[Mlir-commits] [mlir] [mlir][gpu] Support Cluster of Thread Blocks in `gpu.launch_func` (PR #72871)

[llvmlistbot at llvm.org]

================
@@ -550,12 +601,23 @@ def GPU_LaunchFuncOp :GPU_Op<"launch_func", [
     /// The name of the kernel.
     StringAttr getKernelName();
 
+    /// Has cluster
+    bool hasClusterSize() {
+      auto totalSize = getOperands().size();
+      totalSize -= getKernelOperands().size();
+      totalSize -= getAsyncDependencies().size();
+      return totalSize > 7;
+    }
+
     /// The number of operands passed to the kernel function.
     unsigned getNumKernelOperands();
 
     /// The i-th operand passed to the kernel function.
     Value getKernelOperand(unsigned i);
 
+    /// Get the SSA values passed as operands to specify the cluster size.
----------------
qcolombet wrote:

Could you add in the comment what happens if we call this function when `hasClusterSize()` is `false`?

https://github.com/llvm/llvm-project/pull/72871