-
Notifications
You must be signed in to change notification settings - Fork 15.3k
[NFC][AMDGPU] Auto generate check lines for schedule-regpressure-ilp-metric-spills.mir
#139184
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Closed
shiltian
wants to merge
1
commit into
main
from
users/shiltian/auto-generate-check-lines-for-schedule-regpressure-ilp-metric-spills
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…-metric-spills.mir`
Contributor
Author
This stack of pull requests is managed by Graphite. Learn more about stacking. |
Member
|
@llvm/pr-subscribers-backend-amdgpu Author: Shilei Tian (shiltian) ChangesPatch is 217.37 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/139184.diff 1 Files Affected:
diff --git a/llvm/test/CodeGen/AMDGPU/schedule-regpressure-ilp-metric-spills.mir b/llvm/test/CodeGen/AMDGPU/schedule-regpressure-ilp-metric-spills.mir
index aa0d1fe45e9a8..9435792c93f1e 100644
--- a/llvm/test/CodeGen/AMDGPU/schedule-regpressure-ilp-metric-spills.mir
+++ b/llvm/test/CodeGen/AMDGPU/schedule-regpressure-ilp-metric-spills.mir
@@ -1,3 +1,4 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 5
# RUN: llc -mtriple=amdgcn -mcpu=gfx1100 -start-before=machine-scheduler -stop-after=greedy,2 -verify-machineinstrs -o - %s | FileCheck -check-prefix=GCN %s
# RUN: llc -mtriple=amdgcn -mcpu=gfx1100 -start-before=machine-scheduler -stop-after=greedy,2 -amdgpu-use-amdgpu-trackers=1 -verify-machineinstrs -o - %s | FileCheck -check-prefix=GCN-GCNTRACKER %s
@@ -36,6 +37,1435 @@ body: |
bb.0:
liveins: $vgpr0, $sgpr0_sgpr1, $sgpr15
+ ; GCN-LABEL: name: no_sched_metric_due_to_spills
+ ; GCN: liveins: $sgpr15, $vgpr0, $sgpr0_sgpr1
+ ; GCN-NEXT: {{ $}}
+ ; GCN-NEXT: [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr0
+ ; GCN-NEXT: renamable $sgpr2_sgpr3 = S_LOAD_DWORDX2_IMM renamable $sgpr0_sgpr1, 216, 0 :: (dereferenceable invariant load (s64), addrspace 4)
+ ; GCN-NEXT: renamable $sgpr2 = S_ADD_I32 renamable $sgpr2, 127, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr3 = S_ADD_I32 killed renamable $sgpr3, 255, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr4 = S_ASHR_I32 renamable $sgpr2, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_ASHR_I32 renamable $sgpr3, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr4 = S_LSHR_B32 killed renamable $sgpr4, 25, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_LSHR_B32 killed renamable $sgpr5, 24, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr2 = S_ADD_I32 killed renamable $sgpr2, killed renamable $sgpr4, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr3 = S_ADD_I32 killed renamable $sgpr3, killed renamable $sgpr5, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr4 = S_ASHR_I32 renamable $sgpr2, 7, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr3 = S_ASHR_I32 killed renamable $sgpr3, 8, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = nsw S_MUL_I32 renamable $sgpr3, renamable $sgpr4
+ ; GCN-NEXT: renamable $sgpr6 = S_ASHR_I32 renamable $sgpr5, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_ADD_I32 killed renamable $sgpr5, renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_XOR_B32 killed renamable $sgpr5, killed renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: [[V_CVT_F32_U32_e64_:%[0-9]+]]:vgpr_32 = V_CVT_F32_U32_e64 $sgpr5, 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_RCP_IFLAG_F32_e64_:%[0-9]+]]:vgpr_32 = V_RCP_IFLAG_F32_e64 0, [[V_CVT_F32_U32_e64_]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_MUL_F32_e64_:%[0-9]+]]:vgpr_32 = V_MUL_F32_e64 0, 1333788670, 0, [[V_RCP_IFLAG_F32_e64_]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_CVT_U32_F32_e64_:%[0-9]+]]:vgpr_32 = V_CVT_U32_F32_e64 0, [[V_MUL_F32_e64_]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: renamable $sgpr6 = V_READFIRSTLANE_B32 [[V_CVT_U32_F32_e64_]], implicit $exec
+ ; GCN-NEXT: renamable $sgpr7 = S_SUB_I32 0, renamable $sgpr5, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_ASHR_I32 renamable $sgpr15, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_MUL_I32 killed renamable $sgpr7, renamable $sgpr6
+ ; GCN-NEXT: renamable $sgpr9 = S_ADD_I32 killed renamable $sgpr15, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_MUL_HI_U32 renamable $sgpr6, killed renamable $sgpr7
+ ; GCN-NEXT: renamable $sgpr9 = S_XOR_B32 killed renamable $sgpr9, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_ADD_I32 killed renamable $sgpr6, killed renamable $sgpr7, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_MUL_HI_U32 renamable $sgpr9, killed renamable $sgpr6
+ ; GCN-NEXT: renamable $sgpr6 = S_MUL_I32 killed renamable $sgpr6, renamable $sgpr5
+ ; GCN-NEXT: renamable $sgpr6 = S_SUB_I32 killed renamable $sgpr9, killed renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_SUB_I32 renamable $sgpr6, renamable $sgpr5, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 renamable $sgpr6, renamable $sgpr5, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_CSELECT_B32 killed renamable $sgpr7, killed renamable $sgpr6, implicit $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_SUB_I32 renamable $sgpr6, renamable $sgpr5, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 renamable $sgpr6, killed renamable $sgpr5, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_CSELECT_B32 killed renamable $sgpr7, killed renamable $sgpr6, implicit $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_ASHR_I32 renamable $sgpr3, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_ADD_I32 renamable $sgpr3, renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_XOR_B32 killed renamable $sgpr7, renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: [[V_CVT_F32_U32_e64_1:%[0-9]+]]:vgpr_32 = V_CVT_F32_U32_e64 $sgpr7, 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_RCP_IFLAG_F32_e64_1:%[0-9]+]]:vgpr_32 = V_RCP_IFLAG_F32_e64 0, [[V_CVT_F32_U32_e64_1]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_MUL_F32_e64_1:%[0-9]+]]:vgpr_32 = V_MUL_F32_e64 0, 1333788670, 0, [[V_RCP_IFLAG_F32_e64_1]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_CVT_U32_F32_e64_1:%[0-9]+]]:vgpr_32 = V_CVT_U32_F32_e64 0, [[V_MUL_F32_e64_1]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: renamable $sgpr5 = S_XOR_B32 killed renamable $sgpr5, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr9 = V_READFIRSTLANE_B32 [[V_CVT_U32_F32_e64_1]], implicit $exec
+ ; GCN-NEXT: renamable $sgpr10 = S_LOAD_DWORD_IMM renamable $sgpr0_sgpr1, 192, 0 :: (dereferenceable invariant load (s32), align 8, addrspace 4)
+ ; GCN-NEXT: renamable $sgpr5 = S_SUB_I32 killed renamable $sgpr5, killed renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_SUB_I32 0, renamable $sgpr7, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr11 = S_ASHR_I32 renamable $sgpr5, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_MUL_I32 killed renamable $sgpr8, renamable $sgpr9
+ ; GCN-NEXT: renamable $sgpr12 = S_ADD_I32 renamable $sgpr5, renamable $sgpr11, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_MUL_HI_U32 renamable $sgpr9, killed renamable $sgpr8
+ ; GCN-NEXT: renamable $sgpr12 = S_XOR_B32 killed renamable $sgpr12, renamable $sgpr11, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_ADD_I32 killed renamable $sgpr9, killed renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_MUL_HI_U32 renamable $sgpr12, killed renamable $sgpr8
+ ; GCN-NEXT: renamable $sgpr9 = S_MUL_I32 renamable $sgpr8, renamable $sgpr7
+ ; GCN-NEXT: renamable $sgpr6 = S_XOR_B32 killed renamable $sgpr11, killed renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr9 = S_SUB_I32 killed renamable $sgpr12, killed renamable $sgpr9, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr11 = S_ADD_I32 renamable $sgpr8, 1, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr12 = S_SUB_I32 renamable $sgpr9, renamable $sgpr7, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 renamable $sgpr9, renamable $sgpr7, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_CSELECT_B32 killed renamable $sgpr11, killed renamable $sgpr8, implicit $scc
+ ; GCN-NEXT: renamable $sgpr9 = S_CSELECT_B32 killed renamable $sgpr12, killed renamable $sgpr9, implicit $scc
+ ; GCN-NEXT: renamable $sgpr11 = S_ADD_I32 renamable $sgpr8, 1, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 killed renamable $sgpr9, killed renamable $sgpr7, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_CSELECT_B32 killed renamable $sgpr11, killed renamable $sgpr8, implicit $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_ASHR_I32 renamable $sgpr10, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr9 = S_ADD_I32 renamable $sgpr10, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_XOR_B32 killed renamable $sgpr9, killed renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: [[V_CVT_F32_U32_e64_2:%[0-9]+]]:vgpr_32 = V_CVT_F32_U32_e64 $sgpr8, 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_RCP_IFLAG_F32_e64_2:%[0-9]+]]:vgpr_32 = V_RCP_IFLAG_F32_e64 0, [[V_CVT_F32_U32_e64_2]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_MUL_F32_e64_2:%[0-9]+]]:vgpr_32 = V_MUL_F32_e64 0, 1333788670, 0, [[V_RCP_IFLAG_F32_e64_2]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_CVT_U32_F32_e64_2:%[0-9]+]]:vgpr_32 = V_CVT_U32_F32_e64 0, [[V_MUL_F32_e64_2]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: renamable $sgpr9 = V_READFIRSTLANE_B32 [[V_CVT_U32_F32_e64_2]], implicit $exec
+ ; GCN-NEXT: renamable $sgpr11 = S_SUB_I32 0, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr2 = S_ASHR_I32 killed renamable $sgpr2, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr11 = S_MUL_I32 killed renamable $sgpr11, renamable $sgpr9
+ ; GCN-NEXT: renamable $sgpr12 = S_ADD_I32 renamable $sgpr4, renamable $sgpr2, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr11 = S_MUL_HI_U32 renamable $sgpr9, killed renamable $sgpr11
+ ; GCN-NEXT: renamable $sgpr12 = S_XOR_B32 killed renamable $sgpr12, renamable $sgpr2, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr9 = S_ADD_I32 killed renamable $sgpr9, killed renamable $sgpr11, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr11 = S_MUL_HI_U32 renamable $sgpr12, renamable $sgpr9
+ ; GCN-NEXT: renamable $sgpr11 = S_MUL_I32 killed renamable $sgpr11, renamable $sgpr8
+ ; GCN-NEXT: renamable $sgpr7 = S_XOR_B32 killed renamable $sgpr7, renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr11 = S_SUB_I32 killed renamable $sgpr12, killed renamable $sgpr11, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr12 = S_SUB_I32 killed renamable $sgpr7, killed renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_SUB_I32 renamable $sgpr11, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 renamable $sgpr11, renamable $sgpr8, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_CSELECT_B32 killed renamable $sgpr6, killed renamable $sgpr11, implicit $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_SUB_I32 renamable $sgpr6, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 renamable $sgpr6, renamable $sgpr8, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_CSELECT_B32 killed renamable $sgpr7, killed renamable $sgpr6, implicit $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_XOR_B32 killed renamable $sgpr6, renamable $sgpr2, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr2 = S_SUB_I32 killed renamable $sgpr6, killed renamable $sgpr2, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr4 = nsw S_SUB_I32 killed renamable $sgpr4, renamable $sgpr2, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_LT_I32 renamable $sgpr12, killed renamable $sgpr4, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr2 = S_CSELECT_B32 killed renamable $sgpr10, killed renamable $sgpr2, implicit $scc
+ ; GCN-NEXT: renamable $sgpr4 = S_ASHR_I32 renamable $sgpr12, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_ADD_I32 renamable $sgpr12, renamable $sgpr4, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_XOR_B32 killed renamable $sgpr6, renamable $sgpr4, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_MUL_HI_U32 renamable $sgpr6, killed renamable $sgpr9
+ ; GCN-NEXT: renamable $sgpr7 = S_MUL_I32 killed renamable $sgpr7, renamable $sgpr8
+ ; GCN-NEXT: renamable $sgpr9 = S_MUL_I32 renamable $sgpr12, renamable $sgpr3
+ ; GCN-NEXT: renamable $sgpr6 = S_SUB_I32 killed renamable $sgpr6, killed renamable $sgpr7, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_SUB_I32 killed renamable $sgpr5, killed renamable $sgpr9, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_SUB_I32 renamable $sgpr6, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 renamable $sgpr6, renamable $sgpr8, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_CSELECT_B32 killed renamable $sgpr7, killed renamable $sgpr6, implicit $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_SUB_I32 renamable $sgpr6, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 renamable $sgpr6, killed renamable $sgpr8, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_CSELECT_B32 killed renamable $sgpr7, killed renamable $sgpr6, implicit $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_ASHR_I32 renamable $sgpr2, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_ADD_I32 renamable $sgpr2, renamable $sgpr7, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr8 = S_XOR_B32 killed renamable $sgpr8, renamable $sgpr7, implicit-def dead $scc
+ ; GCN-NEXT: [[V_CVT_F32_U32_e64_3:%[0-9]+]]:vgpr_32 = V_CVT_F32_U32_e64 $sgpr8, 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_RCP_IFLAG_F32_e64_3:%[0-9]+]]:vgpr_32 = V_RCP_IFLAG_F32_e64 0, [[V_CVT_F32_U32_e64_3]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: [[V_MUL_F32_e64_3:%[0-9]+]]:vgpr_32 = V_MUL_F32_e64 0, 1333788670, 0, [[V_RCP_IFLAG_F32_e64_3]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: renamable $sgpr6 = S_XOR_B32 killed renamable $sgpr6, renamable $sgpr4, implicit-def dead $scc
+ ; GCN-NEXT: [[V_CVT_U32_F32_e64_3:%[0-9]+]]:vgpr_32 = V_CVT_U32_F32_e64 0, [[V_MUL_F32_e64_3]], 0, 0, implicit $mode, implicit $exec
+ ; GCN-NEXT: renamable $sgpr9 = S_SUB_I32 killed renamable $sgpr6, killed renamable $sgpr4, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr3 = nsw S_MUL_I32 renamable $sgpr9, killed renamable $sgpr3
+ ; GCN-NEXT: renamable $sgpr4 = V_READFIRSTLANE_B32 [[V_CVT_U32_F32_e64_3]], implicit $exec
+ ; GCN-NEXT: renamable $sgpr3 = nsw S_ADD_I32 killed renamable $sgpr3, killed renamable $sgpr5, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_SUB_I32 0, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_ASHR_I32 renamable $sgpr3, 31, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_MUL_I32 killed renamable $sgpr5, renamable $sgpr4
+ ; GCN-NEXT: renamable $sgpr10 = S_ADD_I32 renamable $sgpr3, renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_MUL_HI_U32 renamable $sgpr4, killed renamable $sgpr5
+ ; GCN-NEXT: renamable $sgpr10 = S_XOR_B32 killed renamable $sgpr10, renamable $sgpr6, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr4 = S_ADD_I32 killed renamable $sgpr4, killed renamable $sgpr5, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr4 = S_MUL_HI_U32 renamable $sgpr10, killed renamable $sgpr4
+ ; GCN-NEXT: renamable $sgpr11 = S_LOAD_DWORD_IMM renamable $sgpr0_sgpr1, 32, 0 :: (dereferenceable invariant load (s32), align 8, addrspace 4)
+ ; GCN-NEXT: renamable $sgpr5 = S_MUL_I32 renamable $sgpr4, renamable $sgpr8
+ ; GCN-NEXT: renamable $sgpr13 = S_XOR_B32 killed renamable $sgpr6, killed renamable $sgpr7, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_SUB_I32 killed renamable $sgpr10, killed renamable $sgpr5, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_ADD_I32 renamable $sgpr4, 1, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr7 = S_SUB_I32 renamable $sgpr5, renamable $sgpr8, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 renamable $sgpr5, renamable $sgpr8, implicit-def $scc
+ ; GCN-NEXT: renamable $sgpr10 = S_LOAD_DWORD_IMM renamable $sgpr0_sgpr1, 64, 0 :: (dereferenceable invariant load (s32), align 8, addrspace 4)
+ ; GCN-NEXT: renamable $sgpr4 = S_CSELECT_B32 killed renamable $sgpr6, killed renamable $sgpr4, implicit $scc
+ ; GCN-NEXT: renamable $sgpr5 = S_CSELECT_B32 killed renamable $sgpr7, killed renamable $sgpr5, implicit $scc
+ ; GCN-NEXT: renamable $sgpr6 = S_ADD_I32 renamable $sgpr4, 1, implicit-def dead $scc
+ ; GCN-NEXT: S_CMP_GE_U32 killed renamable $sgpr5, killed renamable $sgpr8, implicit-def $scc
+ ; GCN-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = nuw nsw V_LSHLREV_B32_e64 1, [[COPY]](s32), implicit $exec
+ ; GCN-NEXT: renamable $sgpr8 = S_CSELECT_B32 killed renamable $sgpr6, killed renamable $sgpr4, implicit $scc
+ ; GCN-NEXT: renamable $sgpr4_sgpr5_sgpr6_sgpr7 = S_LOAD_DWORDX4_IMM renamable $sgpr0_sgpr1, 0, 0 :: (dereferenceable invariant load (s128), addrspace 4)
+ ; GCN-NEXT: renamable $sgpr8 = S_XOR_B32 killed renamable $sgpr8, renamable $sgpr13, implicit-def dead $scc
+ ; GCN-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 1, [[COPY]](s32), implicit $exec
+ ; GCN-NEXT: renamable $sgpr8 = S_SUB_I32 killed renamable $sgpr8, killed renamable $sgpr13, implicit-def dead $scc
+ ; GCN-NEXT: [[V_AND_B32_e64_:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 6, [[V_LSHLREV_B32_e64_]], implicit $exec
+ ; GCN-NEXT: renamable $sgpr2 = S_MUL_I32 renamable $sgpr8, killed renamable $sgpr2
+ ; GCN-NEXT: renamable $sgpr9 = nsw S_SUB_I32 killed renamable $sgpr12, killed renamable $sgpr9, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr2 = S_SUB_I32 killed renamable $sgpr3, killed renamable $sgpr2, implicit-def dead $scc
+ ; GCN-NEXT: [[V_AND_B32_e64_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 126, [[V_LSHRREV_B32_e64_]], implicit $exec
+ ; GCN-NEXT: renamable $sgpr2 = S_ADD_I32 killed renamable $sgpr9, killed renamable $sgpr2, implicit-def dead $scc
+ ; GCN-NEXT: renamable $sgpr2 = S_LSHL_B32 killed renamable $sgpr2, 7, implicit-def dead $scc
+ ; GCN-NEXT: undef [[V_LSHLREV_B32_e64_1:%[0-9]+]].sub0:vreg_64 = nuw nsw V_LSHLREV_B32_e64 3, [[V_AND_B32_e64_]], implicit $exec
+ ; GCN-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = nsw V_ADD_U32_e64 $sgpr2, [[V_AND_B32_e64_1]], 0, implicit $exec
+ ; GCN-NEXT: early-clobber %157:vreg_64, $sgpr_null = V_MAD_U64_U32_gfx11_e64 [[V_ADD_U32_e64_]], $sgpr11, [[V_LSHLREV_B32_e64_1]], 0, implicit $exec
+ ; GCN-NEXT: renamable $sgpr15 = S_MOV_B32 268566528
+ ; GCN-NEXT: renamable $sgpr12 = COPY renamable $sgpr4
+ ; GCN-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 8, %157.sub0, 0, implicit $exec
+ ; GCN-NEXT: [[V_LSHLREV_B32_e64_2:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 1, %157.sub0, implicit $exec
+ ; GCN-NEXT: renamable $sgpr13 = COPY renamable $sgpr5
+ ; GCN-NEXT: renamable $sgpr14 = S_LSHL_B32 killed renamable $sgpr10, 1, implicit-def dead $scc
+ ; GCN-NEXT: [[V_ADD_LSHL_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_LSHL_U32_e64 %157.sub0, $sgpr11, 1, implicit $exec
+ ; GCN-NEXT: [[V_ADD_LSHL_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_LSHL_U32_e64 [[V_ADD_U32_e64_1]], killed $sgpr11, 1, implicit $exec
+ ; GCN-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[V_LSHLREV_B32_e64_2]], renamable $sgpr12_sgpr13_sgpr14_sgpr15, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s128), align 1, addrspace 7)
+ ; GCN-NEXT: [[V_LSHLREV_B32_e64_3:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 1, [[V_ADD_U32_e64_1]], implicit $exec
+ ; GCN-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[V_ADD_LSHL_U32_e64_]], renamable $sgpr12_sgpr13_sgpr14_sgpr15, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s128), align 1, addrspace 7)
+ ; GCN-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN2:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[V_ADD_LSHL_U32_e64_1]], renamable $sgpr12_sgpr13_sgpr14_sgpr15, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s128), align 1, addrspace 7)
+ ; GCN-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN3:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[V_LSHLREV_B32_e64_3]], renamable $sgpr12_sgpr13_sgpr14_sgpr15, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s128), align 1, addrspace 7)
+ ; GCN-NEXT: renamable $sgpr4 = S_LOAD_DWORD_IMM renamable $sgpr0_sgpr1, 84, 0 :: (dereferenceable invariant load (s32), addrspace 4)
+ ; GCN-NEXT: [[V_AND_B32_e64_2:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 252, [[COPY]](s32), implicit $exec
+ ; GCN-NEXT...
[truncated]
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.

No description provided.