[clang] [CUDA] Add support for sm101 target architecture (Tegra Blackwell) (PR #127187)

[Sebastian Jodłowski via cfe-commits]

https://github.com/jodelek created https://github.com/llvm/llvm-project/pull/127187

Add support for sm101 target architecture (Tegra Blackwell). It requires CUDA 12.8.



>From a55c76bac1bd70878c777b5930553fac114d2fd5 Mon Sep 17 00:00:00 2001
From: Sebastian Jodlowski <sjodlowski at nuro.ai>
Date: Fri, 14 Feb 2025 01:34:16 -0800
Subject: [PATCH] Add support for sm101 target architecture (Tegra Blackwell)

---
 clang/include/clang/Basic/BuiltinsNVPTX.td      | 4 ++++
 clang/include/clang/Basic/Cuda.h                | 2 ++
 clang/lib/Basic/Cuda.cpp                        | 4 ++++
 clang/lib/Basic/Targets/NVPTX.cpp               | 5 +++++
 clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp        | 2 ++
 clang/test/Misc/target-invalid-cpu-note/nvptx.c | 2 ++
 6 files changed, 19 insertions(+)

diff --git a/clang/include/clang/Basic/BuiltinsNVPTX.td b/clang/include/clang/Basic/BuiltinsNVPTX.td
index 327dc88cffdb4..3853e7dc8fbaf 100644
--- a/clang/include/clang/Basic/BuiltinsNVPTX.td
+++ b/clang/include/clang/Basic/BuiltinsNVPTX.td
@@ -21,6 +21,10 @@ class SM<string version, list<SMFeatures> newer_list> : SMFeatures {
                         !strconcat(f, "|", newer.Features));
 }
 
+let Features = "sm_101a" in def SM_101a : SMFeatures;
+
+def SM_101 : SM<"101", [SM_101a]>;
+
 let Features = "sm_100a" in def SM_100a : SMFeatures;
 
 def SM_100 : SM<"100", [SM_100a]>;
diff --git a/clang/include/clang/Basic/Cuda.h b/clang/include/clang/Basic/Cuda.h
index f33ba46233a7a..380f51fed22a2 100644
--- a/clang/include/clang/Basic/Cuda.h
+++ b/clang/include/clang/Basic/Cuda.h
@@ -82,6 +82,8 @@ enum class OffloadArch {
   SM_90a,
   SM_100,
   SM_100a,
+  SM_101,
+  SM_101a,
   GFX600,
   GFX601,
   GFX602,
diff --git a/clang/lib/Basic/Cuda.cpp b/clang/lib/Basic/Cuda.cpp
index 1bfec0b37c5ee..e92a12b3ce3be 100644
--- a/clang/lib/Basic/Cuda.cpp
+++ b/clang/lib/Basic/Cuda.cpp
@@ -100,6 +100,8 @@ static const OffloadArchToStringMap arch_names[] = {
     SM(90a),                         // Hopper
     SM(100),                         // Blackwell
     SM(100a),                        // Blackwell
+    SM(101),                         // Blackwell
+    SM(101a),                        // Blackwell
     GFX(600),  // gfx600
     GFX(601),  // gfx601
     GFX(602),  // gfx602
@@ -230,6 +232,8 @@ CudaVersion MinVersionForOffloadArch(OffloadArch A) {
     return CudaVersion::CUDA_120;
   case OffloadArch::SM_100:
   case OffloadArch::SM_100a:
+  case OffloadArch::SM_101:
+  case OffloadArch::SM_101a:
     return CudaVersion::CUDA_128;
   default:
     llvm_unreachable("invalid enum");
diff --git a/clang/lib/Basic/Targets/NVPTX.cpp b/clang/lib/Basic/Targets/NVPTX.cpp
index 7d13c1f145440..3f3c1bb653b04 100644
--- a/clang/lib/Basic/Targets/NVPTX.cpp
+++ b/clang/lib/Basic/Targets/NVPTX.cpp
@@ -292,6 +292,9 @@ void NVPTXTargetInfo::getTargetDefines(const LangOptions &Opts,
       case OffloadArch::SM_100:
       case OffloadArch::SM_100a:
         return "1000";
+      case OffloadArch::SM_101:
+      case OffloadArch::SM_101a:
+         return "1010";
       }
       llvm_unreachable("unhandled OffloadArch");
     }();
@@ -300,6 +303,8 @@ void NVPTXTargetInfo::getTargetDefines(const LangOptions &Opts,
       Builder.defineMacro("__CUDA_ARCH_FEAT_SM90_ALL", "1");
     if (GPU == OffloadArch::SM_100a)
       Builder.defineMacro("__CUDA_ARCH_FEAT_SM100_ALL", "1");
+    if (GPU == OffloadArch::SM_101a)
+      Builder.defineMacro("__CUDA_ARCH_FEAT_SM101_ALL", "1");
   }
 }
 
diff --git a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
index c13928f61a748..2cac1eb73b438 100644
--- a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
+++ b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
@@ -2278,6 +2278,8 @@ void CGOpenMPRuntimeGPU::processRequiresDirective(const OMPRequiresDecl *D) {
       case OffloadArch::SM_90a:
       case OffloadArch::SM_100:
       case OffloadArch::SM_100a:
+      case OffloadArch::SM_101:
+      case OffloadArch::SM_101a:
       case OffloadArch::GFX600:
       case OffloadArch::GFX601:
       case OffloadArch::GFX602:
diff --git a/clang/test/Misc/target-invalid-cpu-note/nvptx.c b/clang/test/Misc/target-invalid-cpu-note/nvptx.c
index 3afcdf8c9fe5c..ed9d4865c3ec9 100644
--- a/clang/test/Misc/target-invalid-cpu-note/nvptx.c
+++ b/clang/test/Misc/target-invalid-cpu-note/nvptx.c
@@ -28,6 +28,8 @@
 // CHECK-SAME: {{^}}, sm_90a
 // CHECK-SAME: {{^}}, sm_100
 // CHECK-SAME: {{^}}, sm_100a
+// CHECK-SAME: {{^}}, sm_101
+// CHECK-SAME: {{^}}, sm_101a
 // CHECK-SAME: {{^}}, gfx600
 // CHECK-SAME: {{^}}, gfx601
 // CHECK-SAME: {{^}}, gfx602