summaryrefslogtreecommitdiffstats
path: root/clang/test/CodeGenOpenCL
diff options
context:
space:
mode:
Diffstat (limited to 'clang/test/CodeGenOpenCL')
-rw-r--r--clang/test/CodeGenOpenCL/amdgpu-features.cl12
-rw-r--r--clang/test/CodeGenOpenCL/builtins-amdgcn-dl-insts-err.cl25
-rw-r--r--clang/test/CodeGenOpenCL/builtins-amdgcn-dl-insts.cl36
3 files changed, 73 insertions, 0 deletions
diff --git a/clang/test/CodeGenOpenCL/amdgpu-features.cl b/clang/test/CodeGenOpenCL/amdgpu-features.cl
new file mode 100644
index 00000000000..a1815cedcd4
--- /dev/null
+++ b/clang/test/CodeGenOpenCL/amdgpu-features.cl
@@ -0,0 +1,12 @@
+// REQUIRES: amdgpu-registered-target
+
+// Check that appropriate features are defined for every supported AMDGPU
+// "-target" and "-mcpu" options.
+
+// RUN: %clang_cc1 -triple amdgcn -target-cpu gfx904 -S -emit-llvm -o - %s | FileCheck --check-prefix=GFX904 %s
+// RUN: %clang_cc1 -triple amdgcn -target-cpu gfx906 -S -emit-llvm -o - %s | FileCheck --check-prefix=GFX906 %s
+
+// GFX904: "target-features"="+16-bit-insts,+dpp,+fp32-denormals,+fp64-fp16-denormals,+gfx9-insts,+s-memrealtime"
+// GFX906: "target-features"="+16-bit-insts,+dl-insts,+dpp,+fp32-denormals,+fp64-fp16-denormals,+gfx9-insts,+s-memrealtime"
+
+kernel void test() {}
diff --git a/clang/test/CodeGenOpenCL/builtins-amdgcn-dl-insts-err.cl b/clang/test/CodeGenOpenCL/builtins-amdgcn-dl-insts-err.cl
new file mode 100644
index 00000000000..bfc811e2955
--- /dev/null
+++ b/clang/test/CodeGenOpenCL/builtins-amdgcn-dl-insts-err.cl
@@ -0,0 +1,25 @@
+// REQUIRES: amdgpu-registered-target
+
+// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-cpu gfx900 -verify -S -emit-llvm -o - %s
+
+typedef unsigned int uint;
+typedef half __attribute__((ext_vector_type(2))) half2;
+typedef short __attribute__((ext_vector_type(2))) short2;
+typedef unsigned short __attribute__((ext_vector_type(2))) ushort2;
+
+kernel void builtins_amdgcn_dl_insts_err(
+ global float *fOut, global int *siOut, global uint *uiOut,
+ half2 v2hA, half2 v2hB, float fC,
+ short2 v2ssA, short2 v2ssB, int siA, int siB, int siC,
+ ushort2 v2usA, ushort2 v2usB, uint uiA, uint uiB, uint uiC) {
+ fOut[0] = __builtin_amdgcn_fdot2(v2hA, v2hB, fC); // expected-error {{'__builtin_amdgcn_fdot2' needs target feature dl-insts}}
+
+ siOut[0] = __builtin_amdgcn_sdot2(v2ssA, v2ssB, siC); // expected-error {{'__builtin_amdgcn_sdot2' needs target feature dl-insts}}
+ uiOut[0] = __builtin_amdgcn_udot2(v2usA, v2usB, uiC); // expected-error {{'__builtin_amdgcn_udot2' needs target feature dl-insts}}
+
+ siOut[1] = __builtin_amdgcn_sdot4(siA, siB, siC); // expected-error {{'__builtin_amdgcn_sdot4' needs target feature dl-insts}}
+ uiOut[1] = __builtin_amdgcn_udot4(uiA, uiB, uiC); // expected-error {{'__builtin_amdgcn_udot4' needs target feature dl-insts}}
+
+ siOut[2] = __builtin_amdgcn_sdot8(siA, siB, siC); // expected-error {{'__builtin_amdgcn_sdot8' needs target feature dl-insts}}
+ uiOut[2] = __builtin_amdgcn_udot8(uiA, uiB, uiC); // expected-error {{'__builtin_amdgcn_udot8' needs target feature dl-insts}}
+}
diff --git a/clang/test/CodeGenOpenCL/builtins-amdgcn-dl-insts.cl b/clang/test/CodeGenOpenCL/builtins-amdgcn-dl-insts.cl
new file mode 100644
index 00000000000..c407bd6fe15
--- /dev/null
+++ b/clang/test/CodeGenOpenCL/builtins-amdgcn-dl-insts.cl
@@ -0,0 +1,36 @@
+// REQUIRES: amdgpu-registered-target
+
+// RUN: %clang_cc1 -triple amdgcn-unknown-unknown -target-cpu gfx906 -S -emit-llvm -o - %s | FileCheck %s
+
+typedef unsigned int uint;
+typedef half __attribute__((ext_vector_type(2))) half2;
+typedef short __attribute__((ext_vector_type(2))) short2;
+typedef unsigned short __attribute__((ext_vector_type(2))) ushort2;
+
+// CHECK-LABEL: @builtins_amdgcn_dl_insts
+// CHECK: call float @llvm.amdgcn.fdot2
+
+// CHECK: call i32 @llvm.amdgcn.sdot2
+// CHECK: call i32 @llvm.amdgcn.udot2
+
+// CHECK: call i32 @llvm.amdgcn.sdot4
+// CHECK: call i32 @llvm.amdgcn.udot4
+
+// CHECK: call i32 @llvm.amdgcn.sdot8
+// CHECK: call i32 @llvm.amdgcn.udot8
+kernel void builtins_amdgcn_dl_insts(
+ global float *fOut, global int *siOut, global uint *uiOut,
+ half2 v2hA, half2 v2hB, float fC,
+ short2 v2ssA, short2 v2ssB, int siA, int siB, int siC,
+ ushort2 v2usA, ushort2 v2usB, uint uiA, uint uiB, uint uiC) {
+ fOut[0] = __builtin_amdgcn_fdot2(v2hA, v2hB, fC);
+
+ siOut[0] = __builtin_amdgcn_sdot2(v2ssA, v2ssB, siC);
+ uiOut[0] = __builtin_amdgcn_udot2(v2usA, v2usB, uiC);
+
+ siOut[1] = __builtin_amdgcn_sdot4(siA, siB, siC);
+ uiOut[1] = __builtin_amdgcn_udot4(uiA, uiB, uiC);
+
+ siOut[2] = __builtin_amdgcn_sdot8(siA, siB, siC);
+ uiOut[2] = __builtin_amdgcn_udot8(uiA, uiB, uiC);
+}
OpenPOWER on IntegriCloud