1 files changed, 39 insertions, 0 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/simplify-demanded-bits-build-pair.ll b/llvm/test/CodeGen/AMDGPU/simplify-demanded-bits-build-pair.ll
new file mode 100644
index 00000000000..dffee70b6b0
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/simplify-demanded-bits-build-pair.ll
@@ -0,0 +1,39 @@
+; XFAIL: *
+; RUN: llc -verify-machineinstrs -march=amdgcn -mcpu=SI -mattr=-promote-alloca < %s | FileCheck -check-prefix=SI %s
+; RUN: llc -verify-machineinstrs -march=amdgcn -mcpu=tonga -mattr=-promote-alloca < %s | FileCheck -check-prefix=SI %s
+
+; 64-bit select was originally lowered with a build_pair, and this
+; could be simplified to 1 cndmask instead of 2, but that broken when
+; it started being implemented with a v2i32 build_vector and
+; bitcasting.
+define void @trunc_select_i64(i32 addrspace(1)* %out, i64 %a, i64 %b, i32 %c) {
+  %cmp = icmp eq i32 %c, 0
+  %select = select i1 %cmp, i64 %a, i64 %b
+  %trunc = trunc i64 %select to i32
+  store i32 %trunc, i32 addrspace(1)* %out, align 4
+  ret void
+}
+
+; FIXME: Fix truncating store for local memory
+; SI-LABEL: {{^}}trunc_load_alloca_i64:
+; SI: v_movrels_b32
+; SI-NOT: v_movrels_b32
+; SI: s_endpgm
+define void @trunc_load_alloca_i64(i64 addrspace(1)* %out, i32 %a, i32 %b) {
+  %idx = add i32 %a, %b
+  %alloca = alloca i64, i32 4
+  %gep0 = getelementptr i64, i64* %alloca, i64 0
+  %gep1 = getelementptr i64, i64* %alloca, i64 1
+  %gep2 = getelementptr i64, i64* %alloca, i64 2
+  %gep3 = getelementptr i64, i64* %alloca, i64 3
+  store i64 24, i64* %gep0, align 8
+  store i64 9334, i64* %gep1, align 8
+  store i64 3935, i64* %gep2, align 8
+  store i64 9342, i64* %gep3, align 8
+  %gep = getelementptr i64, i64* %alloca, i32 %idx
+  %load = load i64, i64* %gep, align 8
+  %mask = and i64 %load, 4294967296
+  %add = add i64 %mask, -1
+  store i64 %add, i64 addrspace(1)* %out, align 4
+  ret void
+}