svenvh created this revision.
svenvh added a reviewer: Anastasia.
svenvh added a project: clang.
Herald added a subscriber: yaxunl.
svenvh requested review of this revision.
Herald added a subscriber: cfe-commits.
Add the builtin functions brought by the cl_khr_subgroup_extended_types
extension to `-fdeclare-opencl-builtins`.
Repository:
rG LLVM Github Monorepo
https://reviews.llvm.org/D96279
Files:
clang/lib/Sema/OpenCLBuiltins.td
clang/test/SemaOpenCL/fdeclare-opencl-builtins.cl
Index: clang/test/SemaOpenCL/fdeclare-opencl-builtins.cl
===================================================================
--- clang/test/SemaOpenCL/fdeclare-opencl-builtins.cl
+++ clang/test/SemaOpenCL/fdeclare-opencl-builtins.cl
@@ -39,6 +39,7 @@
// Enable extensions that are enabled in opencl-c-base.h.
#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200)
+#define cl_khr_subgroup_extended_types 1
#define cl_khr_subgroup_ballot 1
#define cl_khr_subgroup_non_uniform_arithmetic 1
#define cl_khr_subgroup_clustered_reduce 1
@@ -157,15 +158,18 @@
#endif
}
-kernel void extended_subgroup(global uint4 *out, global int *scalar) {
+kernel void extended_subgroup(global uint4 *out, global int *scalar, global
char2 *c2) {
out[0] = get_sub_group_eq_mask();
scalar[0] = sub_group_non_uniform_scan_inclusive_or(3);
scalar[1] = sub_group_clustered_reduce_logical_xor(2, 4);
+ *c2 = sub_group_broadcast(*c2, 2);
#if __OPENCL_C_VERSION__ < CL_VERSION_2_0 && !defined(__OPENCL_CPP_VERSION__)
- // expected-error@-4{{implicit declaration of function
'get_sub_group_eq_mask' is invalid in OpenCL}}
- // expected-error@-5{{implicit conversion changes signedness}}
- // expected-error@-5{{implicit declaration of function
'sub_group_non_uniform_scan_inclusive_or' is invalid in OpenCL}}
- // expected-error@-5{{implicit declaration of function
'sub_group_clustered_reduce_logical_xor' is invalid in OpenCL}}
+ // expected-error@-5{{implicit declaration of function
'get_sub_group_eq_mask' is invalid in OpenCL}}
+ // expected-error@-6{{implicit conversion changes signedness}}
+ // expected-error@-6{{implicit declaration of function
'sub_group_non_uniform_scan_inclusive_or' is invalid in OpenCL}}
+ // expected-error@-6{{implicit declaration of function
'sub_group_clustered_reduce_logical_xor' is invalid in OpenCL}}
+ // expected-error@-6{{implicit declaration of function 'sub_group_broadcast'
is invalid in OpenCL}}
+ // expected-error@-7{{implicit conversion loses integer precision}}
#endif
}
Index: clang/lib/Sema/OpenCLBuiltins.td
===================================================================
--- clang/lib/Sema/OpenCLBuiltins.td
+++ clang/lib/Sema/OpenCLBuiltins.td
@@ -53,6 +53,7 @@
// FunctionExtension definitions.
def FuncExtNone : FunctionExtension<"">;
def FuncExtKhrSubgroups :
FunctionExtension<"cl_khr_subgroups">;
+def FuncExtKhrSubgroupExtendedTypes :
FunctionExtension<"cl_khr_subgroup_extended_types">;
def FuncExtKhrSubgroupNonUniformVote :
FunctionExtension<"cl_khr_subgroup_non_uniform_vote">;
def FuncExtKhrSubgroupBallot :
FunctionExtension<"cl_khr_subgroup_ballot">;
def FuncExtKhrSubgroupNonUniformArithmetic:
FunctionExtension<"cl_khr_subgroup_non_uniform_arithmetic">;
@@ -381,6 +382,9 @@
def FGenTypeN : GenericType<"FGenTypeN", TLFloat, VecAndScalar>;
// (u)int, (u)long, and all floats
def IntLongFloatGenType1 : GenericType<"IntLongFloatGenType1",
TLIntLongFloats, Vec1>;
+// (u)char and (u)short
+def CharShortGenType1 : GenericType<"CharShortGenType1",
+ TypeList<[Char, UChar, Short, UShort]>, Vec1>;
// GenType definitions for every single base type (e.g. fp32 only).
// Names are like: GenTypeFloatVecAndScalar.
@@ -1519,7 +1523,19 @@
// OpenCL Extension v3.0 s38 - Extended Subgroup Functions
// Section 38.4.1 - cl_khr_subgroup_extended_types
-// TODO
+let Extension = FuncExtKhrSubgroupExtendedTypes in {
+ // For sub_group_broadcast, add scalar char, uchar, short, and ushort
support,
+ def : Builtin<"sub_group_broadcast", [CharShortGenType1, CharShortGenType1,
UInt], Attr.Convergent>;
+ // gentype may additionally be one of the supported built-in vector data
types.
+ def : Builtin<"sub_group_broadcast", [AGenTypeNNoScalar, AGenTypeNNoScalar,
UInt], Attr.Convergent>;
+
+ foreach name = ["sub_group_reduce_", "sub_group_scan_exclusive_",
+ "sub_group_scan_inclusive_"] in {
+ foreach op = ["add", "min", "max"] in {
+ def : Builtin<name # op, [CharShortGenType1, CharShortGenType1],
Attr.Convergent>;
+ }
+ }
+}
// Section 38.5.1 - cl_khr_subgroup_non_uniform_vote
let Extension = FuncExtKhrSubgroupNonUniformVote in {
Index: clang/test/SemaOpenCL/fdeclare-opencl-builtins.cl
===================================================================
--- clang/test/SemaOpenCL/fdeclare-opencl-builtins.cl
+++ clang/test/SemaOpenCL/fdeclare-opencl-builtins.cl
@@ -39,6 +39,7 @@
// Enable extensions that are enabled in opencl-c-base.h.
#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200)
+#define cl_khr_subgroup_extended_types 1
#define cl_khr_subgroup_ballot 1
#define cl_khr_subgroup_non_uniform_arithmetic 1
#define cl_khr_subgroup_clustered_reduce 1
@@ -157,15 +158,18 @@
#endif
}
-kernel void extended_subgroup(global uint4 *out, global int *scalar) {
+kernel void extended_subgroup(global uint4 *out, global int *scalar, global char2 *c2) {
out[0] = get_sub_group_eq_mask();
scalar[0] = sub_group_non_uniform_scan_inclusive_or(3);
scalar[1] = sub_group_clustered_reduce_logical_xor(2, 4);
+ *c2 = sub_group_broadcast(*c2, 2);
#if __OPENCL_C_VERSION__ < CL_VERSION_2_0 && !defined(__OPENCL_CPP_VERSION__)
- // expected-error@-4{{implicit declaration of function 'get_sub_group_eq_mask' is invalid in OpenCL}}
- // expected-error@-5{{implicit conversion changes signedness}}
- // expected-error@-5{{implicit declaration of function 'sub_group_non_uniform_scan_inclusive_or' is invalid in OpenCL}}
- // expected-error@-5{{implicit declaration of function 'sub_group_clustered_reduce_logical_xor' is invalid in OpenCL}}
+ // expected-error@-5{{implicit declaration of function 'get_sub_group_eq_mask' is invalid in OpenCL}}
+ // expected-error@-6{{implicit conversion changes signedness}}
+ // expected-error@-6{{implicit declaration of function 'sub_group_non_uniform_scan_inclusive_or' is invalid in OpenCL}}
+ // expected-error@-6{{implicit declaration of function 'sub_group_clustered_reduce_logical_xor' is invalid in OpenCL}}
+ // expected-error@-6{{implicit declaration of function 'sub_group_broadcast' is invalid in OpenCL}}
+ // expected-error@-7{{implicit conversion loses integer precision}}
#endif
}
Index: clang/lib/Sema/OpenCLBuiltins.td
===================================================================
--- clang/lib/Sema/OpenCLBuiltins.td
+++ clang/lib/Sema/OpenCLBuiltins.td
@@ -53,6 +53,7 @@
// FunctionExtension definitions.
def FuncExtNone : FunctionExtension<"">;
def FuncExtKhrSubgroups : FunctionExtension<"cl_khr_subgroups">;
+def FuncExtKhrSubgroupExtendedTypes : FunctionExtension<"cl_khr_subgroup_extended_types">;
def FuncExtKhrSubgroupNonUniformVote : FunctionExtension<"cl_khr_subgroup_non_uniform_vote">;
def FuncExtKhrSubgroupBallot : FunctionExtension<"cl_khr_subgroup_ballot">;
def FuncExtKhrSubgroupNonUniformArithmetic: FunctionExtension<"cl_khr_subgroup_non_uniform_arithmetic">;
@@ -381,6 +382,9 @@
def FGenTypeN : GenericType<"FGenTypeN", TLFloat, VecAndScalar>;
// (u)int, (u)long, and all floats
def IntLongFloatGenType1 : GenericType<"IntLongFloatGenType1", TLIntLongFloats, Vec1>;
+// (u)char and (u)short
+def CharShortGenType1 : GenericType<"CharShortGenType1",
+ TypeList<[Char, UChar, Short, UShort]>, Vec1>;
// GenType definitions for every single base type (e.g. fp32 only).
// Names are like: GenTypeFloatVecAndScalar.
@@ -1519,7 +1523,19 @@
// OpenCL Extension v3.0 s38 - Extended Subgroup Functions
// Section 38.4.1 - cl_khr_subgroup_extended_types
-// TODO
+let Extension = FuncExtKhrSubgroupExtendedTypes in {
+ // For sub_group_broadcast, add scalar char, uchar, short, and ushort support,
+ def : Builtin<"sub_group_broadcast", [CharShortGenType1, CharShortGenType1, UInt], Attr.Convergent>;
+ // gentype may additionally be one of the supported built-in vector data types.
+ def : Builtin<"sub_group_broadcast", [AGenTypeNNoScalar, AGenTypeNNoScalar, UInt], Attr.Convergent>;
+
+ foreach name = ["sub_group_reduce_", "sub_group_scan_exclusive_",
+ "sub_group_scan_inclusive_"] in {
+ foreach op = ["add", "min", "max"] in {
+ def : Builtin<name # op, [CharShortGenType1, CharShortGenType1], Attr.Convergent>;
+ }
+ }
+}
// Section 38.5.1 - cl_khr_subgroup_non_uniform_vote
let Extension = FuncExtKhrSubgroupNonUniformVote in {
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
