[clang] [llvm] [AMDGPU] Add amdgcn builtin for s_prefetch_inst (PR #211642)
Robert Imschweiler via cfe-commits
cfe-commits at lists.llvm.org
Fri Jul 24 11:48:29 PDT 2026
================
@@ -793,3 +800,16 @@ using the scale factor ``scale``, then converts the values to a packed
integers.
}];
}
+
+//===----------------------------------------------------------------------===//
+// Prefetch Builtins
+//===----------------------------------------------------------------------===//
+
+def DocPrefetchInst_GFX12 : Documentation {
+ let Category = DocCatAMDGPUPrefetch;
+ let Content = [{
+Prefetch instructions into I-Cache from the address given by ``ptr`` pointing
+to instruction memory, with the length specified by ``len``, which should be
+0-31 (1-32 chunks, units of 128 bytes).
+}];
+}
----------------
ro-i wrote:
eof
https://github.com/llvm/llvm-project/pull/211642
More information about the cfe-commits
mailing list