_threads clause expression
are evaluated in the context outside of the parallel construct,
This reverts commit r360073.
llvm-svn: 360326
default(none).
If the combined directive has default(none) clause and has clauses for
inner directive that reference some variables, for which data-sharing
attributes are not specified, the error messages should be emitted for
such variables.
llvm-svn: 360073
If the `default(none)` was specified for the construct, we might miss
diagnostic for the globals without explicitly specified data-sharing
attributes. Patch fixes this problem.
llvm-svn: 360061
counters.
According to the OpenMP 5.0, For any associated loop where the b or lb
expression is not loop invariant with respect to the outermost loop, the
var-outer that appears in the expression may not have a random access
iterator type.
llvm-svn: 359340
loop nests.
Added a checks that the initializer/condition expressions depend only
only of the single previous loop iteration variable.
llvm-svn: 359200
Summary: The requires directive containing target related clauses must appear before any target region in the compilation unit.
Reviewers: ABataev, AlexEichenberger, caomhin
Reviewed By: ABataev
Subscribers: guansong, jfb, jdoerfert, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D60875
llvm-svn: 358709
According to OpenMP 5.0, 2.11.4 allocate Clause, Restrictions, allocate
clauses that appear on a target construct or on constructs in a target
region must specify an allocator expression unless a requires directive
with the dynamic_allocators clause is present in the same compilation
unit. Patch adds a check for this restriction.
llvm-svn: 357412
According to OpenMP 5.0 standard, 2.11.4 allocate Clause, Restrictions,
For any list item that is specified in the allocate clause on a
directive, a data-sharing attribute clause that may create a private
copy of that list item must be specified on the same directive. Patch
adds the checks for this restriction.
llvm-svn: 357390
target and task-based directives.
According to OpenMP 5.0, 2.11.4 allocate Clause, Restrictions, For task,
taskloop or target directives, allocation requests to memory allocators
with the trait access set to thread result in unspecified behavior.
Patch introduces a check for omp_thread_mem_alloc predefined allocator
on target- and trask-based directives.
llvm-svn: 357205
The various CorrectionCandidateCallbacks are currently heap-allocated
unconditionally. This was needed because of delayed typo correction.
However these allocations represent currently 15.4% of all allocations
(number of allocations) when parsing all of Boost (!), mostly because
of ParseCastExpression, ParseStatementOrDeclarationAfterAttrtibutes
and isCXXDeclarationSpecifier. Note that all of these callback objects
are small. Let's not do this.
Instead initially allocate the callback on the stack, and only do a
heap allocation if we are going to do some typo correction. Do this by:
1. Adding a clone function to each callback, which will do a polymorphic
clone of the callback. This clone function is required to be implemented
by every callback (of which there is a fair amount). Make sure this is
the case by making it pure virtual.
2. Use this clone function when we are going to try to correct a typo.
This additionally cut the time of -fsyntax-only on all of Boost by 0.5%
(not that much, but still something). No functional changes intended.
Differential Revision: https://reviews.llvm.org/D58827
Reviewed By: rnk
llvm-svn: 356925
dynamic_allocators.
According to the OpenMP 5.0, 2.11.3 allocate Directive, Restrictions,
allocate directives that appear in a target region must specify an
allocator clause unless a requires directive with the dynamic_allocators
clause is present in the same compilation unit. Patch adds a check for a
presence of the requires directive with the dynamic_allocators clause.
llvm-svn: 356758
clause in target region.
According to the OpenMP 5.0, 2.11.3 allocate Directive, Restrictions,
allocate directives that appear in a target region must specify an
allocator clause unless a requires directive with the dynamic_allocators
clause is present in the same compilation unit.
llvm-svn: 356752
Previously implemented check required the reevaluation of the already
evaluated predefined allocator kind for the global variables. Patch
simplifies this evaluation and removes extra code.
llvm-svn: 356699
allocators.
It is better to deduce omp_allocator_handle_t type from the predefined
allocators, because omp.h header might not define it explicitly. Plus,
it allows to identify the predefined allocators correctly when trying to
build the allcoator for the global variables.
llvm-svn: 356607
Summary:
https://www.openmp.org/wp-content/uploads/OpenMP-API-Specification-5.0.pdf, page 3:
```
structured block
For C/C++, an executable statement, possibly compound, with a single entry at the
top and a single exit at the bottom, or an OpenMP construct.
COMMENT: See Section 2.1 on page 38 for restrictions on structured
blocks.
```
```
2.1 Directive Format
Some executable directives include a structured block. A structured block:
• may contain infinite loops where the point of exit is never reached;
• may halt due to an IEEE exception;
• may contain calls to exit(), _Exit(), quick_exit(), abort() or functions with a
_Noreturn specifier (in C) or a noreturn attribute (in C/C++);
• may be an expression statement, iteration statement, selection statement, or try block, provided
that the corresponding compound statement obtained by enclosing it in { and } would be a
structured block; and
Restrictions
Restrictions to structured blocks are as follows:
• Entry to a structured block must not be the result of a branch.
• The point of exit cannot be a branch out of the structured block.
C / C++
• The point of entry to a structured block must not be a call to setjmp().
• longjmp() and throw() must not violate the entry/exit criteria.
```
Of particular note here is the fact that OpenMP structured blocks are as-if `noexcept`,
in the same sense as with the normal `noexcept` functions in C++.
I.e. if throw happens, and it attempts to travel out of the `noexcept` function
(here: out of the current structured-block), then the program terminates.
Now, one of course can say that since it is explicitly prohibited by the Specification,
then any and all programs that violate this Specification contain undefined behavior,
and are unspecified, and thus no one should care about them. Just don't write broken code /s
But i'm not sure this is a reasonable approach.
I have personally had oss-fuzz issues of this origin - exception thrown inside
of an OpenMP structured-block that is not caught, thus causing program termination.
This issue isn't all that hard to catch, it's not any particularly different from
diagnosing the same situation with the normal `noexcept` function.
Now, clang static analyzer does not presently model exceptions.
But clang-tidy has a simplisic [[ https://clang.llvm.org/extra/clang-tidy/checks/bugprone-exception-escape.html | bugprone-exception-escape ]] check,
and it is even refactored as a `ExceptionAnalyzer` class for reuse.
So it would be trivial to use that analyzer to check for
exceptions escaping out of OpenMP structured blocks. (D59466)
All that sounds too great to be true. Indeed, there is a caveat.
Presently, it's practically impossible to do. To check a OpenMP structured block
you need to somehow 'get' the OpenMP structured block, and you can't because
it's simply not modelled in AST. `CapturedStmt`/`CapturedDecl` is not it's representation.
Now, it is of course possible to write e.g. some AST matcher that would e.g.
match every OpenMP executable directive, and then return the whatever `Stmt` is
the structured block of said executable directive, if any.
But i said //practically//. This isn't practical for the following reasons:
1. This **will** bitrot. That matcher will need to be kept up-to-date,
and refreshed with every new OpenMP spec version.
2. Every single piece of code that would want that knowledge would need to
have such matcher. Well, okay, if it is an AST matcher, it could be shared.
But then you still have `RecursiveASTVisitor` and friends.
`2 > 1`, so now you have code duplication.
So it would be reasonable (and is fully within clang AST spirit) to not
force every single consumer to do that work, but instead store that knowledge
in the correct, and appropriate place - AST, class structure.
Now, there is another hoop we need to get through.
It isn't fully obvious //how// to model this.
The best solution would of course be to simply add a `OMPStructuredBlock` transparent
node. It would be optimal, it would give us two properties:
* Given this `OMPExecutableDirective`, what's it OpenMP structured block?
* It is trivial to check whether the `Stmt*` is a OpenMP structured block (`isa<OMPStructuredBlock>(ptr)`)
But OpenMP structured block isn't **necessarily** the first, direct child of `OMP*Directive`.
(even ignoring the clang's `CapturedStmt`/`CapturedDecl` that were inserted inbetween).
So i'm not sure whether or not we could re-create AST statements after they were already created?
There would be other costs to a new AST node: https://bugs.llvm.org/show_bug.cgi?id=40563#c12
```
1. You will need to break the representation of loops. The body should be replaced by the "structured block" entity.
2. You will need to support serialization/deserialization.
3. You will need to support template instantiation.
4. You will need to support codegen and take this new construct to account in each OpenMP directive.
```
Instead, there **is** an functionally-equivalent, alternative solution, consisting of two parts.
Part 1:
* Add a member function `isStandaloneDirective()` to the `OMPExecutableDirective` class,
that will tell whether this directive is stand-alone or not, as per the spec.
We need it because we can't just check for the existance of associated statements,
see code comment.
* Add a member function `getStructuredBlock()` to the OMPExecutableDirective` class itself,
that assert that this is not a stand-alone directive, and either return the correct loop body
if this is a loop-like directive, or the captured statement.
This way, given an `OMPExecutableDirective`, we can get it's structured block.
Also, since the knowledge is ingrained into the clang OpenMP implementation,
it will not cause any duplication, and //hopefully// won't bitrot.
Great we achieved 1 of 2 properties of `OMPStructuredBlock` approach.
Thus, there is a second part needed:
* How can we check whether a given `Stmt*` is `OMPStructuredBlock`?
Well, we can't really, in general. I can see this workaround:
```
class FunctionASTVisitor : public RecursiveASTVisitor<FunctionASTVisitor> {
using Base = RecursiveASTVisitor<FunctionASTVisitor>;
public:
bool VisitOMPExecDir(OMPExecDir *D) {
OmpStructuredStmts.emplace_back(D.getStructuredStmt());
}
bool VisitSOMETHINGELSE(???) {
if(InOmpStructuredStmt)
HI!
}
bool TraverseStmt(Stmt *Node) {
if (!Node)
return Base::TraverseStmt(Node);
if (OmpStructuredStmts.back() == Node)
++InOmpStructuredStmt;
Base::TraverseStmt(Node);
if (OmpStructuredStmts.back() == Node) {
OmpStructuredStmts.pop_back();
--InOmpStructuredStmt;
}
return true;
}
std::vector<Stmt*> OmpStructuredStmts;
int InOmpStructuredStmt = 0;
};
```
But i really don't see using it in practice.
It's just too intrusive; and again, requires knowledge duplication.
.. but no. The solution lies right on the ground.
Why don't we simply store this `i'm a openmp structured block` in the bitfield of the `Stmt` itself?
This does not appear to have any impact on the memory footprint of the clang AST,
since it's just a single extra bit in the bitfield. At least the static assertions don't fail.
Thus, indeed, we can achieve both of the properties without a new AST node.
We can cheaply set that bit right in sema, at the end of `Sema::ActOnOpenMPExecutableDirective()`,
by just calling the `getStructuredBlock()` that we just added.
Test coverage that demonstrates all this has been added.
This isn't as great with serialization though. Most of it does not use abbrevs,
so we do end up paying the full price (4 bytes?) instead of a single bit.
That price, of course, can be reclaimed by using abbrevs.
In fact, i suspect that //might// not just reclaim these bytes, but pack these PCH significantly.
I'm not seeing a third solution. If there is one, it would be interesting to hear about it.
("just don't write code that would require `isa<OMPStructuredBlock>(ptr)`" is not a solution.)
Fixes [[ https://bugs.llvm.org/show_bug.cgi?id=40563 | PR40563 ]].
Reviewers: ABataev, rjmccall, hfinkel, rsmith, riccibruno, gribozavr
Reviewed By: ABataev, gribozavr
Subscribers: mgorny, aaron.ballman, steveire, guansong, jfb, jdoerfert, cfe-commits
Tags: #clang, #openmp
Differential Revision: https://reviews.llvm.org/D59214
llvm-svn: 356570
If the allocator was specified for the variable and next one is found
with the different allocator, the warning is emitted, and the allocator
is ignored.
llvm-svn: 356513
According to OpenMP, 2.11.3 allocate Directive, Restrictions, C / C++,
if a list item has a static storage type, the allocator expression in
the allocator clause must be a constant expression that evaluates to
one of the predefined memory allocator values. Added check for this
restriction.
llvm-svn: 356496
Summary:
This patch refactors several instances of cast<> used in if
conditionals. Since cast<> asserts on failure, the else branch can
never be taken.
In some cases, the fix is to replace cast<> with dyn_cast<>. While
others required the removal of the conditional and some minor
refactoring.
A discussion can be seen here: http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20190318/265044.html
Differential Revision: https://reviews.llvm.org/D59529
llvm-svn: 356441
If the doacross lop construct is used and the loop counter is declare
outside of the loop, the compiler might crash trying to get the address
of the loop counter. Patch fixes this problem.
llvm-svn: 356198
If the declare target link global is used in the target region
indirectly (used in the inner parallel, teams, etc. regions), we may
miss this variable and it leads to incorrect codegen.
llvm-svn: 355858
This patch implements the parsing and sema support for the OpenMP
'from'-clause with potential user-defined mappers attached.
User-defined mappers are a new feature in OpenMP 5.0. A 'from'-clause
can have an explicit or implicit associated mapper, which instructs the
compiler to generate and use customized mapping functions. An example is
shown below:
struct S { int len; int *d; };
#pragma omp declare mapper(id: struct S s) map(s, s.d[0:s.len])
struct S ss;
#pragma omp target update from(mapper(id): ss) // use the mapper with name 'id' to map ss from device
Contributed-by: Lingda Li <lildmh@gmail.com>
Differential Revision: https://reviews.llvm.org/D58638
llvm-svn: 354817
This patch implements the parsing and sema support for OpenMP to clause
with potential user-defined mappers attached. User defined mapper is a
new feature in OpenMP 5.0. A to/from clause can have an explicit or
implicit associated mapper, which instructs the compiler to generate and
use customized mapping functions. An example is shown below:
struct S { int len; int *d; };
#pragma omp declare mapper(id: struct S s) map(s, s.d[0:s.len])
struct S ss;
#pragma omp target update to(mapper(id): ss) // use the mapper with name 'id' to map ss to device
Contributed-by: <lildmh@gmail.com>
Differential Revision: https://reviews.llvm.org/D58523
llvm-svn: 354698
This patch implements the parsing and sema support for OpenMP map
clauses with potential user-defined mapper attached. User defined mapper
is a new feature in OpenMP 5.0. A map clause can have an explicit or
implicit associated mapper, which instructs the compiler to generate
extra data mapping. An example is shown below:
struct S { int len; int *d; };
#pragma omp declare mapper(id: struct S s) map(s, s.d[0:s.len])
struct S ss;
#pragma omp target map(mapper(id) tofrom: ss) // use the mapper with name 'id' to map ss
Contributed-by: Lingda Li <lildmh@gmail.com>
Differential Revision: https://reviews.llvm.org/D58074
llvm-svn: 354347
Fixed diagnostic emission for the exceptions support in case of the
compilation of OpenMP code for the devices. From now on, it uses delayed
diagnostics mechanism, previously used for CUDA only. It allow to
diagnose not allowed used of exceptions only in functions that are going
to be codegen'ed.
llvm-svn: 353542
It is important to delay the emission of the diagnostic messages for the
functions unless it is proved that the function is going to be used on
the device side. It is required to support compilation with some of the
target-specific system headers.
llvm-svn: 353540
The fix is to issue error messages if there are more than one
teams construct inside a target constructs.
#pragma omp target
{
#pragma omp teams
{ ... }
#pragma omp teams
{ ... }
}
llvm-svn: 353186
edge cases.
Currently, EmitCall emits a call instruction with a function type
derived from the pointee-type of the callee. This *should* be the same
as the type created from the CallInfo parameter, but in some cases an
incorrect CallInfo was being passed.
All of these fixes were discovered by the addition of the assert in
EmitCall which verifies that the passed-in CallInfo matches the
Callee's function type.
As far as I know, these issues caused no bugs at the moment, as the
correct types were ultimately being emitted. But, some would become
problematic when pointee types are removed.
List of fixes:
* arrangeCXXConstructorCall was passing an incorrect value for the
number of Required args, when calling an inheriting constructor
where the inherited constructor is variadic. (The inheriting
constructor doesn't actually get passed any of the user's args, but
the code was calculating it as if it did).
* arrangeFreeFunctionLikeCall was not including the count of the
pass_object_size arguments in the count of required args.
* OpenCL uses other address spaces for the "this" pointer. However,
commonEmitCXXMemberOrOperatorCall was not annotating the address
space on the "this" argument of the call.
* Destructor calls were being created with EmitCXXMemberOrOperatorCall
instead of EmitCXXDestructorCall in a few places. This was a problem
because the calling convention sometimes has destructors returning
"this" rather than void, and the latter function knows about that,
and sets up the types properly (through calling
arrangeCXXStructorDeclaration), while the former does not.
* generateObjCGetterBody: the 'objc_getProperty' function returns type
'id', but was being called as if it returned the particular
property's type. (That is of course the *dynamic* return type, and
there's a downcast immediately after.)
* OpenMP user-defined reduction functions (#pragma omp declare
reduction) can be called with a subclass of the declared type. In
such case, the call was being setup as if the function had been
actually declared to take the subtype, rather than the base type.
Differential Revision: https://reviews.llvm.org/D57664
llvm-svn: 353181
Further reviews (D57594, D57615) have revealed that this was not reviewed,
and that the differential's description was not read during the review,
thus rendering this commit invalid.
This reverts commit r352882.
llvm-svn: 352933
This patch implements parsing and sema for "omp declare mapper"
directive. User defined mapper, i.e., declare mapper directive, is a new
feature in OpenMP 5.0. It is introduced to extend existing map clauses
for the purpose of simplifying the copy of complex data structures
between host and device (i.e., deep copy). An example is shown below:
struct S { int len; int *d; };
#pragma omp declare mapper(struct S s) map(s, s.d[0:s.len]) // Memory region that d points to is also mapped using this mapper.
Contributed-by: Lingda Li <lildmh@gmail.com>
Differential Revision: https://reviews.llvm.org/D56326
llvm-svn: 352906
Summary:
I'm working on a clang-tidy check, much like existing [[ http://clang.llvm.org/extra/clang-tidy/checks/bugprone-exception-escape.html | bugprone-exception-escape ]],
to detect when an exception might escape out of an OpenMP construct it isn't supposed to escape from.
For that i will be using the `nothrow` bit of `CapturedDecl`s.
While that bit is already correctly set for some constructs, e.g. `#pragma omp parallel`: https://godbolt.org/z/2La7pv
it isn't set for the `#pragma omp sections`, or `#pragma omp section`: https://godbolt.org/z/qZ-EbP
If i'm reading [[ https://www.openmp.org/wp-content/uploads/OpenMP-API-Specification-5.0.pdf | `OpenMP Application Programming Interface Version 5.0 November 2018` ]] correctly,
they should be, as per `2.8.1 sections Construct`, starting with page 86:
* The sections construct is a non-iterative worksharing construct that contains a set of **structured blocks**
that are to be distributed among and executed by the threads in a team. Each **structured block** is executed
once by one of the threads in the team in the context of its implicit task.
* The syntax of the sections construct is as follows:
#pragma omp sections [clause[ [,] clause] ... ] new-line
{
[#pragma omp section new-line]
**structured-block**
...
* Description
Each **structured block** in the sections construct is preceded by a section directive except
possibly **the first block**, for which a preceding section directive is optional.
* Restrictions
• The code enclosed in a sections construct must be a **structured block**.
* A throw executed inside a sections region must cause execution to resume within the same
section of the sections region, and the same thread that threw the exception must catch it.
Reviewers: ABataev, #openmp
Reviewed By: ABataev
Subscribers: guansong, openmp-commits, cfe-commits
Tags: #clang, #openmp
Differential Revision: https://reviews.llvm.org/D57585
llvm-svn: 352882
We don't need to use the predetermined data-sharing attributes for the
loop counters if the user explicitly specified correct data-sharing
attributes for such variables.
llvm-svn: 352543
According to the report, better to keep the original strict compare
operation as the loop condition with unsigned loop counters to make the
loop countable. This allows further loop transformations.
llvm-svn: 352526
to reflect the new license.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351636
Summary: Change the strategy for computing loop index variables after collapsing a loop nest via the collapse clause by replacing the expensive remainder operation with multiplications and additions.
Reviewers: ABataev, caomhin
Reviewed By: ABataev
Subscribers: guansong, arphaman, cfe-commits
Differential Revision: https://reviews.llvm.org/D56413
llvm-svn: 350759
Summary: Introduce a compiler flag for cases when the user knows that the collapsed loop counter can be safely represented using at most 32 bits. This will prevent the emission of expensive mathematical operations (such as the div operation) on the iteration variable using 64 bits where 32 bit operations are sufficient.
Reviewers: ABataev, caomhin
Reviewed By: ABataev
Subscribers: hfinkel, kkwli0, guansong, cfe-commits
Differential Revision: https://reviews.llvm.org/D55928
llvm-svn: 350758
Summary:
There is a minor issue in how the implicit data-sharings for nested tasks are computed.
For the following example:
```
int x;
#pragma omp task shared(x)
#pragma omp task
x++;
```
We compute an implicit data-sharing of shared for `x` in the second task although I think that it should be firstprivate. Below you can find the part of the OpenMP spec that covers this example:
- // In a task generating construct, if no default clause is present, a variable for which the data-sharing attribute is not determined by the rules above and that in the enclosing context is determined to be shared by all implicit tasks bound to the current team is shared.//
- //In a task generating construct, if no default clause is present, a variable for which the data-sharing attribute is not determined by the rules above is firstprivate.//
Since each implicit-task has its own copy of `x`, we shouldn't apply the first rule.
Reviewers: ABataev
Reviewed By: ABataev
Subscribers: cfe-commits, rogfer01
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D56430
llvm-svn: 350734
As discussed in D56113, this patch refactors the implementation of the
const restriction for linear to reuse a function introduced by D56113.
A side effect is that, if a variable has mutable members, this
diagnostic is now skipped, and the diagnostic for the variable not
being an integer or pointer is reported instead.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D56299
llvm-svn: 350441
As discussed in D56113, this patch refactors the implementation of the
const restriction for reductions to reuse a function introduced by
D56113. A side effect is that diagnostics sometimes now say
"variable" instead of "list item" when a list item is a variable.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D56298
llvm-svn: 350440
The following appears in OpenMP 3.1 sec. 2.9.1.1 as a predetermined
data-sharing attribute:
> Variables with const-qualified type having no mutable member are
> shared.
It does not appear in OpenmP 4.0, 4.5, or 5.0. This patch removes the
implementation of that attribute when the requested OpenMP version is
greater than 3.1.
One effect of that removal is that `default(none)` affects const
variables without mutable members.
Also, without this patch, if a const variable without mutable members
was explicitly lastprivate or private, it was an error because it was
predetermined shared. Now, clang instead complains that it's const
without mutable fields, which is a more intelligible diagnostic. That
should be fine for all of the above versions because they all have
something like the following, which is quoted from OpenMP 5.0
sec. 2.19.3:
> A variable that is privatized must not have a const-qualified type
> unless it is of class type with a mutable member. This restriction does
> not apply to the firstprivate clause.
reduction and linear clauses already have separate checks for const
variables. Future patches will merge the implementations.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D56113
llvm-svn: 350439
Rather than sprinkle calls to DiagnoseUnusedExprResult() around in places where we want diagnostics, we now diagnose unused expression statements and full expressions in a more generic way when acting on the final expression statement. This results in more appropriate diagnostics for [[nodiscard]] where we were previously lacking them, such as when the body of a for loop is not a compound statement.
This patch fixes PR39837.
llvm-svn: 350404
For constants with the predefined data-sharing clauses we may had
troubles with the target combined directives. It may cause compiler
crash in some corner cases.
llvm-svn: 350127
Since CallExpr::setNumArgs has been removed, it is now possible to store the
callee expression and the argument expressions of CallExpr in a trailing array.
This saves one pointer per CallExpr, CXXOperatorCallExpr, CXXMemberCallExpr,
CUDAKernelCallExpr and UserDefinedLiteral.
Given that CallExpr is used as a base of the above classes we cannot use
llvm::TrailingObjects. Instead we store the offset in bytes from the this pointer
to the start of the trailing objects and manually do the casts + arithmetic.
Some notes:
1.) I did not try to fit the number of arguments in the bit-fields of Stmt.
This leaves some space for future additions and avoid the discussion about
whether x bits are sufficient to hold the number of arguments.
2.) It would be perfectly possible to recompute the offset to the trailing
objects before accessing the trailing objects. However the trailing objects
are frequently accessed and benchmarks show that it is slightly faster to
just load the offset from the bit-fields. Additionally, because of 1),
we have plenty of space in the bit-fields of Stmt.
Differential Revision: https://reviews.llvm.org/D55771
Reviewed By: rjmccall
llvm-svn: 349910
Need to mark the loop as started when the initialization statement is
found. It is required to prevent possible incorrect loop iteraton
variable detection during template instantiation and fix the compiler
crash during the codegen.
llvm-svn: 349657
Without this patch, clang doesn't complain that X needs explicit data
sharing attributes in the following:
```
#pragma omp target teams default(none)
{
#pragma omp parallel num_threads(X)
;
}
```
However, clang does produce that complaint after the braces are
removed. With this patch, clang complains in both cases.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D55861
llvm-svn: 349635
A map clause with the close map-type-modifier is a hint to
prefer that the variables are mapped using a copy into faster
memory.
Patch by Ahsan Saghir (saghir)
Differential Revision: https://reviews.llvm.org/D55719
llvm-svn: 349551
It seems the two failing tests can be simply fixed after r348037
Fix 3 cases in Analysis/builtin-functions.cpp
Delete the bad CodeGen/builtin-constant-p.c for now
llvm-svn: 348053
Kept the "indirect_builtin_constant_p" test case in test/SemaCXX/constant-expression-cxx1y.cpp
while we are investigating why the following snippet fails:
extern char extern_var;
struct { int a; } a = {__builtin_constant_p(extern_var)};
llvm-svn: 348039
This was reverted in r347656 due to me thinking it caused a miscompile of
Chromium. Turns out it was the Chromium code that was broken.
llvm-svn: 347756
This caused a miscompile in Chrome (see crbug.com/908372) that's
illustrated by this small reduction:
static bool f(int *a, int *b) {
return !__builtin_constant_p(b - a) || (!(b - a));
}
int arr[] = {1,2,3};
bool g() {
return f(arr, arr + 3);
}
$ clang -O2 -S -emit-llvm a.cc -o -
g() should return true, but after r347417 it became false for some reason.
This also reverts the follow-up commits.
r347417:
> Re-Reinstate 347294 with a fix for the failures.
>
> Don't try to emit a scalar expression for a non-scalar argument to
> __builtin_constant_p().
>
> Third time's a charm!
r347446:
> The result of is.constant() is unsigned.
r347480:
> A __builtin_constant_p() returns 0 with a function type.
r347512:
> isEvaluatable() implies a constant context.
>
> Assume that we're in a constant context if we're asking if the expression can
> be compiled into a constant initializer. This fixes the issue where a
> __builtin_constant_p() in a compound literal was diagnosed as not being
> constant, even though it's always possible to convert the builtin into a
> constant.
r347531:
> A "constexpr" is evaluated in a constant context. Make sure this is reflected
> if a __builtin_constant_p() is a part of a constexpr.
llvm-svn: 347656
Loop-control variables with the default data-sharing attributes should
not be captured in the OpenMP region as they are private by default.
Also, default attributes should be emitted for such variables in the
inner OpenMP regions for the correct data sharing during codegen.
llvm-svn: 347409
forms of random access iterator
In OpenMP 4.5, only 4 relational operators are supported: <, <=, >,
and >=. This work is to enable support for relational operator
!= (not-equal) as one of the canonical forms.
Patch by Anh Tuyen Tran
Differential Revision: https://reviews.llvm.org/D54441
llvm-svn: 347405
If lambda is used inside of the OpenMP region and captures `this`, we
should recapture it in the OpenMP region also. But we should do this
only if the OpenMP region is used in the context of the same class, just
like the lambda.
llvm-svn: 347096
A ConstantExpr class represents a full expression that's in a context where a
constant expression is required. This class reflects the path the evaluator
took to reach the expression rather than the syntactic context in which the
expression occurs.
In the future, the class will be expanded to cache the result of the evaluated
expression so that it's not needlessly re-evaluated
Reviewed By: rsmith
Differential Revision: https://reviews.llvm.org/D53475
llvm-svn: 345692
Added support for mapping of lambdas in the target regions. It scans all
the captures by reference in the lambda, implicitly maps those variables
in the target region and then later reinstate the addresses of
references in lambda to the correct addresses of the captured|privatized
variables.
llvm-svn: 345609
being shared.
According to the standard, the variables with unspecified data-sharing
attributes in presence of `default(none)` clause must be reported to
users. Compiler did not generate error reports for the variables used in
other OpenMP regions. Patch fixes this.
llvm-svn: 345533
Summary: Iteration variable must be strictly less than the number of iterations. This fixes a bug introduced by previous patch D53448.
Reviewers: ABataev, caomhin
Reviewed By: ABataev
Subscribers: guansong, cfe-commits
Differential Revision: https://reviews.llvm.org/D53827
llvm-svn: 345527
Summary: This patch adds a new code generation path for bound sharing directives containing distribute parallel for. The new code generation scheme applies to chunked schedules on distribute and parallel for directives. The scheme simplifies the code that is being generated by eliminating the need for an outer for loop over chunks for both distribute and parallel for directives. In the case of distribute it applies to any sized chunk while in the parallel for case it only applies when chunk size is 1.
Reviewers: ABataev, caomhin
Reviewed By: ABataev
Subscribers: jholewinski, guansong, cfe-commits
Differential Revision: https://reviews.llvm.org/D53448
llvm-svn: 345509
If the loop counter is not declared in the context of the loop and it is
private, such loop counters should not be captured in the outlined
regions.
llvm-svn: 345505
According to the OpenMP standard, In a task generating construct, if no
default clause is present, a variable for which the data-sharing
attribute is not determined by the rules above is firstprivatized.
Compiler tries to implement this, but if the variable is not directly
used in the task context, this variable may not be firstprivatized.
Patch fixes this problem.
llvm-svn: 345277
The compiler is crashing if we trying to post-capture the fields
implicitly captured inside of the task constructs. Seems, this kind of
processing is not supported and such fields should not be
firstprivatized.
llvm-svn: 345177
According to OpenMP, the reduction item must be shared in parent region.
But the item can be an array section or array subscript. In this case,
we should not check for the datasharing of the base declaration.
llvm-svn: 343356
Add support for OMP5.0 requires directive and unified_address clause.
Patches to follow will include support for additional clauses.
Differential Revision: https://reviews.llvm.org/D52359
llvm-svn: 343063
declare reduction.
If the declare reduction construct with the non-dependent type is
defined in the template construct, the compiler might crash on the
template instantition. Reworked the whole instantiation scheme for the
declare reduction constructs to fix this problem correctly.
llvm-svn: 342151
Sema analysis should not mark functions as an implicit declare target,
it may break codegen. Simplified semantic analysis and removed extra
code for implicit declare target functions.
llvm-svn: 341939
Add the capability to nest multiple declare target directives
- including header files within a declare target region.
Differential Revision: https://reviews.llvm.org/D51378
Patch by Patrick Lyster
llvm-svn: 341766
declare target.
According to OpenMP 5.0, variables captured in lambdas in declare target
regions must be considered as implicitly declare target.
llvm-svn: 339152
When the parsing of the functions happens inside of the declare target
region, we may erroneously mark local variables as declare target
thought they are not. This attribute can be applied only to global
variables.
llvm-svn: 336592
A reduction for an incomplete array type used to produce an assert
fail during codegen. Now it produces a diagnostic.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D48735
llvm-svn: 335911
This is similar to the LLVM change https://reviews.llvm.org/D46290.
We've been running doxygen with the autobrief option for a couple of
years now. This makes the \brief markers into our comments
redundant. Since they are a visual distraction and we don't want to
encourage more \brief markers in new code either, this patch removes
them all.
Patch produced by
for i in $(git grep -l '\@brief'); do perl -pi -e 's/\@brief //g' $i & done
for i in $(git grep -l '\\brief'); do perl -pi -e 's/\\brief //g' $i & done
Differential Revision: https://reviews.llvm.org/D46320
llvm-svn: 331834
We should not emit warning that the parameters are not marked as declare
target, these declaration are local and cannot be marked as declare
target.
llvm-svn: 331211
If the link clause is used on the declare target directive, the object
should be linked on target or target data directives, not during the
codegen. Patch adds support for this clause.
llvm-svn: 328544
If the generic codegen is enabled and private copy of the original
variable escapes the declaration context, this private copy should be
globalized just like it was the original variable.
llvm-svn: 327985
Added initial codegen for device side of declarations inside `omp
declare target` construct + codegen for implicit `declare target`
functions, which are used in the target regions.
llvm-svn: 327636
This relands r326965.
There was a null dereference in typo correction that was triggered in
Sema/diagnose_if.c. We are not always in a function scope when doing
typo correction. The fix is to add a null check.
LLVM's optimizer made it hard to find this bug. I wrote it up in a
not-very-well-editted blog post here:
http://qinsb.blogspot.com/2018/03/ub-will-delete-your-null-checks.html
llvm-svn: 327334
This reverts r326965. It seems to have caused repeating test failures in
clang/test/Sema/diagnose_if.c on some buildbots.
I cannot reproduce the problem, and it's not immediately obvious what
the problem is, so let's revert to green.
llvm-svn: 326974
Summary:
Before this patch, Sema pre-allocated a FunctionScopeInfo and kept it in
the first, always present element of the FunctionScopes stack. This
meant that Sema::getCurFunction would return a pointer to this
pre-allocated object when parsing code outside a function body. This is
pretty much always a bug, so this patch moves the pre-allocated object
into a separate unique_ptr. This should make bugs like PR36536 a lot
more obvious.
As you can see from this patch, there were a number of places that
unconditionally assumed they were always called inside a function.
However, there are also many places that null checked the result of
getCurFunction(), so I think this is a reasonable direction.
Reviewers: rsmith
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D44039
llvm-svn: 326965
In CUDA mode all local variables are actually thread
local|threadprivate, not private, and, thus, they cannot be shared
between threads|lanes.
llvm-svn: 326590
So I wrote a clang-tidy check to lint out redundant `isa`, `cast`, and
`dyn_cast`s for fun. This is a portion of what it found for clang; I
plan to do similar cleanups in LLVM and other subprojects when I find
time.
Because of the volume of changes, I explicitly avoided making any change
that wasn't highly local and obviously correct to me (e.g. we still have
a number of foo(cast<Bar>(baz)) that I didn't touch, since overloading
is a thing and the cast<Bar> did actually change the type -- just up the
class hierarchy).
I also tried to leave the types we were cast<>ing to somewhere nearby,
in cases where it wasn't locally obvious what we were dealing with
before.
llvm-svn: 326416
If several member expressions are mapped and they reference the same
address as a base, but access different members, this must be allowed.
llvm-svn: 326212
constructs.
The compiler may emit some extra warnings for functions, that are
implicit specialization of the templates, declared in the target region.
llvm-svn: 325391
parallel for simd` directives.
Added codegen for `depend` clauses on `#pragma omp target teams
distribute parallel for simd` directives.
llvm-svn: 322587
simd`.
Added host codegen + codegen for devices with default codegen for
`#pragma omp target teams distribute parallel for simd` directive.
llvm-svn: 322515
Patch fixes incorrect capturing of the expressions in clauses with
expressions that must be captured for the combined constructs. Incorrect
capturing may lead to compiler crash during codegen phase.
llvm-svn: 321820
If the clause is applied to the combined construct and has captured
expression, try to capture this expression by value rather than by
reference.
llvm-svn: 321386
Adding the new enumerator forced a bunch more changes into this patch than I
would have liked. The -Wtautological-compare warning was extended to properly
check the new comparison operator, clang-format needed updating because it uses
precedence levels as weights for determining where to break lines (and several
operators increased their precedence levels with this change), thread-safety
analysis needed changes to build its own IL properly for the new operator.
All "real" semantic checking for this operator has been deferred to a future
patch. For now, we use the relational comparison rules and arbitrarily give
the builtin form of the operator a return type of 'void'.
llvm-svn: 320707
This patch is to add diagnose when a function name is
specified on the link clause. According to the OpenMP
spec, only the list items that exclude the function
name are allowed on the link clause.
Differential Revision: https://reviews.llvm.org/D40968
llvm-svn: 320521
If the error is generated during analysis of implicitly or explicitly
mapped variables, it may cause compiler crash because of incorrect
analysis.
llvm-svn: 319774
distribute directives.
OpenMP standard does not allow to mark the variables as firstprivate and lastprivate at the same time in distribute-based directives. Patch fixes this problem.
llvm-svn: 319560
Clang asserts on undeclared variables on the to or link clause in the declare
target directive. The patch is to properly diagnose the error.
// foo1 and foo2 are not declared
#pragma omp declare target to(foo1)
#pragma omp declare target link(foo2)
Differential Revision: https://reviews.llvm.org/D40588
llvm-svn: 319458
teams region.
If the inner teams region is not correct, it may cause an assertion when
processing outer target region. Patch fixes this problem.
llvm-svn: 319450
directives.
According to the OpenMP standard, only loop control variables can be
used in linear clauses of distribute-based simd directives.
llvm-svn: 319362
OpenMP 5.0 introduces asynchronous data update/dependecies clauses on
target data directives. Patch adds initial support for outer task
regions to use task-based codegen for future async target data
directives.
llvm-svn: 318781
https://reviews.llvm.org/D40187
This patch implements code gen for 'teams distribute parallel for' on the host, including all its clauses and related regression tests.
llvm-svn: 318692
Some target devices (e.g. Nvidia GPUs) don't support dynamic stack
allocation and hence no VLAs. Print errors with description instead
of failing in the backend or generating code that doesn't work.
This patch handles explicit uses of VLAs (local variable in target
or declare target region) or implicitly generated (private) VLAs
for reductions on VLAs or on array sections with non-constant size.
Differential Revision: https://reviews.llvm.org/D39505
llvm-svn: 318601
We can generate constant sized arrays whenever the array section has constant
length, even if the base expression itself is a VLA.
Differential Revision: https://reviews.llvm.org/D39504
llvm-svn: 317207
In some cases the compiler can deduce the length of an array section
as constants. With this information, VLAs can be avoided in place of
a constant sized array or even a scalar value if the length is 1.
Example:
int a[4], b[2];
pragma omp parallel reduction(+: a[1:2], b[1:1])
{ }
For chained array sections, this optimization is restricted to cases
where all array sections except the last have a constant length 1.
This trivially guarantees that there are no holes in the memory region
that needs to be privatized.
Example:
int c[3][4];
pragma omp parallel reduction(+: c[1:1][1:2])
{ }
This relands commit r316229 that I reverted in r316235 because it
failed on some bots. During investigation I found that this was because
Clang and GCC evaluate the two arguments to emplace_back() in
ReductionCodeGen::emitSharedLValue() in a different order, hence
leading to a different order of generated instructions in the final
LLVM IR. Fix this by passing in the arguments from temporary variables
that are evaluated in a defined order.
Differential Revision: https://reviews.llvm.org/D39136
llvm-svn: 316362
In some cases the compiler can deduce the length of an array section
as constants. With this information, VLAs can be avoided in place of
a constant sized array or even a scalar value if the length is 1.
Example:
int a[4], b[2];
pragma omp parallel reduction(+: a[1:2], b[1:1])
{ }
For chained array sections, this optimization is restricted to cases
where all array sections except the last have a constant length 1.
This trivially guarantees that there are no holes in the memory region
that needs to be privatized.
Example:
int c[3][4];
pragma omp parallel reduction(+: c[1:1][1:2])
{ }
Differential Revision: https://reviews.llvm.org/D39136
llvm-svn: 316229
When declaring an entity in the "purview" of a module, it's never a
redeclaration of an entity in the purview of a default module or in no module
("in the global module"). Don't consider those other declarations as possible
redeclaration targets if they're not visible, and reject any cases where we
pick a prior visible declaration that violates this rule.
This reinstates r315251 and r315256, reverted in r315309 and r315308
respectively, tweaked to avoid triggering a linkage calculation when declaring
implicit special members (this exposed our pre-existing issue with typedef
names for linkage changing the linkage of types whose linkage has already been
computed and cached in more cases). A testcase for that regression has been
added in r315366.
llvm-svn: 315379
When declaring an entity in the "purview" of a module, it's never a
redeclaration of an entity in the purview of a default module or in no module
("in the global module"). Don't consider those other declarations as possible
redeclaration targets if they're not visible, and reject any cases where we
pick a prior visible declaration that violates this rule.
llvm-svn: 315251
Previously we may erroneously try to capture locally declared static
variables, which will lead to crash for target-based constructs.
Patch fixes this problem.
llvm-svn: 315076
If the `defaultmap(tofrom:scalar)` clause is specified, the scalars must
be mapped with 'tofrom' modifiers, otherwise they must be captured as
firstprivates.
llvm-svn: 314995
https://reviews.llvm.org/D38371
This patch implements codegen for the combined 'teams distribute" OpenMP pragma and adds regression tests for all its clauses.
llvm-svn: 314905
declaration.
Patch allows using of the `#pragma omp declare target`| `#pragma omp end
declare target` directives inside the structures if we need to mark as
declare target only some static members.
llvm-svn: 314833
directives.
The argument of the `device` clause in target-based executable
directives must be captured to support codegen for the `target`
directives with the `depend` clauses.
llvm-svn: 314686
directives.
If the variable is used in the target-based region but is not found in
any private|mapping clause, then generate implicit firstprivate|map
clauses for these implicitly mapped variables.
llvm-svn: 314205
If the captured variable has some redeclarations we may run into the
situation where the redeclaration is used instead of the canonical
declaration and we may consider this variable as one not captured
before.
llvm-svn: 313880
When the value specified for n in ordered(n) is larger than the number of loops a segmentation fault can occur in one of two ways when attempting to print out a diagnostic for an associated depend(sink : vec):
1) The iteration vector vec contains less than n items
2) The iteration vector vec contains a variable that is not a loop control variable
This patch addresses both of these issues.
Differential Revision: https://reviews.llvm.org/D38049
llvm-svn: 313675
move constructor.
Previously user-defined reduction initializer was considered as an
assignment expression, not as initializer. Fixed this by treating the
initializer expression as an initializer.
llvm-svn: 312638
step>1.
If the loop is a loot with random access iterators and the iteration
construct is represented it += n, then the compiler crashed because of
reusing of the same MaterializedTemporaryExpr around N. Patch fixes it
by using the expression as written, without any special kind of
wrappings.
llvm-svn: 312292
Capturing of the global variables occurs only in target regions. Patch
fixes it and allows capturing of globals in all target executable
directives.
llvm-svn: 312024
Arguments, passed to the outlined function, must have correct address
space info for proper Debug info support. Patch sets global address
space for arguments that are mapped and passed by reference.
Also, cuda-gdb does not handle reference types correctly, so reference
arguments are represented as pointers.
llvm-svn: 310387
Arguments, passed to the outlined function, must have correct address
space info for proper Debug info support. Patch sets global address
space for arguments that are mapped and passed by reference.
Also, cuda-gdb does not handle reference types correctly, so reference
arguments are represented as pointers.
llvm-svn: 310377
Arguments, passed to the outlined function, must have correct address
space info for proper Debug info support. Patch sets global address
space for arguments that are mapped and passed by reference.
Also, cuda-gdb does not handle reference types correctly, so reference
arguments are represented as pointers.
llvm-svn: 310360
Arguments, passed to the outlined function, must have correct address
space info for proper Debug info support. Patch sets global address
space for arguments that are mapped and passed by reference.
Also, cuda-gdb does not handle reference types correctly, so reference
arguments are represented as pointers.
llvm-svn: 310104
If the member declaration is captured in the OMPCapturedExprDecl, we may
loose data-sharing attribute info for this declaration. Patch fixes this
bug.
llvm-svn: 308629
Added checks for the reduction clauses in the taskloop directives:
1. Only addressable items must be used in reduction clauses.
2. Reduction clauses cannot be used with nogroup clauses.
llvm-svn: 307693
Combined directives like 'target parallel' have two captured statements.
Sema has to check the right one from the right direction.
Previously, Sema::IsOpenMPCapturedByRef would return false for mapped
scalars on combined directives. This results in a wrong signature of
the outlined function which triggers an assertion:
void llvm::CallInst::init(llvm::FunctionType *, llvm::Value *, ArrayRef<llvm::Value *>, ArrayRef<OperandBundleDef>, const llvm::Twine &): Assertion `(i >= FTy->getNumParams() || FTy->getParamType(i) == Args[i]->getType()) && "Calling a function with a bad signature!"' failed.
Fixes PR30975 (and PR31985). New function was taken from clang-ykt.
Differential Revision: https://reviews.llvm.org/D34888
llvm-svn: 306956
According to OpenMP 5.0 at least one 'map' or 'use_device_ptr' clause
must be specified for 'target data' construct. Patch adds support for
this feature.
llvm-svn: 304216
If some function template is instantiated during handling of OpenMP
code, currently it may cause crash of compiler because of trying of
capturing variables in non-capturing function scopes. Patch fixes this
bug.
llvm-svn: 301416
https://reviews.llvm.org/D32237
This patch prepares sema with additional fields to support all those composite and combined constructs of OpenMP that include pragma 'distribute' and 'for', such as 'distribute parallel for'. It also extends the regression tests for 'distribute parallel for' and adds a new one.
llvm-svn: 300802
- also replace direct equality checks against the ConstantEvaluated enumerator with isConstantEvaluted(), in anticipation of adding finer granularity to the various ConstantEvaluated contexts and reinstating certain restrictions on where lambda expressions can occur in C++17.
- update the clang tablegen backend that uses these Enumerators, and add the relevant scope where needed.
llvm-svn: 299316
checkNestingOfRegions uses CancelRegion to determine whether cancel and
cancellation point are valid in the given nesting. This leads to unuseful
diagnostics if CancelRegion is invalid. The given test case has produced:
region cannot be closely nested inside 'parallel' region
As a solution, introduce checkCancelRegion and call it first to get the
expected error:
one of 'for', 'parallel', 'sections' or 'taskgroup' is expected
Differential Revision: https://reviews.llvm.org/D30135
llvm-svn: 295808
https://reviews.llvm.org/D29922
This patch adds two fields for use in the implementation of 'distribute parallel for':
The increment expression for the distribute loop. As the chunk assigned to a team is executed by multiple threads within the 'parallel for' region, the increment expression has to correspond to the value returned by the related runtime call (for_static_init).
The upper bound of the innermost loop ('for' in 'distribute parallel for') is not the globalUB expression normally used for pragma 'for' when found in isolation. It is instead the upper bound of the chunk assigned to the team ('distribute' loop). In this way, we prevent teams from executing chunks assigned to other teams.
The use of these two fields can be see in a related explanatory patch:
https://reviews.llvm.org/D29508
llvm-svn: 295497
The thread_limit-clause on the combined directive applies to the
'teams' region of this construct. We modify the ThreadLimitClause
class to capture the clause expression within the 'target' region.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29087
llvm-svn: 293049
The num_teams-clause on the combined directive applies to the
'teams' region of this construct. We modify the NumTeamsClause
class to capture the clause expression within the 'target' region.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29085
llvm-svn: 293048
This patch adds support for codegen of 'target teams' on the host.
This combined directive has two captured statements, one for the
'teams' region, and the other for the 'parallel'.
This target teams region is offloaded using the __tgt_target_teams()
call. The patch sets the number of teams as an argument to
this call.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29084
llvm-svn: 293005
This patch adds support for codegen of 'target teams' on the host.
This combined directive has two captured statements, one for the
'teams' region, and the other for the 'parallel'.
This target teams region is offloaded using the __tgt_target_teams()
call. The patch sets the number of teams as an argument to
this call.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29084
llvm-svn: 293001
The num_threads-clause on the combined directive applies to the
'parallel' region of this construct. We modify the NumThreadsClause
class to capture the clause expression within the 'target' region.
The offload runtime call for 'target parallel' is changed to
__tgt_target_teams() with 1 team and the number of threads set by
this clause or a default if none.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29082
llvm-svn: 292997
The DSAChecker code in SemaOpenMP looks at the captured statement
associated with an OpenMP directive. A combined directive such as
'target parallel' has nested capture statements, which have to be
fully traversed before executing the DSAChecker. This is a patch
to perform the traversal for such combined directives.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29026
llvm-svn: 292794
The if-clause on the combined directive potentially applies to both the
'target' and the 'parallel' regions. Codegen'ing the if-clause on the
combined directive requires additional support because the expression in
the clause must be captured by the 'target' capture statement but not
the 'parallel' capture statement. Note that this situation arises for
other clauses such as num_threads.
The OMPIfClause class inherits OMPClauseWithPreInit to support capturing
of expressions in the clause. A member CaptureRegion is added to
OMPClauseWithPreInit to indicate which captured statement (in this case
'target' but not 'parallel') captures these expressions.
To ensure correct codegen of captured expressions in the presence of
combined 'target' directives, OMPParallelScope was added to 'parallel'
codegen.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D28781
llvm-svn: 292437
This patch adds support for codegen of 'target parallel' on the host.
It is also the first combined directive that requires two or more
captured statements. Support for this functionality is included in
the patch.
A combined directive such as 'target parallel' has two captured
statements, one for the 'target' and the other for the 'parallel'
region. Two captured statements are required because each has
different implicit parameters (see SemaOpenMP.cpp). For example,
the 'parallel' has 'global_tid' and 'bound_tid' while the 'target'
does not. The patch adds support for handling multiple captured
statements based on the combined directive.
When codegen'ing the 'target parallel' directive, the 'target'
outlined function is created using the outer captured statement
and the 'parallel' outlined function is created using the inner
captured statement.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D28753
llvm-svn: 292419
This patch adds support for codegen of 'target parallel' on the host.
It is also the first combined directive that requires two or more
captured statements. Support for this functionality is included in
the patch.
A combined directive such as 'target parallel' has two captured
statements, one for the 'target' and the other for the 'parallel'
region. Two captured statements are required because each has
different implicit parameters (see SemaOpenMP.cpp). For example,
the 'parallel' has 'global_tid' and 'bound_tid' while the 'target'
does not. The patch adds support for handling multiple captured
statements based on the combined directive.
When codegen'ing the 'target parallel' directive, the 'target'
outlined function is created using the outer captured statement
and the 'parallel' outlined function is created using the inner
captured statement.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D28753
llvm-svn: 292374
This flag serves no purpose other than to prevent us walking through a type to
check whether it contains an 'auto' specifier; this duplication of information
is error-prone, does not appear to provide any performance benefit, and will
become less practical once we support C++1z deduced class template types and
eventually constrained types from the Concepts TS.
No functionality change intended.
llvm-svn: 291737
This patch is to implement sema and parsing for 'target teams distribute simd’ pragma.
Differential Revision: https://reviews.llvm.org/D28252
llvm-svn: 291579
This patch is to add support of the 'is_device_ptr' clause with the 'target parallel for' pragma.
Differential Revision: https://reviews.llvm.org/D28255
llvm-svn: 291540
This patch is to add support of the 'is_device_ptr' clause with the 'target parallel for simd' pragma.
Differential Revision: https://reviews.llvm.org/D28402
llvm-svn: 291537
This patch is to implement sema and parsing for 'target teams distribute parallel for simd’ pragma.
Differential Revision: https://reviews.llvm.org/D28202
llvm-svn: 290862
This patch is to implement sema and parsing for 'target teams distribute parallel for’ pragma.
Differential Revision: https://reviews.llvm.org/D28160
llvm-svn: 290725
This patch is to implement sema and parsing for 'target teams distribute' pragma.
Differential Revision: https://reviews.llvm.org/D28015
llvm-svn: 290508
based coverage compilation
Added source location info to captured expression declaration + fixed
source location info for loop based directives.
llvm-svn: 290181
This patch is to add support of the 'is_device_ptr' clause in the 'target parallel' pragma.
Differential Revision: https://reviews.llvm.org/D27821
llvm-svn: 289989
This patch is to implement sema and parsing for 'teams distribute parallel for' pragma.
Differential Revision: https://reviews.llvm.org/D27345
llvm-svn: 289179
This patch is to implement sema and parsing for 'teams distribute parallel for simd' pragma.
Differential Revision: https://reviews.llvm.org/D27084
llvm-svn: 288294
If member expression is used in the task region and the base expression
is a DeclRefExp and the variable used in this ref expression is private,
it should be marked as implicitly firstprivate inside this region. Patch
fixes this issue.
llvm-svn: 288039
This reverts commit r279003 as it breaks some of our buildbots (e.g.
clang-cmake-aarch64-quick, clang-x86_64-linux-selfhost-modules).
The error is in OpenMP/teams_distribute_simd_ast_print.cpp:
clang: /home/buildslave/buildslave/clang-cmake-aarch64-quick/llvm/include/llvm/ADT/DenseMap.h:527:
bool llvm::DenseMapBase<DerivedT, KeyT, ValueT, KeyInfoT, BucketT>::LookupBucketFor(const LookupKeyT&, const BucketT*&) const
[with LookupKeyT = clang::Stmt*; DerivedT = llvm::DenseMap<clang::Stmt*, long unsigned int>;
KeyT = clang::Stmt*; ValueT = long unsigned int;
KeyInfoT = llvm::DenseMapInfo<clang::Stmt*>;
BucketT = llvm::detail::DenseMapPair<clang::Stmt*, long unsigned int>]:
Assertion `!KeyInfoT::isEqual(Val, EmptyKey) && !KeyInfoT::isEqual(Val, TombstoneKey) &&
"Empty/Tombstone value shouldn't be inserted into map!"' failed.
llvm-svn: 279045
This patch is to implement sema and parsing for 'teams distribute simd’ pragma.
This patch is originated by Carlo Bertolli.
Differential Revision: https://reviews.llvm.org/D23528
llvm-svn: 279003
Summary: This patch adds support for the is_device_ptr clause. It expands SEMA to use the mappable expression logic that can only be tested with code generation in place and check conflicts with other data sharing related clauses using the mappable expressions infrastructure.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: caomhin, cfe-commits
Differential Revision: https://reviews.llvm.org/D22788
llvm-svn: 276978
Summary: This patch adds support for the use_device_ptr clause. It includes changes in SEMA that could not be tested without codegen, namely, the use of the first private logic and mappable expressions support.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: caomhin, cfe-commits
Differential Revision: https://reviews.llvm.org/D22691
llvm-svn: 276977
The OpenMP spec mandates that 'a teams construct must be contained within a
target construct'. Currently, this scenario is not diagnosed. This patch is
to add check for orphaned teams construct and issue an error message.
Differential Revision: https://reviews.llvm.org/D22785
llvm-svn: 276726
Clang misdiagnoses atomic captures cases that contains parens.
i.e.
int v, int *p;
#pragma omp atomic capture
{ v = (*p); (*p)++; }
Patch by David S.
Differential Revision: https://reviews.llvm.org/D22487
llvm-svn: 276167
Having the following code pattern will result in incorrect diagnostic
int main() {
int arr[10];
#pragma omp target data map(arr[:])
#pragma omp target map(arr)
{}
}
t.cpp:4:24: error: original storage of expression in data environment is shared
but data environment do not fully contain mapped expression storage
#pragma omp target map(arr)
^~~
t.cpp:3:29: note: used here
#pragma omp target data map(arr[:])
^~~~~~
1 error generated.
Patch by David S.
Differential Revision: https://reviews.llvm.org/D22075
llvm-svn: 275926
This patch adds the check for specifying both simdlen and safelen clauses on the 'distribute simd' or 'distribute parallel for simd' constructs.
Differential Revision: https://reviews.llvm.org/D22384
llvm-svn: 275529
This patch is to implement sema and parsing for 'target parallel for simd' pragma.
Differential Revision: http://reviews.llvm.org/D22096
llvm-svn: 275365
http://reviews.llvm.org/D21904
This patch is similar to the implementation of 'private' clause: it adds a list of private pointers to be used within the target data region to store the device pointers returned by the runtime.
Please refer to the following document for a full description of what the runtime witll return in this case (page 10 and 11):
https://github.com/clang-omp/OffloadingDesign
I am happy to answer any question related to the runtime interface to help reviewing this patch.
llvm-svn: 275271
OpenMP relies on some helper expressions generated during semantic
analysis. But they are required only for codegen and not required in
dependent contexts. Patch removes generation of some of helper
expressions.
llvm-svn: 274745
Summary: This patch is an implementation of sema and parsing for the OpenMP composite pragma 'distribute simd'.
Differential Revision: http://reviews.llvm.org/D22007
llvm-svn: 274604
Summary: This patch is an implementation of sema and parsing for the OpenMP composite pragma 'distribute parallel for simd'.
Differential Revision: http://reviews.llvm.org/D21977
llvm-svn: 274530
[OpenMP] Initial implementation of parse and sema for composite pragma 'distribute parallel for'
This patch is an initial implementation for #distribute parallel for.
The main differences that affect other pragmas are:
The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds.
As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value.
As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound.
llvm-svn: 273884
http://reviews.llvm.org/D21564
This patch is an initial implementation for #distribute parallel for.
The main differences that affect other pragmas are:
The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds.
As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value.
As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound.
llvm-svn: 273705
-Wfor-loop-analysis warnings for a for-loop with a condition variable. In such
a case, the loop condition variable is modified on each iteration of the loop
by definition.
Original commit message:
Rearrange condition handling so that semantic checks on a condition variable
are performed before the other substatements of the construct are parsed,
rather than deferring them until the end. This allows better error recovery
from semantic errors in the condition, improves diagnostic order, and is a
prerequisite for C++17 constexpr if.
llvm-svn: 273600
are performed before the other substatements of the construct are parsed,
rather than deferring them until the end. This allows better error recovery
from semantic errors in the condition, improves diagnostic order, and is a
prerequisite for C++17 constexpr if.
llvm-svn: 273548
instantiation.
Added checks for non-dependent context when trygin to capture
non-constant schedule chunk expression for proper codegen of outlined
functions.
llvm-svn: 272775
These ExprWithCleanups are added for holding a RunCleanupsScope not
for destructor calls; rather, they are for lifetime marks. This requires
ExprWithCleanups to keep a bit to indicate whether it have cleanups with
side effects (e.g. dtor calls).
Differential Revision: http://reviews.llvm.org/D20498
llvm-svn: 272296
Summary:
Target regions require globals to be captured. This patch fixes a bug exposed when that happens in a template function.
Reviewers: hfinkel, arpith-jacob, kkwli0, carlo.bertolli, ABataev
Subscribers: guansong, ABataev, cfe-commits, caomhin, fraggamuffin
Differential Revision: http://reviews.llvm.org/D18110
llvm-svn: 271001
Summary:
The patch contains the parsing and sema support for the `from` clause.
Patch based on the original post by Kelvin Li.
Reviewers: hfinkel, carlo.bertolli, kkwli0, arpith-jacob, ABataev
Subscribers: caomhin, cfe-commits
Differential Revision: http://reviews.llvm.org/D18488
llvm-svn: 270882
Summary:
The patch contains the parsing and sema support for the `to` clause.
Patch based on the original post by Kelvin Li.
Reviewers: carlo.bertolli, hfinkel, kkwli0, arpith-jacob, ABataev
Subscribers: caomhin, cfe-commits
Differential Revision: http://reviews.llvm.org/D18597
llvm-svn: 270880
Summary:
This patch is to add parsing and sema support for `target update` directive. Support for the `to` and `from` clauses will be added by a different patch. This patch also adds support for other clauses that are already implemented upstream and apply to `target update`, e.g. `device` and `if`.
This patch is based on the original post by Kelvin Li.
Reviewers: hfinkel, carlo.bertolli, kkwli0, arpith-jacob, ABataev
Subscribers: caomhin, cfe-commits
Differential Revision: http://reviews.llvm.org/D15944
llvm-svn: 270878
For better performance and to unify code with offloading part we pass
scalar firstprivate values by value, instead of by reference. It will
remove some extra copying operations.
llvm-svn: 269751
Support OpenMP version 4.5 syntax for #pragma omp declare target.
Syntax:
#pragma omp declare target (extended-list) new-line
or
#pragma omp declare target clause[ [,] clause ... ] new-line
Where clause is one of the following:
to(extended-list)
link(list)
Differential Revision: http://reviews.llvm.org/D20011
llvm-svn: 268925
OpenMP 4.5 defines 'taskloop simd' directive, which is combined
directive for 'taskloop' and 'simd' directives. Patch adds initial
codegen support for this directive and its 2 basic clauses 'safelen' and
'simdlen'.
llvm-svn: 267872
Summary:
Implement codegen for the map clause. All the new list items in 4.5 specification are supported.
Fix bug in the generation of array sections that was exposed by some of the map clause tests: for pointer types the offsets have to be calculated from the pointee not the pointer.
Reviewers: hfinkel, kkwli0, carlo.bertolli, arpith-jacob, ABataev
Subscribers: ABataev, cfe-commits, caomhin, fraggamuffin
Differential Revision: http://reviews.llvm.org/D16749
llvm-svn: 267808
Summary:
This patch adds logic to save the components of mappable expressions in the clause that uses it, so that they don't have to be recomputed during codegen. Given that the mappable components are (will be) used in several clauses a new geneneric implementation `OMPMappableExprListClause` is used that extends the existing `OMPVarListClause`.
This patch does not add new tests. The goal is to preserve the existing functionality while storing more info in the clauses.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: cfe-commits, caomhin
Differential Revision: http://reviews.llvm.org/D19382
llvm-svn: 267560
The taskloop construct specifies that the iterations of one or more associated loops will be executed in parallel using OpenMP tasks. The iterations are distributed across tasks created by the construct and scheduled to be executed.
The next code will be generated for the taskloop directive:
#pragma omp taskloop num_tasks(N) lastprivate(j)
for( i=0; i<N*GRAIN*STRIDE-1; i+=STRIDE ) {
int th = omp_get_thread_num();
#pragma omp atomic
counter++;
#pragma omp atomic
th_counter[th]++;
j = i;
}
Generated code:
task = __kmpc_omp_task_alloc(NULL,gtid,1,sizeof(struct
task),sizeof(struct shar),&task_entry);
psh = task->shareds;
psh->pth_counter = &th_counter;
psh->pcounter = &counter;
psh->pj = &j;
task->lb = 0;
task->ub = N*GRAIN*STRIDE-2;
task->st = STRIDE;
__kmpc_taskloop(
NULL, // location
gtid, // gtid
task, // task structure
1, // if clause value
&task->lb, // lower bound
&task->ub, // upper bound
STRIDE, // loop increment
0, // 1 if nogroup specified
2, // schedule type: 0-none, 1-grainsize, 2-num_tasks
N, // schedule value (ignored for type 0)
(void*)&__task_dup_entry // tasks duplication routine
);
llvm-svn: 267395
If loop control variable for simd-based directives is explicitly marked
as linear/lastprivate in clauses, codegen for such construct would
crash. Patch fixes this problem.
llvm-svn: 267101
If the untied clause is present on a task construct, any thread in the
team can resume the task region after a suspension. Patch adds proper
codegen for untied tasks.
llvm-svn: 266853
If the untied clause is present on a task construct, any thread in the
team can resume the task region after a suspension. Patch adds proper
codegen for untied tasks.
llvm-svn: 266754
If the untied clause is present on a task construct, any thread in the team can resume the task region after a suspension. Patch adds proper codegen for untied tasks.
llvm-svn: 266722
The linear clause declares one or more list items to be private to a SIMD lane and to have a linear relationship with respect to the iteration space of a loop.
'linear' '(' <linear-list> [ ':' <linear-step> ] ')'
When a linear-step expression is specified in a linear clause it must be
either a constant integer expression or an integer-typed parameter that is specified in a uniform clause on the directive.
The special this pointer can be used as if was one of the arguments to the function in any of the linear, aligned, or uniform clauses.
llvm-svn: 266056
The aligned clause declares that the object to which each list item points is aligned to the number of bytes expressed in the optional parameter of the aligned clause.
'aligned' '(' <argument-list> [ ':' <alignment> ] ')'
The optional parameter of the aligned clause, alignment, must be a constant positive integer expression. If no optional parameter is specified, implementation-defined default alignments for SIMD instructions on the target platforms are assumed.
The special this pointer can be used as if was one of the arguments to the function in any of the linear, aligned, or uniform clauses.
llvm-svn: 266052
OpenMP 4.0 defines clause 'uniform' in 'declare simd' directive:
'uniform' '(' <argument-list> ')'
The uniform clause declares one or more arguments to have an invariant value for all concurrent invocations of the function in the execution of a single SIMD loop.
The special this pointer can be used as if was one of the arguments to the function in any of the linear, aligned, or uniform clauses.
llvm-svn: 266041
construct.
OpenMP 4.0 defines '#pragma omp declare simd' construct that may have
associated 'simdlen' clause with constant positive expression as an
argument:
simdlen(<const_expr>)
Patch adds parsin and semantic analysis for simdlen clause.
llvm-svn: 265668
Add parsing, sema analysis for 'declare target' construct for OpenMP 4.0
(4.5 support will be added in separate patch).
The declare target directive specifies that variables, functions (C, C++
and Fortran), and subroutines (Fortran) are mapped to a device. The declare
target directive is a declarative directive. In Clang declare target is
implemented as implicit attribute for the declaration.
The syntax of the declare target directive is as follows:
#pragma omp declare target
declarations-definition-seq
#pragma omp end declare target
Based on patch from Michael Wong http://reviews.llvm.org/D15321
llvm-svn: 265530
directives.
OpenMP 4.5 allows privatization of non-static data members in non-static
member functions. Patch allows to use and implicit privatization of data
members used as counters in loop-based directives.
llvm-svn: 265121
Initial parsing/sema/serialization/deserialization support for '#pragma
omp declare simd' directive.
The 'declare simd' construct can be applied to a function to enable the
creation of one or more versions that can process multiple arguments
using SIMD instructions from a single invocation from a SIMD loop.
If the function has any declarations, then the declare simd construct
for any declaration that has one must be equivalent to the one specified
for the definition. Otherwise, the result is unspecified.
This pragma can be applied many times to the same declaration.
Internally this pragma is represented as an attribute. But we need special processing for this pragma because it must be used before function declaration, this directive is applied to.
Differential Revision: http://reviews.llvm.org/D10599
llvm-svn: 264853
For better support of some specific GNU extensions some extra
transformation of AST nodes were introduced. These transformations are
very hard to handle. The code is improved in handling of these
extensions by using captured expressions construct.
llvm-svn: 264709
This patch implements the following aspects:
It extends sema to check that a variable is not reference in both a map clause and firstprivate or private. This is needed to ensure correct functioning at codegen level, apart from being useful for the user.
It implements firstprivate for target in codegen. The implementation applies to both host and nvptx devices.
It adds regression tests for codegen of firstprivate, host and device side when using the host as device, and nvptx side.
Please note that the regression test for nvptx codegen is missing VLAs. This is because VLAs currently require saving and restoring the stack which appears not to be a supported operation by nvptx backend.
It adds a check in sema regression tests for target map, firstprivate, and private clauses.
http://reviews.llvm.org/D18203
llvm-svn: 263837
OpenMP 4.5 allows privatization of non-static data members in non-static
member functions. This patch adds support of private data members in
'copyprivate' clauses.
llvm-svn: 263706
OpenMP 4.0 allows to define custom reduction operations using '#pragma
omp declare reduction' construct. Patch allows to use this custom
defined reduction operations in 'reduction' clauses.
llvm-svn: 263701
Summary: In some cases it can be proved statically that multidimensional array section refer to contiguous storage and can therefore be allowed in a map clause. This patch adds support for those cases in SEMA.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: cfe-commits, fraggamuffin, caomhin
Differential Revision: http://reviews.llvm.org/D17547
llvm-svn: 263019
Alexey Bataev