xref: /llvm-project/llvm/test/CodeGen/NVPTX/cp-async-bulk-tensor-g2s.ll (revision b279f6b098d3849f7f1c1f539b108307d5f8ae2d)

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_90 -mattr=+ptx80| FileCheck --check-prefixes=CHECK-PTX64 %s
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_90 -mattr=+ptx80 --nvptx-short-ptr| FileCheck --check-prefixes=CHECK-PTX-SHARED32 %s
; RUN: %if ptxas-12.3 %{ llc < %s -mtriple=nvptx64 -mcpu=sm_90 -mattr=+ptx80| %ptxas-verify -arch=sm_90 %}
; RUN: %if ptxas-12.3 %{ llc < %s -mtriple=nvptx64 -mcpu=sm_90 -mattr=+ptx80 --nvptx-short-ptr| %ptxas-verify -arch=sm_90 %}

target triple = "nvptx64-nvidia-cuda"

declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.1d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tm, i32 %d0, i16 %mc, i64 %ch, i1 %f1, i1 %f2);
declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tm, i32 %d0, i32 %d1, i16 %mc, i64 %ch, i1 %f1, i1 %f2);
declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i16 %mc, i64 %ch, i1 %f1, i1 %f2);
declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %mc, i64 %ch, i1 %f1, i1 %f2);
declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %mc, i64 %ch, i1 %f1, i1 %f2);

declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i16 %im2col0, i16 %mc, i64 %ch, i1 %f1, i1 %f2);
declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %im2col0, i16 %im2col1, i16 %mc, i64 %ch, i1 %f1, i1 %f2);
declare void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %im2col0, i16 %im2col1, i16 %im2col2, i16 %mc, i64 %ch, i1 %f1, i1 %f2);

; CHECK-LABEL: cp_async_bulk_tensor_g2s_tile_1d
define void @cp_async_bulk_tensor_g2s_tile_1d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i16 %mc, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_g2s_tile_1d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX64-NEXT:    .reg .b32 %r<2>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<5>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_1d_param_0];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_1d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_g2s_tile_1d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_1d_param_3];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%rd1], [%rd3, {%r1}], [%rd2];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd4, [cp_async_bulk_tensor_g2s_tile_1d_param_5];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%rd1], [%rd3, {%r1}], [%rd2], %rd4;
; CHECK-PTX64-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_1d_param_4];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%rd1], [%rd3, {%r1}], [%rd2], %rs1;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%rd1], [%rd3, {%r1}], [%rd2], %rs1, %rd4;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_g2s_tile_1d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<4>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_1d_param_0];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_tile_1d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_1d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_tile_1d_param_3];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%r1], [%rd1, {%r3}], [%r2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_1d_param_5];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%r1], [%rd1, {%r3}], [%r2], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_1d_param_4];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%r1], [%rd1, {%r3}], [%r2], %rs1;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.1d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%r1], [%rd1, {%r3}], [%r2], %rs1, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.1d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i16 undef, i64 undef, i1 0, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.1d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i16 undef, i64 %ch, i1 0, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.1d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i16 %mc, i64 undef, i1 1, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.1d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i16 %mc, i64 %ch, i1 1, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_g2s_tile_2d
define void @cp_async_bulk_tensor_g2s_tile_2d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i16 %mc, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_g2s_tile_2d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX64-NEXT:    .reg .b32 %r<3>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<5>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_2d_param_0];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_2d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_g2s_tile_2d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_2d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_tile_2d_param_4];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%rd1], [%rd3, {%r1, %r2}], [%rd2];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd4, [cp_async_bulk_tensor_g2s_tile_2d_param_6];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%rd1], [%rd3, {%r1, %r2}], [%rd2], %rd4;
; CHECK-PTX64-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_2d_param_5];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%rd1], [%rd3, {%r1, %r2}], [%rd2], %rs1;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%rd1], [%rd3, {%r1, %r2}], [%rd2], %rs1, %rd4;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_g2s_tile_2d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<5>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_2d_param_0];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_tile_2d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_2d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_tile_2d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_tile_2d_param_4];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%r1], [%rd1, {%r3, %r4}], [%r2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_2d_param_6];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%r1], [%rd1, {%r3, %r4}], [%r2], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_2d_param_5];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%r1], [%rd1, {%r3, %r4}], [%r2], %rs1;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.2d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%r1], [%rd1, {%r3, %r4}], [%r2], %rs1, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i16 undef, i64 undef, i1 0, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i16 undef, i64 %ch, i1 0, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i16 %mc, i64 undef, i1 1, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.2d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i16 %mc, i64 %ch, i1 1, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_g2s_tile_3d
define void @cp_async_bulk_tensor_g2s_tile_3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 %mc, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_g2s_tile_3d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX64-NEXT:    .reg .b32 %r<4>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<5>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_3d_param_0];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_3d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_g2s_tile_3d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_3d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_tile_3d_param_4];
; CHECK-PTX64-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_tile_3d_param_5];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%rd1], [%rd3, {%r1, %r2, %r3}], [%rd2];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd4, [cp_async_bulk_tensor_g2s_tile_3d_param_7];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3}], [%rd2], %rd4;
; CHECK-PTX64-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_3d_param_6];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%rd1], [%rd3, {%r1, %r2, %r3}], [%rd2], %rs1;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3}], [%rd2], %rs1, %rd4;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_g2s_tile_3d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<6>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_3d_param_0];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_tile_3d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_3d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_tile_3d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_tile_3d_param_4];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_g2s_tile_3d_param_5];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%r1], [%rd1, {%r3, %r4, %r5}], [%r2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_3d_param_7];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5}], [%r2], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_3d_param_6];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%r1], [%rd1, {%r3, %r4, %r5}], [%r2], %rs1;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5}], [%r2], %rs1, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 undef, i64 undef, i1 0, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 undef, i64 %ch, i1 0, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 %mc, i64 %ch, i1 1, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 %mc, i64 %ch, i1 1, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_g2s_tile_4d
define void @cp_async_bulk_tensor_g2s_tile_4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %mc, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_g2s_tile_4d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX64-NEXT:    .reg .b32 %r<5>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<5>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_4d_param_0];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_4d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_g2s_tile_4d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_4d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_tile_4d_param_4];
; CHECK-PTX64-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_tile_4d_param_5];
; CHECK-PTX64-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_tile_4d_param_6];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%rd1], [%rd3, {%r1, %r2, %r3, %r4}], [%rd2];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd4, [cp_async_bulk_tensor_g2s_tile_4d_param_8];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3, %r4}], [%rd2], %rd4;
; CHECK-PTX64-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_4d_param_7];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%rd1], [%rd3, {%r1, %r2, %r3, %r4}], [%rd2], %rs1;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3, %r4}], [%rd2], %rs1, %rd4;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_g2s_tile_4d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<7>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_4d_param_0];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_tile_4d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_4d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_tile_4d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_tile_4d_param_4];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_g2s_tile_4d_param_5];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r6, [cp_async_bulk_tensor_g2s_tile_4d_param_6];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%r1], [%rd1, {%r3, %r4, %r5, %r6}], [%r2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_4d_param_8];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5, %r6}], [%r2], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_4d_param_7];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%r1], [%rd1, {%r3, %r4, %r5, %r6}], [%r2], %rs1;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5, %r6}], [%r2], %rs1, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 undef, i64 undef, i1 0, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 undef, i64 %ch, i1 0, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %mc, i64 %ch, i1 1, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %mc, i64 %ch, i1 1, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_g2s_tile_5d
define void @cp_async_bulk_tensor_g2s_tile_5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %mc, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_g2s_tile_5d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX64-NEXT:    .reg .b32 %r<6>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<5>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_5d_param_0];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_5d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_g2s_tile_5d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_5d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_tile_5d_param_4];
; CHECK-PTX64-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_tile_5d_param_5];
; CHECK-PTX64-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_tile_5d_param_6];
; CHECK-PTX64-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_g2s_tile_5d_param_7];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%rd1], [%rd3, {%r1, %r2, %r3, %r4, %r5}], [%rd2];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd4, [cp_async_bulk_tensor_g2s_tile_5d_param_9];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3, %r4, %r5}], [%rd2], %rd4;
; CHECK-PTX64-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_5d_param_8];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%rd1], [%rd3, {%r1, %r2, %r3, %r4, %r5}], [%rd2], %rs1;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3, %r4, %r5}], [%rd2], %rs1, %rd4;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_g2s_tile_5d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b16 %rs<2>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<8>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_tile_5d_param_0];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_tile_5d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_tile_5d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_tile_5d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_tile_5d_param_4];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_g2s_tile_5d_param_5];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r6, [cp_async_bulk_tensor_g2s_tile_5d_param_6];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r7, [cp_async_bulk_tensor_g2s_tile_5d_param_7];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes [%r1], [%rd1, {%r3, %r4, %r5, %r6, %r7}], [%r2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_tile_5d_param_9];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5, %r6, %r7}], [%r2], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_tile_5d_param_8];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster [%r1], [%rd1, {%r3, %r4, %r5, %r6, %r7}], [%r2], %rs1;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.tile.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5, %r6, %r7}], [%r2], %rs1, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 undef, i64 undef, i1 0, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 undef, i64 %ch, i1 0, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %mc, i64 %ch, i1 1, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.tile.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %mc, i64 %ch, i1 1, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_g2s_im2col_3d
define void @cp_async_bulk_tensor_g2s_im2col_3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 %im2col0, i16 %mc, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_g2s_im2col_3d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b16 %rs<3>;
; CHECK-PTX64-NEXT:    .reg .b32 %r<4>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<5>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_im2col_3d_param_0];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_im2col_3d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_g2s_im2col_3d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_im2col_3d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_im2col_3d_param_4];
; CHECK-PTX64-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_im2col_3d_param_5];
; CHECK-PTX64-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_im2col_3d_param_6];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes [%rd1], [%rd3, {%r1, %r2, %r3}], [%rd2], {%rs1};
; CHECK-PTX64-NEXT:    ld.param.u64 %rd4, [cp_async_bulk_tensor_g2s_im2col_3d_param_8];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3}], [%rd2], {%rs1}, %rd4;
; CHECK-PTX64-NEXT:    ld.param.u16 %rs2, [cp_async_bulk_tensor_g2s_im2col_3d_param_7];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster [%rd1], [%rd3, {%r1, %r2, %r3}], [%rd2], {%rs1}, %rs2;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3}], [%rd2], {%rs1}, %rs2, %rd4;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_g2s_im2col_3d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b16 %rs<3>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<6>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_im2col_3d_param_0];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_im2col_3d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_im2col_3d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_im2col_3d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_im2col_3d_param_4];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_g2s_im2col_3d_param_5];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_im2col_3d_param_6];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes [%r1], [%rd1, {%r3, %r4, %r5}], [%r2], {%rs1};
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_im2col_3d_param_8];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5}], [%r2], {%rs1}, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs2, [cp_async_bulk_tensor_g2s_im2col_3d_param_7];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster [%r1], [%rd1, {%r3, %r4, %r5}], [%r2], {%rs1}, %rs2;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5}], [%r2], {%rs1}, %rs2, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 %im2col0, i16 undef, i64 undef, i1 0, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 %im2col0, i16 undef, i64 %ch, i1 0, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 %im2col0, i16 %mc, i64 %ch, i1 1, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.3d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i16 %im2col0, i16 %mc, i64 %ch, i1 1, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_g2s_im2col_4d
define void @cp_async_bulk_tensor_g2s_im2col_4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %im2col0, i16 %im2col1, i16 %mc, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_g2s_im2col_4d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b16 %rs<4>;
; CHECK-PTX64-NEXT:    .reg .b32 %r<5>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<5>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_im2col_4d_param_0];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_im2col_4d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_g2s_im2col_4d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_im2col_4d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_im2col_4d_param_4];
; CHECK-PTX64-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_im2col_4d_param_5];
; CHECK-PTX64-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_im2col_4d_param_6];
; CHECK-PTX64-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_im2col_4d_param_7];
; CHECK-PTX64-NEXT:    ld.param.u16 %rs2, [cp_async_bulk_tensor_g2s_im2col_4d_param_8];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes [%rd1], [%rd3, {%r1, %r2, %r3, %r4}], [%rd2], {%rs1, %rs2};
; CHECK-PTX64-NEXT:    ld.param.u64 %rd4, [cp_async_bulk_tensor_g2s_im2col_4d_param_10];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3, %r4}], [%rd2], {%rs1, %rs2}, %rd4;
; CHECK-PTX64-NEXT:    ld.param.u16 %rs3, [cp_async_bulk_tensor_g2s_im2col_4d_param_9];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster [%rd1], [%rd3, {%r1, %r2, %r3, %r4}], [%rd2], {%rs1, %rs2}, %rs3;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3, %r4}], [%rd2], {%rs1, %rs2}, %rs3, %rd4;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_g2s_im2col_4d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b16 %rs<4>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<7>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_im2col_4d_param_0];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_im2col_4d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_im2col_4d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_im2col_4d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_im2col_4d_param_4];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_g2s_im2col_4d_param_5];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r6, [cp_async_bulk_tensor_g2s_im2col_4d_param_6];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_im2col_4d_param_7];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs2, [cp_async_bulk_tensor_g2s_im2col_4d_param_8];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes [%r1], [%rd1, {%r3, %r4, %r5, %r6}], [%r2], {%rs1, %rs2};
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_im2col_4d_param_10];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5, %r6}], [%r2], {%rs1, %rs2}, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs3, [cp_async_bulk_tensor_g2s_im2col_4d_param_9];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster [%r1], [%rd1, {%r3, %r4, %r5, %r6}], [%r2], {%rs1, %rs2}, %rs3;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5, %r6}], [%r2], {%rs1, %rs2}, %rs3, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %im2col0, i16 %im2col1, i16 undef, i64 undef, i1 0, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %im2col0, i16 %im2col1, i16 undef, i64 %ch, i1 0, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %im2col0, i16 %im2col1, i16 %mc, i64 %ch, i1 1, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.4d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i16 %im2col0, i16 %im2col1, i16 %mc, i64 %ch, i1 1, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_g2s_im2col_5d
define void @cp_async_bulk_tensor_g2s_im2col_5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %im2col0, i16 %im2col1, i16 %im2col2, i16 %mc, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_g2s_im2col_5d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b16 %rs<5>;
; CHECK-PTX64-NEXT:    .reg .b32 %r<6>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<5>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_im2col_5d_param_0];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_im2col_5d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_g2s_im2col_5d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_im2col_5d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_im2col_5d_param_4];
; CHECK-PTX64-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_im2col_5d_param_5];
; CHECK-PTX64-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_im2col_5d_param_6];
; CHECK-PTX64-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_g2s_im2col_5d_param_7];
; CHECK-PTX64-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_im2col_5d_param_8];
; CHECK-PTX64-NEXT:    ld.param.u16 %rs2, [cp_async_bulk_tensor_g2s_im2col_5d_param_9];
; CHECK-PTX64-NEXT:    ld.param.u16 %rs3, [cp_async_bulk_tensor_g2s_im2col_5d_param_10];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes [%rd1], [%rd3, {%r1, %r2, %r3, %r4, %r5}], [%rd2], {%rs1, %rs2, %rs3};
; CHECK-PTX64-NEXT:    ld.param.u64 %rd4, [cp_async_bulk_tensor_g2s_im2col_5d_param_12];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3, %r4, %r5}], [%rd2], {%rs1, %rs2, %rs3}, %rd4;
; CHECK-PTX64-NEXT:    ld.param.u16 %rs4, [cp_async_bulk_tensor_g2s_im2col_5d_param_11];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster [%rd1], [%rd3, {%r1, %r2, %r3, %r4, %r5}], [%rd2], {%rs1, %rs2, %rs3}, %rs4;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%rd1], [%rd3, {%r1, %r2, %r3, %r4, %r5}], [%rd2], {%rs1, %rs2, %rs3}, %rs4, %rd4;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_g2s_im2col_5d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b16 %rs<5>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<8>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_g2s_im2col_5d_param_0];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_g2s_im2col_5d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_g2s_im2col_5d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_g2s_im2col_5d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_g2s_im2col_5d_param_4];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_g2s_im2col_5d_param_5];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r6, [cp_async_bulk_tensor_g2s_im2col_5d_param_6];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r7, [cp_async_bulk_tensor_g2s_im2col_5d_param_7];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs1, [cp_async_bulk_tensor_g2s_im2col_5d_param_8];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs2, [cp_async_bulk_tensor_g2s_im2col_5d_param_9];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs3, [cp_async_bulk_tensor_g2s_im2col_5d_param_10];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes [%r1], [%rd1, {%r3, %r4, %r5, %r6, %r7}], [%r2], {%rs1, %rs2, %rs3};
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_g2s_im2col_5d_param_12];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5, %r6, %r7}], [%r2], {%rs1, %rs2, %rs3}, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ld.param.u16 %rs4, [cp_async_bulk_tensor_g2s_im2col_5d_param_11];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster [%r1], [%rd1, {%r3, %r4, %r5, %r6, %r7}], [%r2], {%rs1, %rs2, %rs3}, %rs4;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.shared::cluster.global.im2col.mbarrier::complete_tx::bytes.multicast::cluster.L2::cache_hint [%r1], [%rd1, {%r3, %r4, %r5, %r6, %r7}], [%r2], {%rs1, %rs2, %rs3}, %rs4, %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %im2col0, i16 %im2col1, i16 %im2col2, i16 undef, i64 undef, i1 0, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %im2col0, i16 %im2col1, i16 %im2col2, i16 undef, i64 %ch, i1 0, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %im2col0, i16 %im2col1, i16 %im2col2, i16 %mc, i64 %ch, i1 1, i1 0)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.g2s.im2col.5d(ptr addrspace(3) %d, ptr addrspace(3) %bar, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i16 %im2col0, i16 %im2col1, i16 %im2col2, i16 %mc, i64 %ch, i1 1, i1 1)
  ret void
}