[SYCL][HIP] Support amd-gpu-gfx1034 as an acceptable value for -fsycl…

…-targets (intel#8106) * AMD GFX1034 is ROCm supported architecture and as such amd-gpu-gfx1034 is added as an acceptable value for -fsycl-targets. * Remove `Added in version` for supported architectures.
YuriPlyakhin · Feb 3, 2023 · 5e86a41 · 5e86a41
1 parent 3df87e2
commit 5e86a41
Show file tree

Hide file tree

Showing 7 changed files with 96 additions and 80 deletions.
diff --git a/clang/lib/Driver/ToolChains/SYCL.cpp b/clang/lib/Driver/ToolChains/SYCL.cpp
@@ -693,6 +693,7 @@ StringRef SYCL::gen::resolveGenDevice(StringRef DeviceName) {
                .Case("amd_gpu_gfx1030", "gfx1030")
                .Case("amd_gpu_gfx1031", "gfx1031")
                .Case("amd_gpu_gfx1032", "gfx1032")
+               .Case("amd_gpu_gfx1034", "gfx1034")
                .Default("");
   return Device;
 }
@@ -757,6 +758,7 @@ SmallString<64> SYCL::gen::getGenDeviceMacro(StringRef DeviceName) {
                       .Case("gfx1030", "AMD_GPU_GFX1030")
                       .Case("gfx1031", "AMD_GPU_GFX1031")
                       .Case("gfx1032", "AMD_GPU_GFX1032")
+                      .Case("gfx1034", "AMD_GPU_GFX1034")
                       .Default("");
   if (!Ext.empty()) {
     Macro = "__SYCL_TARGET_";

diff --git a/clang/test/Driver/sycl-oneapi-gpu.cpp b/clang/test/Driver/sycl-oneapi-gpu.cpp
@@ -166,12 +166,13 @@
 // RUN:   FileCheck %s --check-prefixes=DEVICE_AMD,MACRO_AMD -DDEV_STR=gfx1031 -DMAC_STR=GFX1031
 // RUN: %clangxx -fsycl -fsycl-targets=amd_gpu_gfx1032 -### %s 2>&1 | \
 // RUN:   FileCheck %s --check-prefixes=DEVICE_AMD,MACRO_AMD -DDEV_STR=gfx1032 -DMAC_STR=GFX1032
+// RUN: %clangxx -fsycl -fsycl-targets=amd_gpu_gfx1034 -### %s 2>&1 | \
+// RUN:   FileCheck %s --check-prefixes=DEVICE_AMD,MACRO_AMD -DDEV_STR=gfx1034 -DMAC_STR=GFX1034
 // MACRO_AMD: clang{{.*}} "-triple" "amdgcn-amd-amdhsa"
 // MACRO_AMD: "-D__SYCL_TARGET_AMD_GPU_[[MAC_STR]]__"
 // DEVICE_AMD: clang-offload-wrapper{{.*}} "-compile-opts=--offload-arch=[[DEV_STR]]{{.*}}"
 // MACRO_AMD: clang{{.*}} "-fsycl-is-host"
 // MACRO_AMD: "-D__SYCL_TARGET_AMD_GPU_[[MAC_STR]]__"
-
 /// -fsycl-targets=spir64_x86_64 should set a specific macro
 // RUN: %clangxx -c -fsycl -fsycl-targets=spir64_x86_64 -### %s 2>&1 | \
 // RUN:   FileCheck %s --check-prefix=MACRO_X86_64

diff --git a/sycl/doc/UsersManual.md b/sycl/doc/UsersManual.md
@@ -92,6 +92,7 @@ and not recommended to use in production environment.
     * amd_gpu_gfx1030 - AMD GCN GFX10.3 (RDNA 2) architecture
     * amd_gpu_gfx1031 - GCN GFX10.3 (RDNA 2) architecture
     * amd_gpu_gfx1032 - GCN GFX10.3 (RDNA 2) architecture
+    * amd_gpu_gfx1034 - GCN GFX10.3 (RDNA 2) architecture
 
 ## Language options
 

diff --git a/sycl/doc/design/DeviceIf.md b/sycl/doc/design/DeviceIf.md
@@ -108,6 +108,7 @@ recognizes:
 * `amd_gpu_gfx1030`
 * `amd_gpu_gfx1031`
 * `amd_gpu_gfx1032`
+* `amd_gpu_gfx1034`
 
 The above listed device names may not be mixed with the existing target name
 `spir64_gen` on the same command line.  In addition, the user must not pass the
@@ -190,6 +191,7 @@ one of the following corresponding C++ macro names:
 * `__SYCL_TARGET_AMD_GPU_GFX1030__`
 * `__SYCL_TARGET_AMD_GPU_GFX1031__`
 * `__SYCL_TARGET_AMD_GPU_GFX1032__`
+* `__SYCL_TARGET_AMD_GPU_GFX1034__`
 
 If the user invokes the compiler driver with `-fsycl-targets=spir64_x86_64`,
 the compiler driver must predefine the following C++ macro name: