[AMDGPU] Regenerate i64 shift tests.

To show codegen diff due to a future SimplifyDemandedBits patch.

llvm-svn: 350065

llvm/test/CodeGen/AMDGPU/lshl64-to-32.ll

@@ -1,9 +1,19 @@
-; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=amdgcn-- -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
 
-; GCN-LABEL: {{^}}zext_shl64_to_32:
-; GCN: s_lshl_b32
-; GCN-NOT: s_lshl_b64
 define amdgpu_kernel void @zext_shl64_to_32(i64 addrspace(1)* nocapture %out, i32 %x) {
+; GCN-LABEL: zext_shl64_to_32:
+; GCN:       ; %bb.0:
+; GCN-NEXT:    s_load_dword s4, s[0:1], 0xb
+; GCN-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x9
+; GCN-NEXT:    s_mov_b32 s3, 0xf000
+; GCN-NEXT:    s_mov_b32 s2, -1
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_lshl_b32 s4, s4, 2
+; GCN-NEXT:    v_mov_b32_e32 v1, 0
+; GCN-NEXT:    v_mov_b32_e32 v0, s4
+; GCN-NEXT:    buffer_store_dwordx2 v[0:1], off, s[0:3], 0
+; GCN-NEXT:    s_endpgm
   %and = and i32 %x, 1073741823
   %ext = zext i32 %and to i64
   %shl = shl i64 %ext, 2
@@ -11,10 +21,20 @@ define amdgpu_kernel void @zext_shl64_to_32(i64 addrspace(1)* nocapture %out, i3
   ret void
 }
 
-; GCN-LABEL: {{^}}sext_shl64_to_32:
-; GCN: s_lshl_b32
-; GCN-NOT: s_lshl_b64
 define amdgpu_kernel void @sext_shl64_to_32(i64 addrspace(1)* nocapture %out, i32 %x) {
+; GCN-LABEL: sext_shl64_to_32:
+; GCN:       ; %bb.0:
+; GCN-NEXT:    s_load_dword s4, s[0:1], 0xb
+; GCN-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x9
+; GCN-NEXT:    s_mov_b32 s3, 0xf000
+; GCN-NEXT:    s_mov_b32 s2, -1
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_and_b32 s4, s4, 0x1fffffff
+; GCN-NEXT:    s_lshl_b32 s4, s4, 2
+; GCN-NEXT:    v_mov_b32_e32 v1, 0
+; GCN-NEXT:    v_mov_b32_e32 v0, s4
+; GCN-NEXT:    buffer_store_dwordx2 v[0:1], off, s[0:3], 0
+; GCN-NEXT:    s_endpgm
   %and = and i32 %x, 536870911
   %ext = sext i32 %and to i64
   %shl = shl i64 %ext, 2
@@ -22,10 +42,21 @@ define amdgpu_kernel void @sext_shl64_to_32(i64 addrspace(1)* nocapture %out, i3
   ret void
 }
 
-; GCN-LABEL: {{^}}zext_shl64_overflow:
-; GCN: s_lshl_b64
-; GCN-NOT: s_lshl_b32
 define amdgpu_kernel void @zext_shl64_overflow(i64 addrspace(1)* nocapture %out, i32 %x) {
+; GCN-LABEL: zext_shl64_overflow:
+; GCN:       ; %bb.0:
+; GCN-NEXT:    s_load_dword s2, s[0:1], 0xb
+; GCN-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x9
+; GCN-NEXT:    s_mov_b32 s3, 0xf000
+; GCN-NEXT:    s_mov_b32 s5, 0
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_and_b32 s4, s2, 0x7fffffff
+; GCN-NEXT:    s_lshl_b64 s[4:5], s[4:5], 2
+; GCN-NEXT:    s_mov_b32 s2, -1
+; GCN-NEXT:    v_mov_b32_e32 v0, s4
+; GCN-NEXT:    v_mov_b32_e32 v1, s5
+; GCN-NEXT:    buffer_store_dwordx2 v[0:1], off, s[0:3], 0
+; GCN-NEXT:    s_endpgm
   %and = and i32 %x, 2147483647
   %ext = zext i32 %and to i64
   %shl = shl i64 %ext, 2
@@ -33,10 +64,21 @@ define amdgpu_kernel void @zext_shl64_overflow(i64 addrspace(1)* nocapture %out,
   ret void
 }
 
-; GCN-LABEL: {{^}}sext_shl64_overflow:
-; GCN: s_lshl_b64
-; GCN-NOT: s_lshl_b32
 define amdgpu_kernel void @sext_shl64_overflow(i64 addrspace(1)* nocapture %out, i32 %x) {
+; GCN-LABEL: sext_shl64_overflow:
+; GCN:       ; %bb.0:
+; GCN-NEXT:    s_load_dword s2, s[0:1], 0xb
+; GCN-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x9
+; GCN-NEXT:    s_mov_b32 s3, 0xf000
+; GCN-NEXT:    s_mov_b32 s5, 0
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_and_b32 s4, s2, 0x7fffffff
+; GCN-NEXT:    s_lshl_b64 s[4:5], s[4:5], 2
+; GCN-NEXT:    s_mov_b32 s2, -1
+; GCN-NEXT:    v_mov_b32_e32 v0, s4
+; GCN-NEXT:    v_mov_b32_e32 v1, s5
+; GCN-NEXT:    buffer_store_dwordx2 v[0:1], off, s[0:3], 0
+; GCN-NEXT:    s_endpgm
   %and = and i32 %x, 2147483647
   %ext = sext i32 %and to i64
   %shl = shl i64 %ext, 2
@@ -44,10 +86,19 @@ define amdgpu_kernel void @sext_shl64_overflow(i64 addrspace(1)* nocapture %out,
   ret void
 }
 
-; GCN-LABEL: {{^}}mulu24_shl64:
-; GCN: v_mul_u32_u24_e32 [[M:v[0-9]+]], 7, v{{[0-9]+}}
-; GCN: v_lshlrev_b32_e32 v{{[0-9]+}}, 2, [[M]]
 define amdgpu_kernel void @mulu24_shl64(i32 addrspace(1)* nocapture %arg) {
+; GCN-LABEL: mulu24_shl64:
+; GCN:       ; %bb.0: ; %bb
+; GCN-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0x9
+; GCN-NEXT:    v_and_b32_e32 v0, 6, v0
+; GCN-NEXT:    s_mov_b32 s3, 0xf000
+; GCN-NEXT:    v_mul_u32_u24_e32 v0, 7, v0
+; GCN-NEXT:    s_mov_b32 s2, 0
+; GCN-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
+; GCN-NEXT:    v_mov_b32_e32 v1, 0
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    buffer_store_dword v1, v[0:1], s[0:3], 0 addr64
+; GCN-NEXT:    s_endpgm
 bb:
   %tmp = tail call i32 @llvm.amdgcn.workitem.id.x()
   %tmp1 = and i32 %tmp, 6
@@ -58,10 +109,26 @@ bb:
   ret void
 }
 
-; GCN-LABEL: {{^}}muli24_shl64:
-; GCN: v_mul_i32_i24_e32 [[M:v[0-9]+]], -7, v{{[0-9]+}}
-; GCN: v_lshlrev_b32_e32 v{{[0-9]+}}, 3, [[M]]
 define amdgpu_kernel void @muli24_shl64(i64 addrspace(1)* nocapture %arg, i32 addrspace(1)* nocapture readonly %arg1) {
+; GCN-LABEL: muli24_shl64:
+; GCN:       ; %bb.0: ; %bb
+; GCN-NEXT:    s_load_dwordx4 s[0:3], s[0:1], 0x9
+; GCN-NEXT:    s_mov_b32 s7, 0xf000
+; GCN-NEXT:    s_mov_b32 s6, 0
+; GCN-NEXT:    v_lshlrev_b32_e32 v1, 2, v0
+; GCN-NEXT:    v_mov_b32_e32 v2, 0
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_mov_b64 s[4:5], s[2:3]
+; GCN-NEXT:    buffer_load_dword v1, v[1:2], s[4:7], 0 addr64
+; GCN-NEXT:    s_mov_b64 s[2:3], s[6:7]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_or_b32_e32 v1, 0x800000, v1
+; GCN-NEXT:    v_mul_i32_i24_e32 v1, -7, v1
+; GCN-NEXT:    v_lshlrev_b32_e32 v1, 3, v1
+; GCN-NEXT:    v_lshlrev_b32_e32 v3, 3, v0
+; GCN-NEXT:    v_mov_b32_e32 v4, v2
+; GCN-NEXT:    buffer_store_dwordx2 v[1:2], v[3:4], s[0:3], 0 addr64
+; GCN-NEXT:    s_endpgm
 bb:
   %tmp = tail call i32 @llvm.amdgcn.workitem.id.x()
   %tmp2 = sext i32 %tmp to i64