[Mlir-commits] [mlir] [mlir][nvgpu] NVGPU Tutorials (PR #87065)

[Jacques Pienaar]

================
@@ -0,0 +1,68 @@
+# RUN: env SUPPORT_LIB=%mlir_cuda_runtime \
+# RUN:   %PYTHON %s | FileCheck %s
+
+# ===----------------------------------------------------------------------===//
+#  Chapter 1 : 2D Saxpy
+# ===----------------------------------------------------------------------===//
+#
+# This program demonstrates 2D Saxpy
+#
+# This chapter introduces demonstrates:
+#  1. Use MLIR GPU dialect to allocate and copy memory
+#  2. Compute 2D SAXPY kernel
+#  3. Pass numpy arrays to MLIR
+#  4. Verify MLIR with reference computation
+#
+# ===----------------------------------------------------------------------===//
+
+
+from mlir import ir
+from mlir.dialects import gpu, memref
+from tools.nvdsl import *
+import numpy as np
+
+
+ at NVDSL.mlir_func
+def saxpy(x, y, alpha):
+    # 1. Use MLIR GPU dialect to allocate and copy memory
+    token_ty = ir.Type.parse("!gpu.async.token")
----------------
jpienaar wrote:

We really should be hiding this in the dialect at least (e.g., parse is backdoor for where more native support hasn't been exposed and plumbed through, lets at least contain that in the dialect)

https://github.com/llvm/llvm-project/pull/87065