[llvm] [AMDGPU] handling of COPY for wide tuples when v_pk_mov_b32 is used (PR #119966)

[Pravin Jagtap via llvm-commits]

https://github.com/pravinjagtap created https://github.com/llvm/llvm-project/pull/119966

None

>From f92bbe77a30baa0e741889bef8186f626997111b Mon Sep 17 00:00:00 2001
From: Pravin Jagtap <Pravin.Jagtap at amd.com>
Date: Sat, 14 Dec 2024 18:41:43 +0530
Subject: [PATCH] [AMDGPU] handling of COPY for wide tuples when v_pk_mov_b32
 is used

---
 llvm/test/CodeGen/AMDGPU/copy_phys_vgpr1024.mir | 12 ++++++++++++
 1 file changed, 12 insertions(+)
 create mode 100644 llvm/test/CodeGen/AMDGPU/copy_phys_vgpr1024.mir

diff --git a/llvm/test/CodeGen/AMDGPU/copy_phys_vgpr1024.mir b/llvm/test/CodeGen/AMDGPU/copy_phys_vgpr1024.mir
new file mode 100644
index 00000000000000..035116ad4155a5
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/copy_phys_vgpr1024.mir
@@ -0,0 +1,12 @@
+
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+# RUN: llc -mtriple=amdgcn -mcpu=gfx90a -run-pass postrapseudos -verify-machineinstrs %s -o - | FileCheck -check-prefix=GFX90A %s
+
+---
+name: copy_v1024_to_v1024_fwd
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: $vgpr2_vgpr3_vgpr4_vgpr5, $vgpr32
+    $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31 = COPY killed $vgpr32_vgpr33_vgpr34_vgpr35_vgpr36_vgpr37_vgpr38_vgpr39_vgpr40_vgpr41_vgpr42_vgpr43_vgpr44_vgpr45_vgpr46_vgpr47_vgpr48_vgpr49_vgpr50_vgpr51_vgpr52_vgpr53_vgpr54_vgpr55_vgpr56_vgpr57_vgpr58_vgpr59_vgpr60_vgpr61_vgpr62_vgpr63, implicit $exec
+...