8fb42300a0
A SYCL kernel entry point function is a non-member function or a static
member function declared with the `sycl_kernel_entry_point` attribute.
Such functions define a pattern for an offload kernel entry point
function to be generated to enable execution of a SYCL kernel on a
device. A SYCL library implementation orchestrates the invocation of
these functions with corresponding SYCL kernel arguments in response to
calls to SYCL kernel invocation functions specified by the SYCL 2020
specification.
The offload kernel entry point function (sometimes referred to as the
SYCL kernel caller function) is generated from the SYCL kernel entry
point function by a transformation of the function parameters followed
by a transformation of the function body to replace references to the
original parameters with references to the transformed ones. Exactly how
parameters are transformed will be explained in a future change that
implements non-trivial transformations. For now, it suffices to state
that a given parameter of the SYCL kernel entry point function may be
transformed to multiple parameters of the offload kernel entry point as
needed to satisfy offload kernel argument passing requirements.
Parameters that are decomposed in this way are reconstituted as local
variables in the body of the generated offload kernel entry point
function.
For example, given the following SYCL kernel entry point function
definition:
```
template<typename KernelNameType, typename KernelType>
[[clang::sycl_kernel_entry_point(KernelNameType)]]
void sycl_kernel_entry_point(KernelType kernel) {
kernel();
}
```
and the following call:
```
struct Kernel {
int dm1;
int dm2;
void operator()() const;
};
Kernel k;
sycl_kernel_entry_point<class kernel_name>(k);
```
the corresponding offload kernel entry point function that is generated
might look as follows (assuming `Kernel` is a type that requires
decomposition):
```
void offload_kernel_entry_point_for_kernel_name(int dm1, int dm2) {
Kernel kernel{dm1, dm2};
kernel();
}
```
Other details of the generated offload kernel entry point function, such
as its name and calling convention, are implementation details that need
not be reflected in the AST and may differ across target devices. For
that reason, only the transformation described above is represented in
the AST; other details will be filled in during code generation.
These transformations are represented using new AST nodes introduced
with this change. `OutlinedFunctionDecl` holds a sequence of
`ImplicitParamDecl` nodes and a sequence of statement nodes that
correspond to the transformed parameters and function body.
`SYCLKernelCallStmt` wraps the original function body and associates it
with an `OutlinedFunctionDecl` instance. For the example above, the AST
generated for the `sycl_kernel_entry_point<kernel_name>` specialization
would look as follows:
```
FunctionDecl 'sycl_kernel_entry_point<kernel_name>(Kernel)'
TemplateArgument type 'kernel_name'
TemplateArgument type 'Kernel'
ParmVarDecl kernel 'Kernel'
SYCLKernelCallStmt
CompoundStmt
<original statements>
OutlinedFunctionDecl
ImplicitParamDecl 'dm1' 'int'
ImplicitParamDecl 'dm2' 'int'
CompoundStmt
VarDecl 'kernel' 'Kernel'
<initialization of 'kernel' with 'dm1' and 'dm2'>
<transformed statements with redirected references of 'kernel'>
```
Any ODR-use of the SYCL kernel entry point function will (with future
changes) suffice for the offload kernel entry point to be emitted. An
actual call to the SYCL kernel entry point function will result in a
call to the function. However, evaluation of a `SYCLKernelCallStmt`
statement is a no-op, so such calls will have no effect other than to
trigger emission of the offload kernel entry point.
Additionally, as a related change inspired by code review feedback,
these changes disallow use of the `sycl_kernel_entry_point` attribute
with functions defined with a _function-try-block_. The SYCL 2020
specification prohibits the use of C++ exceptions in device functions.
Even if exceptions were not prohibited, it is unclear what the semantics
would be for an exception that escapes the SYCL kernel entry point
function; the boundary between host and device code could be an implicit
noexcept boundary that results in program termination if violated, or
the exception could perhaps be propagated to host code via the SYCL
library. Pending support for C++ exceptions in device code and clear
semantics for handling them at the host-device boundary, this change
makes use of the `sycl_kernel_entry_point` attribute with a function
defined with a _function-try-block_ an error.
158 lines
6.5 KiB
C++
158 lines
6.5 KiB
C++
// Tests without serialization:
|
|
// RUN: %clang_cc1 -std=c++17 -triple x86_64-unknown-unknown -fsycl-is-device \
|
|
// RUN: -ast-dump %s \
|
|
// RUN: | FileCheck --match-full-lines %s
|
|
// RUN: %clang_cc1 -std=c++17 -triple x86_64-unknown-unknown -fsycl-is-host \
|
|
// RUN: -ast-dump %s \
|
|
// RUN: | FileCheck --match-full-lines %s
|
|
//
|
|
// Tests with serialization:
|
|
// RUN: %clang_cc1 -std=c++17 -triple x86_64-unknown-unknown -fsycl-is-device \
|
|
// RUN: -emit-pch -o %t %s
|
|
// RUN: %clang_cc1 -x c++ -std=c++17 -triple x86_64-unknown-unknown -fsycl-is-device \
|
|
// RUN: -include-pch %t -ast-dump-all /dev/null \
|
|
// RUN: | sed -e "s/ <undeserialized declarations>//" -e "s/ imported//" \
|
|
// RUN: | FileCheck --match-full-lines %s
|
|
// RUN: %clang_cc1 -std=c++17 -triple x86_64-unknown-unknown -fsycl-is-host \
|
|
// RUN: -emit-pch -o %t %s
|
|
// RUN: %clang_cc1 -x c++ -std=c++17 -triple x86_64-unknown-unknown -fsycl-is-host \
|
|
// RUN: -include-pch %t -ast-dump-all /dev/null \
|
|
// RUN: | sed -e "s/ <undeserialized declarations>//" -e "s/ imported//" \
|
|
// RUN: | FileCheck --match-full-lines %s
|
|
|
|
// These tests validate the AST produced for functions declared with the
|
|
// sycl_kernel_entry_point attribute.
|
|
|
|
// CHECK: TranslationUnitDecl {{.*}}
|
|
|
|
// A unique kernel name type is required for each declared kernel entry point.
|
|
template<int, int=0> struct KN;
|
|
|
|
__attribute__((sycl_kernel_entry_point(KN<1>)))
|
|
void skep1() {
|
|
}
|
|
// CHECK: |-FunctionDecl {{.*}} skep1 'void ()'
|
|
// CHECK: | `-SYCLKernelEntryPointAttr {{.*}} KN<1>
|
|
|
|
using KN2 = KN<2>;
|
|
__attribute__((sycl_kernel_entry_point(KN2)))
|
|
void skep2() {
|
|
}
|
|
// CHECK: |-FunctionDecl {{.*}} skep2 'void ()'
|
|
// CHECK: | `-SYCLKernelEntryPointAttr {{.*}} KN2
|
|
|
|
template<int I> using KNT = KN<I>;
|
|
__attribute__((sycl_kernel_entry_point(KNT<3>)))
|
|
void skep3() {
|
|
}
|
|
// CHECK: |-FunctionDecl {{.*}} skep3 'void ()'
|
|
// CHECK: | `-SYCLKernelEntryPointAttr {{.*}} KNT<3>
|
|
|
|
template<typename KNT, typename F>
|
|
[[clang::sycl_kernel_entry_point(KNT)]]
|
|
void skep4(F f) {
|
|
f();
|
|
}
|
|
// CHECK: |-FunctionTemplateDecl {{.*}} skep4
|
|
// CHECK-NEXT: | |-TemplateTypeParmDecl {{.*}} KNT
|
|
// CHECK-NEXT: | |-TemplateTypeParmDecl {{.*}} F
|
|
// CHECK-NEXT: | |-FunctionDecl {{.*}} skep4 'void (F)'
|
|
// CHECK: | | `-SYCLKernelEntryPointAttr {{.*}} KNT
|
|
|
|
void test_skep4() {
|
|
skep4<KNT<4>>([]{});
|
|
}
|
|
// CHECK: | `-FunctionDecl {{.*}} used skep4 'void ((lambda at {{.*}}))' implicit_instantiation
|
|
// CHECK-NEXT: | |-TemplateArgument type 'KN<4>'
|
|
// CHECK: | |-TemplateArgument type '(lambda at {{.*}})'
|
|
// CHECK: | `-SYCLKernelEntryPointAttr {{.*}} struct KN<4>
|
|
// CHECK-NEXT: |-FunctionDecl {{.*}} test_skep4 'void ()'
|
|
|
|
template<typename KNT, typename T>
|
|
[[clang::sycl_kernel_entry_point(KNT)]]
|
|
void skep5(T) {
|
|
}
|
|
// CHECK: |-FunctionTemplateDecl {{.*}} skep5
|
|
// CHECK-NEXT: | |-TemplateTypeParmDecl {{.*}} KNT
|
|
// CHECK-NEXT: | |-TemplateTypeParmDecl {{.*}} T
|
|
// CHECK-NEXT: | |-FunctionDecl {{.*}} skep5 'void (T)'
|
|
// CHECK: | | `-SYCLKernelEntryPointAttr {{.*}} KNT
|
|
|
|
// Checks for the explicit template instantiation declaration below.
|
|
// CHECK: | `-FunctionDecl {{.*}} skep5 'void (int)' explicit_instantiation_definition
|
|
// CHECK-NEXT: | |-TemplateArgument type 'KN<5, 4>'
|
|
// CHECK: | |-TemplateArgument type 'int'
|
|
// CHECK: | `-SYCLKernelEntryPointAttr {{.*}} KN<5, 4>
|
|
|
|
// FIXME: C++23 [temp.expl.spec]p12 states:
|
|
// FIXME: ... Similarly, attributes appearing in the declaration of a template
|
|
// FIXME: have no effect on an explicit specialization of that template.
|
|
// FIXME: Clang currently instantiates and propagates attributes from a function
|
|
// FIXME: template to its explicit specializations resulting in the following
|
|
// FIXME: explicit specialization having an attribute incorrectly attached.
|
|
template<>
|
|
void skep5<KN<5,1>>(short) {
|
|
}
|
|
// CHECK: |-FunctionDecl {{.*}} prev {{.*}} skep5 'void (short)' explicit_specialization
|
|
// CHECK-NEXT: | |-TemplateArgument type 'KN<5, 1>'
|
|
// CHECK: | |-TemplateArgument type 'short'
|
|
// CHECK: | `-SYCLKernelEntryPointAttr {{.*}} Inherited struct KN<5, 1>
|
|
|
|
template<>
|
|
[[clang::sycl_kernel_entry_point(KN<5,2>)]]
|
|
void skep5<KN<5,2>>(long) {
|
|
}
|
|
// CHECK: |-FunctionDecl {{.*}} prev {{.*}} skep5 'void (long)' explicit_specialization
|
|
// CHECK-NEXT: | |-TemplateArgument type 'KN<5, 2>'
|
|
// CHECK: | |-TemplateArgument type 'long'
|
|
// CHECK: | `-SYCLKernelEntryPointAttr {{.*}} KN<5, 2>
|
|
|
|
// FIXME: C++23 [temp.expl.spec]p12 states:
|
|
// FIXME: ... Similarly, attributes appearing in the declaration of a template
|
|
// FIXME: have no effect on an explicit specialization of that template.
|
|
// FIXME: Clang currently instantiates a function template specialization from
|
|
// FIXME: the function template declaration and links it as a previous
|
|
// FIXME: declaration of an explicit specialization. The instantiated
|
|
// FIXME: declaration includes attributes instantiated from the function
|
|
// FIXME: template declaration. When the instantiated declaration and the
|
|
// FIXME: explicit specialization both specify a sycl_kernel_entry_point
|
|
// FIXME: attribute with different kernel name types, a spurious diagnostic
|
|
// FIXME: is issued. The following test case is incorrectly diagnosed as
|
|
// FIXME: having conflicting kernel name types (KN<5,3> vs the incorrectly
|
|
// FIXME: inherited KN<5,-1>).
|
|
#if 0
|
|
template<>
|
|
[[clang::sycl_kernel_entry_point(KN<5,3>)]]
|
|
void skep5<KN<5,-1>>(long long) {
|
|
}
|
|
// FIXME-CHECK: |-FunctionDecl {{.*}} prev {{.*}} skep5 'void (long long)' explicit_specialization
|
|
// FIXME-CHECK-NEXT: | |-TemplateArgument type 'KN<5, -1>'
|
|
// FIXME-CHECK: | |-TemplateArgument type 'long long'
|
|
// FIXME-CHECK: | `-SYCLKernelEntryPointAttr {{.*}} KN<5, 3>
|
|
#endif
|
|
|
|
template void skep5<KN<5,4>>(int);
|
|
// Checks are located with the primary template declaration above.
|
|
|
|
// Ensure that matching attributes from multiple declarations are ok.
|
|
[[clang::sycl_kernel_entry_point(KN<6>)]]
|
|
void skep6();
|
|
[[clang::sycl_kernel_entry_point(KN<6>)]]
|
|
void skep6() {
|
|
}
|
|
// CHECK: |-FunctionDecl {{.*}} skep6 'void ()'
|
|
// CHECK-NEXT: | `-SYCLKernelEntryPointAttr {{.*}} KN<6>
|
|
// CHECK-NEXT: |-FunctionDecl {{.*}} prev {{.*}} skep6 'void ()'
|
|
// CHECK: | `-SYCLKernelEntryPointAttr {{.*}} KN<6>
|
|
|
|
// Ensure that matching attributes from the same declaration are ok.
|
|
[[clang::sycl_kernel_entry_point(KN<7>), clang::sycl_kernel_entry_point(KN<7>)]]
|
|
void skep7() {
|
|
}
|
|
// CHECK: |-FunctionDecl {{.*}} skep7 'void ()'
|
|
// CHECK: | |-SYCLKernelEntryPointAttr {{.*}} KN<7>
|
|
// CHECK-NEXT: | `-SYCLKernelEntryPointAttr {{.*}} KN<7>
|
|
|
|
void the_end() {}
|
|
// CHECK: `-FunctionDecl {{.*}} the_end 'void ()'
|