[llvm] 01c53d7 - [AMDGPU] Add an operand folding test case from D114232

[Jay Foad via llvm-commits]

Author: Jay Foad
Date: 2022-09-07T11:16:40+01:00
New Revision: 01c53d7d80eada4db2c63df28bfbeeb0e2f8cad7

URL: https://github.com/llvm/llvm-project/commit/01c53d7d80eada4db2c63df28bfbeeb0e2f8cad7
DIFF: https://github.com/llvm/llvm-project/commit/01c53d7d80eada4db2c63df28bfbeeb0e2f8cad7.diff

LOG: [AMDGPU] Add an operand folding test case from D114232

Added: 
    llvm/test/CodeGen/AMDGPU/fold-multiple-commute.mir

Modified: 
    

Removed: 
    


################################################################################
diff  --git a/llvm/test/CodeGen/AMDGPU/fold-multiple-commute.mir b/llvm/test/CodeGen/AMDGPU/fold-multiple-commute.mir
new file mode 100644
index 000000000000..2f7bd1c63ad0
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/fold-multiple-commute.mir
@@ -0,0 +1,26 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+# RUN: llc -march=amdgcn -mcpu=gfx900 -verify-machineinstrs -run-pass si-fold-operands %s -o - | FileCheck %s
+
+# Check that the addc instructions are commuted to allow folding of the constant
+# 0 into multiple uses as an inline operand.
+---
+name: test_commute
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: $vgpr0, $vgpr1, $vcc
+    ; CHECK-LABEL: name: test_commute
+    ; CHECK: liveins: $vgpr0, $vgpr1, $vcc
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; CHECK-NEXT: [[V_ADDC_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADDC_U32_e32 0, [[COPY]], implicit-def $vcc, implicit $vcc, implicit $exec
+    ; CHECK-NEXT: [[V_ADDC_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADDC_U32_e32 0, [[COPY1]], implicit-def $vcc, implicit $vcc, implicit $exec
+    ; CHECK-NEXT: S_NOP 0, implicit [[V_ADDC_U32_e32_]], implicit [[V_ADDC_U32_e32_1]]
+    %0:vgpr_32 = COPY $vgpr0
+    %1:vgpr_32 = COPY $vgpr1
+    %2:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
+    %3:vgpr_32 = V_ADDC_U32_e32 %0, %2, implicit-def $vcc, implicit $vcc, implicit $exec
+    %4:vgpr_32 = V_ADDC_U32_e32 %1, %2, implicit-def $vcc, implicit $vcc, implicit $exec
+    S_NOP 0, implicit %3, implicit %4
+...