Lines Matching refs:GCN
4 …fy generated machine code' | FileCheck -match-full-lines -strict-whitespace -check-prefix=GCN-O0 %s
6 …fy generated machine code' | FileCheck -match-full-lines -strict-whitespace -check-prefix=GCN-O1 %s
9 …nerated machine code' | FileCheck -match-full-lines -strict-whitespace -check-prefix=GCN-O1-OPTS %s
11 …fy generated machine code' | FileCheck -match-full-lines -strict-whitespace -check-prefix=GCN-O2 %s
13 …fy generated machine code' | FileCheck -match-full-lines -strict-whitespace -check-prefix=GCN-O3 %s
17 ; GCN-O0:Target Library Information
18 ; GCN-O0-NEXT:Target Pass Configuration
19 ; GCN-O0-NEXT:Machine Module Information
20 ; GCN-O0-NEXT:Target Transform Information
21 ; GCN-O0-NEXT:Assumption Cache Tracker
22 ; GCN-O0-NEXT:Profile summary info
23 ; GCN-O0-NEXT:Argument Register Usage Information Storage
24 ; GCN-O0-NEXT:Create Garbage Collector Module Metadata
25 ; GCN-O0-NEXT:Register Usage Information Storage
26 ; GCN-O0-NEXT:Machine Branch Probability Analysis
27 ; GCN-O0-NEXT: ModulePass Manager
28 ; GCN-O0-NEXT: Pre-ISel Intrinsic Lowering
29 ; GCN-O0-NEXT: AMDGPU Printf lowering
30 ; GCN-O0-NEXT: FunctionPass Manager
31 ; GCN-O0-NEXT: Dominator Tree Construction
32 ; GCN-O0-NEXT: Lower ctors and dtors for AMDGPU
33 ; GCN-O0-NEXT: FunctionPass Manager
34 ; GCN-O0-NEXT: Early propagate attributes from kernels to functions
35 ; GCN-O0-NEXT: AMDGPU Lower Intrinsics
36 ; GCN-O0-NEXT: AMDGPU Inline All Functions
37 ; GCN-O0-NEXT: CallGraph Construction
38 ; GCN-O0-NEXT: Call Graph SCC Pass Manager
39 ; GCN-O0-NEXT: Inliner for always_inline functions
40 ; GCN-O0-NEXT: A No-Op Barrier Pass
41 ; GCN-O0-NEXT: Lower OpenCL enqueued blocks
42 ; GCN-O0-NEXT: Lower uses of LDS variables from non-kernel functions
43 ; GCN-O0-NEXT: FunctionPass Manager
44 ; GCN-O0-NEXT: Expand Atomic instructions
45 ; GCN-O0-NEXT: Lower constant intrinsics
46 ; GCN-O0-NEXT: Remove unreachable blocks from the CFG
47 ; GCN-O0-NEXT: Expand vector predication intrinsics
48 ; GCN-O0-NEXT: Scalarize Masked Memory Intrinsics
49 ; GCN-O0-NEXT: Expand reduction intrinsics
50 ; GCN-O0-NEXT: AMDGPU Attributor
51 ; GCN-O0-NEXT: CallGraph Construction
52 ; GCN-O0-NEXT: Call Graph SCC Pass Manager
53 ; GCN-O0-NEXT: AMDGPU Annotate Kernel Features
54 ; GCN-O0-NEXT: FunctionPass Manager
55 ; GCN-O0-NEXT: AMDGPU Lower Kernel Arguments
56 ; GCN-O0-NEXT: Lazy Value Information Analysis
57 ; GCN-O0-NEXT: Lower SwitchInst's to branches
58 ; GCN-O0-NEXT: Lower invoke and unwind, for unwindless code generators
59 ; GCN-O0-NEXT: Remove unreachable blocks from the CFG
60 ; GCN-O0-NEXT: Post-Dominator Tree Construction
61 ; GCN-O0-NEXT: Dominator Tree Construction
62 ; GCN-O0-NEXT: Natural Loop Information
63 ; GCN-O0-NEXT: Legacy Divergence Analysis
64 ; GCN-O0-NEXT: Unify divergent function exit nodes
65 ; GCN-O0-NEXT: Lazy Value Information Analysis
66 ; GCN-O0-NEXT: Lower SwitchInst's to branches
67 ; GCN-O0-NEXT: Dominator Tree Construction
68 ; GCN-O0-NEXT: Natural Loop Information
69 ; GCN-O0-NEXT: Convert irreducible control-flow into natural loops
70 ; GCN-O0-NEXT: Fixup each natural loop to have a single exit block
71 ; GCN-O0-NEXT: Post-Dominator Tree Construction
72 ; GCN-O0-NEXT: Dominance Frontier Construction
73 ; GCN-O0-NEXT: Detect single entry single exit regions
74 ; GCN-O0-NEXT: Region Pass Manager
75 ; GCN-O0-NEXT: Structurize control flow
76 ; GCN-O0-NEXT: Post-Dominator Tree Construction
77 ; GCN-O0-NEXT: Natural Loop Information
78 ; GCN-O0-NEXT: Legacy Divergence Analysis
79 ; GCN-O0-NEXT: Basic Alias Analysis (stateless AA impl)
80 ; GCN-O0-NEXT: Function Alias Analysis Results
81 ; GCN-O0-NEXT: Memory SSA
82 ; GCN-O0-NEXT: AMDGPU Annotate Uniform Values
83 ; GCN-O0-NEXT: SI annotate control flow
84 ; GCN-O0-NEXT: LCSSA Verifier
85 ; GCN-O0-NEXT: Loop-Closed SSA Form Pass
86 ; GCN-O0-NEXT: DummyCGSCCPass
87 ; GCN-O0-NEXT: FunctionPass Manager
88 ; GCN-O0-NEXT: Safe Stack instrumentation pass
89 ; GCN-O0-NEXT: Insert stack protectors
90 ; GCN-O0-NEXT: Dominator Tree Construction
91 ; GCN-O0-NEXT: Post-Dominator Tree Construction
92 ; GCN-O0-NEXT: Natural Loop Information
93 ; GCN-O0-NEXT: Legacy Divergence Analysis
94 ; GCN-O0-NEXT: AMDGPU DAG->DAG Pattern Instruction Selection
95 ; GCN-O0-NEXT: MachineDominator Tree Construction
96 ; GCN-O0-NEXT: SI Fix SGPR copies
97 ; GCN-O0-NEXT: MachinePostDominator Tree Construction
98 ; GCN-O0-NEXT: SI Lower i1 Copies
99 ; GCN-O0-NEXT: Finalize ISel and expand pseudo-instructions
100 ; GCN-O0-NEXT: Local Stack Slot Allocation
101 ; GCN-O0-NEXT: Register Usage Information Propagation
102 ; GCN-O0-NEXT: Eliminate PHI nodes for register allocation
103 ; GCN-O0-NEXT: SI Lower control flow pseudo instructions
104 ; GCN-O0-NEXT: Two-Address instruction pass
105 ; GCN-O0-NEXT: MachineDominator Tree Construction
106 ; GCN-O0-NEXT: Slot index numbering
107 ; GCN-O0-NEXT: Live Interval Analysis
108 ; GCN-O0-NEXT: MachinePostDominator Tree Construction
109 ; GCN-O0-NEXT: SI Whole Quad Mode
110 ; GCN-O0-NEXT: Virtual Register Map
111 ; GCN-O0-NEXT: Live Register Matrix
112 ; GCN-O0-NEXT: SI Pre-allocate WWM Registers
113 ; GCN-O0-NEXT: Fast Register Allocator
114 ; GCN-O0-NEXT: SI lower SGPR spill instructions
115 ; GCN-O0-NEXT: Fast Register Allocator
116 ; GCN-O0-NEXT: SI Fix VGPR copies
117 ; GCN-O0-NEXT: Remove Redundant DEBUG_VALUE analysis
118 ; GCN-O0-NEXT: Fixup Statepoint Caller Saved
119 ; GCN-O0-NEXT: Lazy Machine Block Frequency Analysis
120 ; GCN-O0-NEXT: Machine Optimization Remark Emitter
121 ; GCN-O0-NEXT: Prologue/Epilogue Insertion & Frame Finalization
122 ; GCN-O0-NEXT: Post-RA pseudo instruction expansion pass
123 ; GCN-O0-NEXT: SI post-RA bundler
124 ; GCN-O0-NEXT: Insert fentry calls
125 ; GCN-O0-NEXT: Insert XRay ops
126 ; GCN-O0-NEXT: SI Memory Legalizer
127 ; GCN-O0-NEXT: MachineDominator Tree Construction
128 ; GCN-O0-NEXT: Machine Natural Loop Construction
129 ; GCN-O0-NEXT: MachinePostDominator Tree Construction
130 ; GCN-O0-NEXT: SI insert wait instructions
131 ; GCN-O0-NEXT: Insert required mode register values
132 ; GCN-O0-NEXT: SI Final Branch Preparation
133 ; GCN-O0-NEXT: Post RA hazard recognizer
134 ; GCN-O0-NEXT: Branch relaxation pass
135 ; GCN-O0-NEXT: Register Usage Information Collector Pass
136 ; GCN-O0-NEXT: Live DEBUG_VALUE analysis
137 ; GCN-O0-NEXT: Function register usage analysis
138 ; GCN-O0-NEXT: FunctionPass Manager
139 ; GCN-O0-NEXT: Lazy Machine Block Frequency Analysis
140 ; GCN-O0-NEXT: Machine Optimization Remark Emitter
141 ; GCN-O0-NEXT: AMDGPU Assembly Printer
142 ; GCN-O0-NEXT: Free MachineFunction
143 ; GCN-O0-NEXT:Pass Arguments: -domtree
144 ; GCN-O0-NEXT: FunctionPass Manager
145 ; GCN-O0-NEXT: Dominator Tree Construction
147 ; GCN-O1:Target Library Information
148 ; GCN-O1-NEXT:Target Pass Configuration
149 ; GCN-O1-NEXT:Machine Module Information
150 ; GCN-O1-NEXT:Target Transform Information
151 ; GCN-O1-NEXT:Assumption Cache Tracker
152 ; GCN-O1-NEXT:Profile summary info
153 ; GCN-O1-NEXT:AMDGPU Address space based Alias Analysis
154 ; GCN-O1-NEXT:External Alias Analysis
155 ; GCN-O1-NEXT:Type-Based Alias Analysis
156 ; GCN-O1-NEXT:Scoped NoAlias Alias Analysis
157 ; GCN-O1-NEXT:Argument Register Usage Information Storage
158 ; GCN-O1-NEXT:Create Garbage Collector Module Metadata
159 ; GCN-O1-NEXT:Machine Branch Probability Analysis
160 ; GCN-O1-NEXT:Register Usage Information Storage
161 ; GCN-O1-NEXT:Default Regalloc Eviction Advisor
162 ; GCN-O1-NEXT: ModulePass Manager
163 ; GCN-O1-NEXT: Pre-ISel Intrinsic Lowering
164 ; GCN-O1-NEXT: AMDGPU Printf lowering
165 ; GCN-O1-NEXT: FunctionPass Manager
166 ; GCN-O1-NEXT: Dominator Tree Construction
167 ; GCN-O1-NEXT: Lower ctors and dtors for AMDGPU
168 ; GCN-O1-NEXT: FunctionPass Manager
169 ; GCN-O1-NEXT: Early propagate attributes from kernels to functions
170 ; GCN-O1-NEXT: AMDGPU Lower Intrinsics
171 ; GCN-O1-NEXT: AMDGPU Inline All Functions
172 ; GCN-O1-NEXT: CallGraph Construction
173 ; GCN-O1-NEXT: Call Graph SCC Pass Manager
174 ; GCN-O1-NEXT: Inliner for always_inline functions
175 ; GCN-O1-NEXT: A No-Op Barrier Pass
176 ; GCN-O1-NEXT: Lower OpenCL enqueued blocks
177 ; GCN-O1-NEXT: Lower uses of LDS variables from non-kernel functions
178 ; GCN-O1-NEXT: FunctionPass Manager
179 ; GCN-O1-NEXT: Infer address spaces
180 ; GCN-O1-NEXT: Expand Atomic instructions
181 ; GCN-O1-NEXT: AMDGPU Promote Alloca
182 ; GCN-O1-NEXT: Dominator Tree Construction
183 ; GCN-O1-NEXT: SROA
184 ; GCN-O1-NEXT: Post-Dominator Tree Construction
185 ; GCN-O1-NEXT: Natural Loop Information
186 ; GCN-O1-NEXT: Legacy Divergence Analysis
187 ; GCN-O1-NEXT: AMDGPU IR optimizations
188 ; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
189 ; GCN-O1-NEXT: Canonicalize natural loops
190 ; GCN-O1-NEXT: Scalar Evolution Analysis
191 ; GCN-O1-NEXT: Loop Pass Manager
192 ; GCN-O1-NEXT: Canonicalize Freeze Instructions in Loops
193 ; GCN-O1-NEXT: Induction Variable Users
194 ; GCN-O1-NEXT: Loop Strength Reduction
195 ; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
196 ; GCN-O1-NEXT: Function Alias Analysis Results
197 ; GCN-O1-NEXT: Merge contiguous icmps into a memcmp
198 ; GCN-O1-NEXT: Natural Loop Information
199 ; GCN-O1-NEXT: Lazy Branch Probability Analysis
200 ; GCN-O1-NEXT: Lazy Block Frequency Analysis
201 ; GCN-O1-NEXT: Expand memcmp() to load/stores
202 ; GCN-O1-NEXT: Lower constant intrinsics
203 ; GCN-O1-NEXT: Remove unreachable blocks from the CFG
204 ; GCN-O1-NEXT: Natural Loop Information
205 ; GCN-O1-NEXT: Post-Dominator Tree Construction
206 ; GCN-O1-NEXT: Branch Probability Analysis
207 ; GCN-O1-NEXT: Block Frequency Analysis
208 ; GCN-O1-NEXT: Constant Hoisting
209 ; GCN-O1-NEXT: Replace intrinsics with calls to vector library
210 ; GCN-O1-NEXT: Partially inline calls to library functions
211 ; GCN-O1-NEXT: Expand vector predication intrinsics
212 ; GCN-O1-NEXT: Scalarize Masked Memory Intrinsics
213 ; GCN-O1-NEXT: Expand reduction intrinsics
214 ; GCN-O1-NEXT: Natural Loop Information
215 ; GCN-O1-NEXT: TLS Variable Hoist
216 ; GCN-O1-NEXT: AMDGPU Attributor
217 ; GCN-O1-NEXT: CallGraph Construction
218 ; GCN-O1-NEXT: Call Graph SCC Pass Manager
219 ; GCN-O1-NEXT: AMDGPU Annotate Kernel Features
220 ; GCN-O1-NEXT: FunctionPass Manager
221 ; GCN-O1-NEXT: AMDGPU Lower Kernel Arguments
222 ; GCN-O1-NEXT: Dominator Tree Construction
223 ; GCN-O1-NEXT: Natural Loop Information
224 ; GCN-O1-NEXT: CodeGen Prepare
225 ; GCN-O1-NEXT: Lazy Value Information Analysis
226 ; GCN-O1-NEXT: Lower SwitchInst's to branches
227 ; GCN-O1-NEXT: Lower invoke and unwind, for unwindless code generators
228 ; GCN-O1-NEXT: Remove unreachable blocks from the CFG
229 ; GCN-O1-NEXT: Dominator Tree Construction
230 ; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
231 ; GCN-O1-NEXT: Function Alias Analysis Results
232 ; GCN-O1-NEXT: Flatten the CFG
233 ; GCN-O1-NEXT: Dominator Tree Construction
234 ; GCN-O1-NEXT: Post-Dominator Tree Construction
235 ; GCN-O1-NEXT: Natural Loop Information
236 ; GCN-O1-NEXT: Legacy Divergence Analysis
237 ; GCN-O1-NEXT: AMDGPU IR late optimizations
238 ; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
239 ; GCN-O1-NEXT: Function Alias Analysis Results
240 ; GCN-O1-NEXT: Code sinking
241 ; GCN-O1-NEXT: Legacy Divergence Analysis
242 ; GCN-O1-NEXT: Unify divergent function exit nodes
243 ; GCN-O1-NEXT: Lazy Value Information Analysis
244 ; GCN-O1-NEXT: Lower SwitchInst's to branches
245 ; GCN-O1-NEXT: Dominator Tree Construction
246 ; GCN-O1-NEXT: Natural Loop Information
247 ; GCN-O1-NEXT: Convert irreducible control-flow into natural loops
248 ; GCN-O1-NEXT: Fixup each natural loop to have a single exit block
249 ; GCN-O1-NEXT: Post-Dominator Tree Construction
250 ; GCN-O1-NEXT: Dominance Frontier Construction
251 ; GCN-O1-NEXT: Detect single entry single exit regions
252 ; GCN-O1-NEXT: Region Pass Manager
253 ; GCN-O1-NEXT: Structurize control flow
254 ; GCN-O1-NEXT: Post-Dominator Tree Construction
255 ; GCN-O1-NEXT: Natural Loop Information
256 ; GCN-O1-NEXT: Legacy Divergence Analysis
257 ; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
258 ; GCN-O1-NEXT: Function Alias Analysis Results
259 ; GCN-O1-NEXT: Memory SSA
260 ; GCN-O1-NEXT: AMDGPU Annotate Uniform Values
261 ; GCN-O1-NEXT: SI annotate control flow
262 ; GCN-O1-NEXT: LCSSA Verifier
263 ; GCN-O1-NEXT: Loop-Closed SSA Form Pass
264 ; GCN-O1-NEXT: DummyCGSCCPass
265 ; GCN-O1-NEXT: FunctionPass Manager
266 ; GCN-O1-NEXT: Safe Stack instrumentation pass
267 ; GCN-O1-NEXT: Insert stack protectors
268 ; GCN-O1-NEXT: Dominator Tree Construction
269 ; GCN-O1-NEXT: Post-Dominator Tree Construction
270 ; GCN-O1-NEXT: Natural Loop Information
271 ; GCN-O1-NEXT: Legacy Divergence Analysis
272 ; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
273 ; GCN-O1-NEXT: Function Alias Analysis Results
274 ; GCN-O1-NEXT: Branch Probability Analysis
275 ; GCN-O1-NEXT: Lazy Branch Probability Analysis
276 ; GCN-O1-NEXT: Lazy Block Frequency Analysis
277 ; GCN-O1-NEXT: AMDGPU DAG->DAG Pattern Instruction Selection
278 ; GCN-O1-NEXT: MachineDominator Tree Construction
279 ; GCN-O1-NEXT: SI Fix SGPR copies
280 ; GCN-O1-NEXT: MachinePostDominator Tree Construction
281 ; GCN-O1-NEXT: SI Lower i1 Copies
282 ; GCN-O1-NEXT: Finalize ISel and expand pseudo-instructions
283 ; GCN-O1-NEXT: Lazy Machine Block Frequency Analysis
284 ; GCN-O1-NEXT: Early Tail Duplication
285 ; GCN-O1-NEXT: Optimize machine instruction PHIs
286 ; GCN-O1-NEXT: Slot index numbering
287 ; GCN-O1-NEXT: Merge disjoint stack slots
288 ; GCN-O1-NEXT: Local Stack Slot Allocation
289 ; GCN-O1-NEXT: Remove dead machine instructions
290 ; GCN-O1-NEXT: MachineDominator Tree Construction
291 ; GCN-O1-NEXT: Machine Natural Loop Construction
292 ; GCN-O1-NEXT: Machine Block Frequency Analysis
293 ; GCN-O1-NEXT: Early Machine Loop Invariant Code Motion
294 ; GCN-O1-NEXT: MachineDominator Tree Construction
295 ; GCN-O1-NEXT: Machine Block Frequency Analysis
296 ; GCN-O1-NEXT: Machine Common Subexpression Elimination
297 ; GCN-O1-NEXT: MachinePostDominator Tree Construction
298 ; GCN-O1-NEXT: Machine Cycle Info Analysis
299 ; GCN-O1-NEXT: Machine code sinking
300 ; GCN-O1-NEXT: Peephole Optimizations
301 ; GCN-O1-NEXT: Remove dead machine instructions
302 ; GCN-O1-NEXT: SI Fold Operands
303 ; GCN-O1-NEXT: GCN DPP Combine
304 ; GCN-O1-NEXT: SI Load Store Optimizer
305 ; GCN-O1-NEXT: Remove dead machine instructions
306 ; GCN-O1-NEXT: SI Shrink Instructions
307 ; GCN-O1-NEXT: Register Usage Information Propagation
308 ; GCN-O1-NEXT: Detect Dead Lanes
309 ; GCN-O1-NEXT: Remove dead machine instructions
310 ; GCN-O1-NEXT: Process Implicit Definitions
311 ; GCN-O1-NEXT: Remove unreachable machine basic blocks
312 ; GCN-O1-NEXT: Live Variable Analysis
313 ; GCN-O1-NEXT: MachineDominator Tree Construction
314 ; GCN-O1-NEXT: SI Optimize VGPR LiveRange
315 ; GCN-O1-NEXT: Eliminate PHI nodes for register allocation
316 ; GCN-O1-NEXT: SI Lower control flow pseudo instructions
317 ; GCN-O1-NEXT: Two-Address instruction pass
318 ; GCN-O1-NEXT: Slot index numbering
319 ; GCN-O1-NEXT: Live Interval Analysis
320 ; GCN-O1-NEXT: Machine Natural Loop Construction
321 ; GCN-O1-NEXT: Simple Register Coalescing
322 ; GCN-O1-NEXT: Rename Disconnected Subregister Components
323 ; GCN-O1-NEXT: Machine Instruction Scheduler
324 ; GCN-O1-NEXT: MachinePostDominator Tree Construction
325 ; GCN-O1-NEXT: SI Whole Quad Mode
326 ; GCN-O1-NEXT: Virtual Register Map
327 ; GCN-O1-NEXT: Live Register Matrix
328 ; GCN-O1-NEXT: SI Pre-allocate WWM Registers
329 ; GCN-O1-NEXT: SI optimize exec mask operations pre-RA
330 ; GCN-O1-NEXT: Machine Natural Loop Construction
331 ; GCN-O1-NEXT: Machine Block Frequency Analysis
332 ; GCN-O1-NEXT: Debug Variable Analysis
333 ; GCN-O1-NEXT: Live Stack Slot Analysis
334 ; GCN-O1-NEXT: Virtual Register Map
335 ; GCN-O1-NEXT: Live Register Matrix
336 ; GCN-O1-NEXT: Bundle Machine CFG Edges
337 ; GCN-O1-NEXT: Spill Code Placement Analysis
338 ; GCN-O1-NEXT: Lazy Machine Block Frequency Analysis
339 ; GCN-O1-NEXT: Machine Optimization Remark Emitter
340 ; GCN-O1-NEXT: Greedy Register Allocator
341 ; GCN-O1-NEXT: Virtual Register Rewriter
342 ; GCN-O1-NEXT: SI lower SGPR spill instructions
343 ; GCN-O1-NEXT: Virtual Register Map
344 ; GCN-O1-NEXT: Live Register Matrix
345 ; GCN-O1-NEXT: Greedy Register Allocator
346 ; GCN-O1-NEXT: GCN NSA Reassign
347 ; GCN-O1-NEXT: Virtual Register Rewriter
348 ; GCN-O1-NEXT: Stack Slot Coloring
349 ; GCN-O1-NEXT: Machine Copy Propagation Pass
350 ; GCN-O1-NEXT: Machine Loop Invariant Code Motion
351 ; GCN-O1-NEXT: SI Fix VGPR copies
352 ; GCN-O1-NEXT: SI optimize exec mask operations
353 ; GCN-O1-NEXT: Remove Redundant DEBUG_VALUE analysis
354 ; GCN-O1-NEXT: Fixup Statepoint Caller Saved
355 ; GCN-O1-NEXT: PostRA Machine Sink
356 ; GCN-O1-NEXT: MachineDominator Tree Construction
357 ; GCN-O1-NEXT: Machine Natural Loop Construction
358 ; GCN-O1-NEXT: Machine Block Frequency Analysis
359 ; GCN-O1-NEXT: MachinePostDominator Tree Construction
360 ; GCN-O1-NEXT: Lazy Machine Block Frequency Analysis
361 ; GCN-O1-NEXT: Machine Optimization Remark Emitter
362 ; GCN-O1-NEXT: Shrink Wrapping analysis
363 ; GCN-O1-NEXT: Prologue/Epilogue Insertion & Frame Finalization
364 ; GCN-O1-NEXT: Control Flow Optimizer
365 ; GCN-O1-NEXT: Lazy Machine Block Frequency Analysis
366 ; GCN-O1-NEXT: Tail Duplication
367 ; GCN-O1-NEXT: Machine Copy Propagation Pass
368 ; GCN-O1-NEXT: Post-RA pseudo instruction expansion pass
369 ; GCN-O1-NEXT: SI Shrink Instructions
370 ; GCN-O1-NEXT: SI post-RA bundler
371 ; GCN-O1-NEXT: MachineDominator Tree Construction
372 ; GCN-O1-NEXT: Machine Natural Loop Construction
373 ; GCN-O1-NEXT: PostRA Machine Instruction Scheduler
374 ; GCN-O1-NEXT: Machine Block Frequency Analysis
375 ; GCN-O1-NEXT: MachinePostDominator Tree Construction
376 ; GCN-O1-NEXT: Branch Probability Basic Block Placement
377 ; GCN-O1-NEXT: Insert fentry calls
378 ; GCN-O1-NEXT: Insert XRay ops
379 ; GCN-O1-NEXT: GCN Create VOPD Instructions
380 ; GCN-O1-NEXT: SI Memory Legalizer
381 ; GCN-O1-NEXT: MachineDominator Tree Construction
382 ; GCN-O1-NEXT: Machine Natural Loop Construction
383 ; GCN-O1-NEXT: MachinePostDominator Tree Construction
384 ; GCN-O1-NEXT: SI insert wait instructions
385 ; GCN-O1-NEXT: Insert required mode register values
386 ; GCN-O1-NEXT: SI Insert Hard Clauses
387 ; GCN-O1-NEXT: SI Final Branch Preparation
388 ; GCN-O1-NEXT: SI peephole optimizations
389 ; GCN-O1-NEXT: Post RA hazard recognizer
390 ; GCN-O1-NEXT: AMDGPU Insert Delay ALU
391 ; GCN-O1-NEXT: Branch relaxation pass
392 ; GCN-O1-NEXT: Register Usage Information Collector Pass
393 ; GCN-O1-NEXT: Live DEBUG_VALUE analysis
394 ; GCN-O1-NEXT: Function register usage analysis
395 ; GCN-O1-NEXT: FunctionPass Manager
396 ; GCN-O1-NEXT: Lazy Machine Block Frequency Analysis
397 ; GCN-O1-NEXT: Machine Optimization Remark Emitter
398 ; GCN-O1-NEXT: AMDGPU Assembly Printer
399 ; GCN-O1-NEXT: Free MachineFunction
400 ; GCN-O1-NEXT:Pass Arguments: -domtree
401 ; GCN-O1-NEXT: FunctionPass Manager
402 ; GCN-O1-NEXT: Dominator Tree Construction
404 ; GCN-O1-OPTS:Target Library Information
405 ; GCN-O1-OPTS-NEXT:Target Pass Configuration
406 ; GCN-O1-OPTS-NEXT:Machine Module Information
407 ; GCN-O1-OPTS-NEXT:Target Transform Information
408 ; GCN-O1-OPTS-NEXT:Assumption Cache Tracker
409 ; GCN-O1-OPTS-NEXT:Profile summary info
410 ; GCN-O1-OPTS-NEXT:AMDGPU Address space based Alias Analysis
411 ; GCN-O1-OPTS-NEXT:External Alias Analysis
412 ; GCN-O1-OPTS-NEXT:Type-Based Alias Analysis
413 ; GCN-O1-OPTS-NEXT:Scoped NoAlias Alias Analysis
414 ; GCN-O1-OPTS-NEXT:Argument Register Usage Information Storage
415 ; GCN-O1-OPTS-NEXT:Create Garbage Collector Module Metadata
416 ; GCN-O1-OPTS-NEXT:Machine Branch Probability Analysis
417 ; GCN-O1-OPTS-NEXT:Register Usage Information Storage
418 ; GCN-O1-OPTS-NEXT:Default Regalloc Eviction Advisor
419 ; GCN-O1-OPTS-NEXT: ModulePass Manager
420 ; GCN-O1-OPTS-NEXT: Pre-ISel Intrinsic Lowering
421 ; GCN-O1-OPTS-NEXT: AMDGPU Printf lowering
422 ; GCN-O1-OPTS-NEXT: FunctionPass Manager
423 ; GCN-O1-OPTS-NEXT: Dominator Tree Construction
424 ; GCN-O1-OPTS-NEXT: Lower ctors and dtors for AMDGPU
425 ; GCN-O1-OPTS-NEXT: FunctionPass Manager
426 ; GCN-O1-OPTS-NEXT: Early propagate attributes from kernels to functions
427 ; GCN-O1-OPTS-NEXT: AMDGPU Lower Intrinsics
428 ; GCN-O1-OPTS-NEXT: AMDGPU Inline All Functions
429 ; GCN-O1-OPTS-NEXT: CallGraph Construction
430 ; GCN-O1-OPTS-NEXT: Call Graph SCC Pass Manager
431 ; GCN-O1-OPTS-NEXT: Inliner for always_inline functions
432 ; GCN-O1-OPTS-NEXT: A No-Op Barrier Pass
433 ; GCN-O1-OPTS-NEXT: Lower OpenCL enqueued blocks
434 ; GCN-O1-OPTS-NEXT: Lower uses of LDS variables from non-kernel functions
435 ; GCN-O1-OPTS-NEXT: FunctionPass Manager
436 ; GCN-O1-OPTS-NEXT: Infer address spaces
437 ; GCN-O1-OPTS-NEXT: Expand Atomic instructions
438 ; GCN-O1-OPTS-NEXT: AMDGPU Promote Alloca
439 ; GCN-O1-OPTS-NEXT: Dominator Tree Construction
440 ; GCN-O1-OPTS-NEXT: SROA
441 ; GCN-O1-OPTS-NEXT: Basic Alias Analysis (stateless AA impl)
442 ; GCN-O1-OPTS-NEXT: Function Alias Analysis Results
443 ; GCN-O1-OPTS-NEXT: Memory SSA
444 ; GCN-O1-OPTS-NEXT: Natural Loop Information
445 ; GCN-O1-OPTS-NEXT: Canonicalize natural loops
446 ; GCN-O1-OPTS-NEXT: LCSSA Verifier
447 ; GCN-O1-OPTS-NEXT: Loop-Closed SSA Form Pass
448 ; GCN-O1-OPTS-NEXT: Scalar Evolution Analysis
449 ; GCN-O1-OPTS-NEXT: Lazy Branch Probability Analysis
450 ; GCN-O1-OPTS-NEXT: Lazy Block Frequency Analysis
451 ; GCN-O1-OPTS-NEXT: Loop Pass Manager
452 ; GCN-O1-OPTS-NEXT: Loop Invariant Code Motion
453 ; GCN-O1-OPTS-NEXT: Split GEPs to a variadic base and a constant offset for better CSE
454 ; GCN-O1-OPTS-NEXT: Speculatively execute instructions
455 ; GCN-O1-OPTS-NEXT: Scalar Evolution Analysis
456 ; GCN-O1-OPTS-NEXT: Straight line strength reduction
457 ; GCN-O1-OPTS-NEXT: Early CSE
458 ; GCN-O1-OPTS-NEXT: Scalar Evolution Analysis
459 ; GCN-O1-OPTS-NEXT: Nary reassociation
460 ; GCN-O1-OPTS-NEXT: Early CSE
461 ; GCN-O1-OPTS-NEXT: Post-Dominator Tree Construction
462 ; GCN-O1-OPTS-NEXT: Legacy Divergence Analysis
463 ; GCN-O1-OPTS-NEXT: AMDGPU IR optimizations
464 ; GCN-O1-OPTS-NEXT: Canonicalize natural loops
465 ; GCN-O1-OPTS-NEXT: Scalar Evolution Analysis
466 ; GCN-O1-OPTS-NEXT: Loop Pass Manager
467 ; GCN-O1-OPTS-NEXT: Canonicalize Freeze Instructions in Loops
468 ; GCN-O1-OPTS-NEXT: Induction Variable Users
469 ; GCN-O1-OPTS-NEXT: Loop Strength Reduction
470 ; GCN-O1-OPTS-NEXT: Basic Alias Analysis (stateless AA impl)
471 ; GCN-O1-OPTS-NEXT: Function Alias Analysis Results
472 ; GCN-O1-OPTS-NEXT: Merge contiguous icmps into a memcmp
473 ; GCN-O1-OPTS-NEXT: Natural Loop Information
474 ; GCN-O1-OPTS-NEXT: Lazy Branch Probability Analysis
475 ; GCN-O1-OPTS-NEXT: Lazy Block Frequency Analysis
476 ; GCN-O1-OPTS-NEXT: Expand memcmp() to load/stores
477 ; GCN-O1-OPTS-NEXT: Lower constant intrinsics
478 ; GCN-O1-OPTS-NEXT: Remove unreachable blocks from the CFG
479 ; GCN-O1-OPTS-NEXT: Natural Loop Information
480 ; GCN-O1-OPTS-NEXT: Post-Dominator Tree Construction
481 ; GCN-O1-OPTS-NEXT: Branch Probability Analysis
482 ; GCN-O1-OPTS-NEXT: Block Frequency Analysis
483 ; GCN-O1-OPTS-NEXT: Constant Hoisting
484 ; GCN-O1-OPTS-NEXT: Replace intrinsics with calls to vector library
485 ; GCN-O1-OPTS-NEXT: Partially inline calls to library functions
486 ; GCN-O1-OPTS-NEXT: Expand vector predication intrinsics
487 ; GCN-O1-OPTS-NEXT: Scalarize Masked Memory Intrinsics
488 ; GCN-O1-OPTS-NEXT: Expand reduction intrinsics
489 ; GCN-O1-OPTS-NEXT: Natural Loop Information
490 ; GCN-O1-OPTS-NEXT: TLS Variable Hoist
491 ; GCN-O1-OPTS-NEXT: Early CSE
492 ; GCN-O1-OPTS-NEXT: AMDGPU Attributor
493 ; GCN-O1-OPTS-NEXT: CallGraph Construction
494 ; GCN-O1-OPTS-NEXT: Call Graph SCC Pass Manager
495 ; GCN-O1-OPTS-NEXT: AMDGPU Annotate Kernel Features
496 ; GCN-O1-OPTS-NEXT: FunctionPass Manager
497 ; GCN-O1-OPTS-NEXT: AMDGPU Lower Kernel Arguments
498 ; GCN-O1-OPTS-NEXT: Dominator Tree Construction
499 ; GCN-O1-OPTS-NEXT: Natural Loop Information
500 ; GCN-O1-OPTS-NEXT: CodeGen Prepare
501 ; GCN-O1-OPTS-NEXT: Dominator Tree Construction
502 ; GCN-O1-OPTS-NEXT: Basic Alias Analysis (stateless AA impl)
503 ; GCN-O1-OPTS-NEXT: Function Alias Analysis Results
504 ; GCN-O1-OPTS-NEXT: Natural Loop Information
505 ; GCN-O1-OPTS-NEXT: Scalar Evolution Analysis
506 ; GCN-O1-OPTS-NEXT: GPU Load and Store Vectorizer
507 ; GCN-O1-OPTS-NEXT: Lazy Value Information Analysis
508 ; GCN-O1-OPTS-NEXT: Lower SwitchInst's to branches
509 ; GCN-O1-OPTS-NEXT: Lower invoke and unwind, for unwindless code generators
510 ; GCN-O1-OPTS-NEXT: Remove unreachable blocks from the CFG
511 ; GCN-O1-OPTS-NEXT: Dominator Tree Construction
512 ; GCN-O1-OPTS-NEXT: Basic Alias Analysis (stateless AA impl)
513 ; GCN-O1-OPTS-NEXT: Function Alias Analysis Results
514 ; GCN-O1-OPTS-NEXT: Flatten the CFG
515 ; GCN-O1-OPTS-NEXT: Dominator Tree Construction
516 ; GCN-O1-OPTS-NEXT: Post-Dominator Tree Construction
517 ; GCN-O1-OPTS-NEXT: Natural Loop Information
518 ; GCN-O1-OPTS-NEXT: Legacy Divergence Analysis
519 ; GCN-O1-OPTS-NEXT: AMDGPU IR late optimizations
520 ; GCN-O1-OPTS-NEXT: Basic Alias Analysis (stateless AA impl)
521 ; GCN-O1-OPTS-NEXT: Function Alias Analysis Results
522 ; GCN-O1-OPTS-NEXT: Code sinking
523 ; GCN-O1-OPTS-NEXT: Legacy Divergence Analysis
524 ; GCN-O1-OPTS-NEXT: Unify divergent function exit nodes
525 ; GCN-O1-OPTS-NEXT: Lazy Value Information Analysis
526 ; GCN-O1-OPTS-NEXT: Lower SwitchInst's to branches
527 ; GCN-O1-OPTS-NEXT: Dominator Tree Construction
528 ; GCN-O1-OPTS-NEXT: Natural Loop Information
529 ; GCN-O1-OPTS-NEXT: Convert irreducible control-flow into natural loops
530 ; GCN-O1-OPTS-NEXT: Fixup each natural loop to have a single exit block
531 ; GCN-O1-OPTS-NEXT: Post-Dominator Tree Construction
532 ; GCN-O1-OPTS-NEXT: Dominance Frontier Construction
533 ; GCN-O1-OPTS-NEXT: Detect single entry single exit regions
534 ; GCN-O1-OPTS-NEXT: Region Pass Manager
535 ; GCN-O1-OPTS-NEXT: Structurize control flow
536 ; GCN-O1-OPTS-NEXT: Post-Dominator Tree Construction
537 ; GCN-O1-OPTS-NEXT: Natural Loop Information
538 ; GCN-O1-OPTS-NEXT: Legacy Divergence Analysis
539 ; GCN-O1-OPTS-NEXT: Basic Alias Analysis (stateless AA impl)
540 ; GCN-O1-OPTS-NEXT: Function Alias Analysis Results
541 ; GCN-O1-OPTS-NEXT: Memory SSA
542 ; GCN-O1-OPTS-NEXT: AMDGPU Annotate Uniform Values
543 ; GCN-O1-OPTS-NEXT: SI annotate control flow
544 ; GCN-O1-OPTS-NEXT: LCSSA Verifier
545 ; GCN-O1-OPTS-NEXT: Loop-Closed SSA Form Pass
546 ; GCN-O1-OPTS-NEXT: DummyCGSCCPass
547 ; GCN-O1-OPTS-NEXT: FunctionPass Manager
548 ; GCN-O1-OPTS-NEXT: Safe Stack instrumentation pass
549 ; GCN-O1-OPTS-NEXT: Insert stack protectors
550 ; GCN-O1-OPTS-NEXT: Dominator Tree Construction
551 ; GCN-O1-OPTS-NEXT: Post-Dominator Tree Construction
552 ; GCN-O1-OPTS-NEXT: Natural Loop Information
553 ; GCN-O1-OPTS-NEXT: Legacy Divergence Analysis
554 ; GCN-O1-OPTS-NEXT: Basic Alias Analysis (stateless AA impl)
555 ; GCN-O1-OPTS-NEXT: Function Alias Analysis Results
556 ; GCN-O1-OPTS-NEXT: Branch Probability Analysis
557 ; GCN-O1-OPTS-NEXT: Lazy Branch Probability Analysis
558 ; GCN-O1-OPTS-NEXT: Lazy Block Frequency Analysis
559 ; GCN-O1-OPTS-NEXT: AMDGPU DAG->DAG Pattern Instruction Selection
560 ; GCN-O1-OPTS-NEXT: MachineDominator Tree Construction
561 ; GCN-O1-OPTS-NEXT: SI Fix SGPR copies
562 ; GCN-O1-OPTS-NEXT: MachinePostDominator Tree Construction
563 ; GCN-O1-OPTS-NEXT: SI Lower i1 Copies
564 ; GCN-O1-OPTS-NEXT: Finalize ISel and expand pseudo-instructions
565 ; GCN-O1-OPTS-NEXT: Lazy Machine Block Frequency Analysis
566 ; GCN-O1-OPTS-NEXT: Early Tail Duplication
567 ; GCN-O1-OPTS-NEXT: Optimize machine instruction PHIs
568 ; GCN-O1-OPTS-NEXT: Slot index numbering
569 ; GCN-O1-OPTS-NEXT: Merge disjoint stack slots
570 ; GCN-O1-OPTS-NEXT: Local Stack Slot Allocation
571 ; GCN-O1-OPTS-NEXT: Remove dead machine instructions
572 ; GCN-O1-OPTS-NEXT: MachineDominator Tree Construction
573 ; GCN-O1-OPTS-NEXT: Machine Natural Loop Construction
574 ; GCN-O1-OPTS-NEXT: Machine Block Frequency Analysis
575 ; GCN-O1-OPTS-NEXT: Early Machine Loop Invariant Code Motion
576 ; GCN-O1-OPTS-NEXT: MachineDominator Tree Construction
577 ; GCN-O1-OPTS-NEXT: Machine Block Frequency Analysis
578 ; GCN-O1-OPTS-NEXT: Machine Common Subexpression Elimination
579 ; GCN-O1-OPTS-NEXT: MachinePostDominator Tree Construction
580 ; GCN-O1-OPTS-NEXT: Machine Cycle Info Analysis
581 ; GCN-O1-OPTS-NEXT: Machine code sinking
582 ; GCN-O1-OPTS-NEXT: Peephole Optimizations
583 ; GCN-O1-OPTS-NEXT: Remove dead machine instructions
584 ; GCN-O1-OPTS-NEXT: SI Fold Operands
585 ; GCN-O1-OPTS-NEXT: GCN DPP Combine
586 ; GCN-O1-OPTS-NEXT: SI Load Store Optimizer
587 ; GCN-O1-OPTS-NEXT: SI Peephole SDWA
588 ; GCN-O1-OPTS-NEXT: Machine Block Frequency Analysis
589 ; GCN-O1-OPTS-NEXT: MachineDominator Tree Construction
590 ; GCN-O1-OPTS-NEXT: Early Machine Loop Invariant Code Motion
591 ; GCN-O1-OPTS-NEXT: MachineDominator Tree Construction
592 ; GCN-O1-OPTS-NEXT: Machine Block Frequency Analysis
593 ; GCN-O1-OPTS-NEXT: Machine Common Subexpression Elimination
594 ; GCN-O1-OPTS-NEXT: SI Fold Operands
595 ; GCN-O1-OPTS-NEXT: Remove dead machine instructions
596 ; GCN-O1-OPTS-NEXT: SI Shrink Instructions
597 ; GCN-O1-OPTS-NEXT: Register Usage Information Propagation
598 ; GCN-O1-OPTS-NEXT: Detect Dead Lanes
599 ; GCN-O1-OPTS-NEXT: Remove dead machine instructions
600 ; GCN-O1-OPTS-NEXT: Process Implicit Definitions
601 ; GCN-O1-OPTS-NEXT: Remove unreachable machine basic blocks
602 ; GCN-O1-OPTS-NEXT: Live Variable Analysis
603 ; GCN-O1-OPTS-NEXT: SI Optimize VGPR LiveRange
604 ; GCN-O1-OPTS-NEXT: Eliminate PHI nodes for register allocation
605 ; GCN-O1-OPTS-NEXT: SI Lower control flow pseudo instructions
606 ; GCN-O1-OPTS-NEXT: Two-Address instruction pass
607 ; GCN-O1-OPTS-NEXT: Slot index numbering
608 ; GCN-O1-OPTS-NEXT: Live Interval Analysis
609 ; GCN-O1-OPTS-NEXT: Machine Natural Loop Construction
610 ; GCN-O1-OPTS-NEXT: Simple Register Coalescing
611 ; GCN-O1-OPTS-NEXT: Rename Disconnected Subregister Components
612 ; GCN-O1-OPTS-NEXT: AMDGPU Pre-RA optimizations
613 ; GCN-O1-OPTS-NEXT: Machine Instruction Scheduler
614 ; GCN-O1-OPTS-NEXT: MachinePostDominator Tree Construction
615 ; GCN-O1-OPTS-NEXT: SI Whole Quad Mode
616 ; GCN-O1-OPTS-NEXT: Virtual Register Map
617 ; GCN-O1-OPTS-NEXT: Live Register Matrix
618 ; GCN-O1-OPTS-NEXT: SI Pre-allocate WWM Registers
619 ; GCN-O1-OPTS-NEXT: SI optimize exec mask operations pre-RA
620 ; GCN-O1-OPTS-NEXT: Machine Natural Loop Construction
621 ; GCN-O1-OPTS-NEXT: Machine Block Frequency Analysis
622 ; GCN-O1-OPTS-NEXT: Debug Variable Analysis
623 ; GCN-O1-OPTS-NEXT: Live Stack Slot Analysis
624 ; GCN-O1-OPTS-NEXT: Virtual Register Map
625 ; GCN-O1-OPTS-NEXT: Live Register Matrix
626 ; GCN-O1-OPTS-NEXT: Bundle Machine CFG Edges
627 ; GCN-O1-OPTS-NEXT: Spill Code Placement Analysis
628 ; GCN-O1-OPTS-NEXT: Lazy Machine Block Frequency Analysis
629 ; GCN-O1-OPTS-NEXT: Machine Optimization Remark Emitter
630 ; GCN-O1-OPTS-NEXT: Greedy Register Allocator
631 ; GCN-O1-OPTS-NEXT: Virtual Register Rewriter
632 ; GCN-O1-OPTS-NEXT: SI lower SGPR spill instructions
633 ; GCN-O1-OPTS-NEXT: Virtual Register Map
634 ; GCN-O1-OPTS-NEXT: Live Register Matrix
635 ; GCN-O1-OPTS-NEXT: Greedy Register Allocator
636 ; GCN-O1-OPTS-NEXT: GCN NSA Reassign
637 ; GCN-O1-OPTS-NEXT: Virtual Register Rewriter
638 ; GCN-O1-OPTS-NEXT: Stack Slot Coloring
639 ; GCN-O1-OPTS-NEXT: Machine Copy Propagation Pass
640 ; GCN-O1-OPTS-NEXT: Machine Loop Invariant Code Motion
641 ; GCN-O1-OPTS-NEXT: SI Fix VGPR copies
642 ; GCN-O1-OPTS-NEXT: SI optimize exec mask operations
643 ; GCN-O1-OPTS-NEXT: Remove Redundant DEBUG_VALUE analysis
644 ; GCN-O1-OPTS-NEXT: Fixup Statepoint Caller Saved
645 ; GCN-O1-OPTS-NEXT: PostRA Machine Sink
646 ; GCN-O1-OPTS-NEXT: MachineDominator Tree Construction
647 ; GCN-O1-OPTS-NEXT: Machine Natural Loop Construction
648 ; GCN-O1-OPTS-NEXT: Machine Block Frequency Analysis
649 ; GCN-O1-OPTS-NEXT: MachinePostDominator Tree Construction
650 ; GCN-O1-OPTS-NEXT: Lazy Machine Block Frequency Analysis
651 ; GCN-O1-OPTS-NEXT: Machine Optimization Remark Emitter
652 ; GCN-O1-OPTS-NEXT: Shrink Wrapping analysis
653 ; GCN-O1-OPTS-NEXT: Prologue/Epilogue Insertion & Frame Finalization
654 ; GCN-O1-OPTS-NEXT: Control Flow Optimizer
655 ; GCN-O1-OPTS-NEXT: Lazy Machine Block Frequency Analysis
656 ; GCN-O1-OPTS-NEXT: Tail Duplication
657 ; GCN-O1-OPTS-NEXT: Machine Copy Propagation Pass
658 ; GCN-O1-OPTS-NEXT: Post-RA pseudo instruction expansion pass
659 ; GCN-O1-OPTS-NEXT: SI Shrink Instructions
660 ; GCN-O1-OPTS-NEXT: SI post-RA bundler
661 ; GCN-O1-OPTS-NEXT: MachineDominator Tree Construction
662 ; GCN-O1-OPTS-NEXT: Machine Natural Loop Construction
663 ; GCN-O1-OPTS-NEXT: PostRA Machine Instruction Scheduler
664 ; GCN-O1-OPTS-NEXT: Machine Block Frequency Analysis
665 ; GCN-O1-OPTS-NEXT: MachinePostDominator Tree Construction
666 ; GCN-O1-OPTS-NEXT: Branch Probability Basic Block Placement
667 ; GCN-O1-OPTS-NEXT: Insert fentry calls
668 ; GCN-O1-OPTS-NEXT: Insert XRay ops
669 ; GCN-O1-OPTS-NEXT: GCN Create VOPD Instructions
670 ; GCN-O1-OPTS-NEXT: SI Memory Legalizer
671 ; GCN-O1-OPTS-NEXT: MachineDominator Tree Construction
672 ; GCN-O1-OPTS-NEXT: Machine Natural Loop Construction
673 ; GCN-O1-OPTS-NEXT: MachinePostDominator Tree Construction
674 ; GCN-O1-OPTS-NEXT: SI insert wait instructions
675 ; GCN-O1-OPTS-NEXT: Insert required mode register values
676 ; GCN-O1-OPTS-NEXT: SI Insert Hard Clauses
677 ; GCN-O1-OPTS-NEXT: SI Final Branch Preparation
678 ; GCN-O1-OPTS-NEXT: SI peephole optimizations
679 ; GCN-O1-OPTS-NEXT: Post RA hazard recognizer
680 ; GCN-O1-OPTS-NEXT: AMDGPU Insert Delay ALU
681 ; GCN-O1-OPTS-NEXT: Branch relaxation pass
682 ; GCN-O1-OPTS-NEXT: Register Usage Information Collector Pass
683 ; GCN-O1-OPTS-NEXT: Live DEBUG_VALUE analysis
684 ; GCN-O1-OPTS-NEXT: Function register usage analysis
685 ; GCN-O1-OPTS-NEXT: FunctionPass Manager
686 ; GCN-O1-OPTS-NEXT: Lazy Machine Block Frequency Analysis
687 ; GCN-O1-OPTS-NEXT: Machine Optimization Remark Emitter
688 ; GCN-O1-OPTS-NEXT: AMDGPU Assembly Printer
689 ; GCN-O1-OPTS-NEXT: Free MachineFunction
690 ; GCN-O1-OPTS-NEXT:Pass Arguments: -domtree
691 ; GCN-O1-OPTS-NEXT: FunctionPass Manager
692 ; GCN-O1-OPTS-NEXT: Dominator Tree Construction
694 ; GCN-O2:Target Library Information
695 ; GCN-O2-NEXT:Target Pass Configuration
696 ; GCN-O2-NEXT:Machine Module Information
697 ; GCN-O2-NEXT:Target Transform Information
698 ; GCN-O2-NEXT:Assumption Cache Tracker
699 ; GCN-O2-NEXT:Profile summary info
700 ; GCN-O2-NEXT:AMDGPU Address space based Alias Analysis
701 ; GCN-O2-NEXT:External Alias Analysis
702 ; GCN-O2-NEXT:Type-Based Alias Analysis
703 ; GCN-O2-NEXT:Scoped NoAlias Alias Analysis
704 ; GCN-O2-NEXT:Argument Register Usage Information Storage
705 ; GCN-O2-NEXT:Create Garbage Collector Module Metadata
706 ; GCN-O2-NEXT:Machine Branch Probability Analysis
707 ; GCN-O2-NEXT:Register Usage Information Storage
708 ; GCN-O2-NEXT:Default Regalloc Eviction Advisor
709 ; GCN-O2-NEXT: ModulePass Manager
710 ; GCN-O2-NEXT: Pre-ISel Intrinsic Lowering
711 ; GCN-O2-NEXT: AMDGPU Printf lowering
712 ; GCN-O2-NEXT: FunctionPass Manager
713 ; GCN-O2-NEXT: Dominator Tree Construction
714 ; GCN-O2-NEXT: Lower ctors and dtors for AMDGPU
715 ; GCN-O2-NEXT: FunctionPass Manager
716 ; GCN-O2-NEXT: Early propagate attributes from kernels to functions
717 ; GCN-O2-NEXT: AMDGPU Lower Intrinsics
718 ; GCN-O2-NEXT: AMDGPU Inline All Functions
719 ; GCN-O2-NEXT: CallGraph Construction
720 ; GCN-O2-NEXT: Call Graph SCC Pass Manager
721 ; GCN-O2-NEXT: Inliner for always_inline functions
722 ; GCN-O2-NEXT: A No-Op Barrier Pass
723 ; GCN-O2-NEXT: Lower OpenCL enqueued blocks
724 ; GCN-O2-NEXT: Lower uses of LDS variables from non-kernel functions
725 ; GCN-O2-NEXT: FunctionPass Manager
726 ; GCN-O2-NEXT: Infer address spaces
727 ; GCN-O2-NEXT: Expand Atomic instructions
728 ; GCN-O2-NEXT: AMDGPU Promote Alloca
729 ; GCN-O2-NEXT: Dominator Tree Construction
730 ; GCN-O2-NEXT: SROA
731 ; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
732 ; GCN-O2-NEXT: Function Alias Analysis Results
733 ; GCN-O2-NEXT: Memory SSA
734 ; GCN-O2-NEXT: Natural Loop Information
735 ; GCN-O2-NEXT: Canonicalize natural loops
736 ; GCN-O2-NEXT: LCSSA Verifier
737 ; GCN-O2-NEXT: Loop-Closed SSA Form Pass
738 ; GCN-O2-NEXT: Scalar Evolution Analysis
739 ; GCN-O2-NEXT: Lazy Branch Probability Analysis
740 ; GCN-O2-NEXT: Lazy Block Frequency Analysis
741 ; GCN-O2-NEXT: Loop Pass Manager
742 ; GCN-O2-NEXT: Loop Invariant Code Motion
743 ; GCN-O2-NEXT: Split GEPs to a variadic base and a constant offset for better CSE
744 ; GCN-O2-NEXT: Speculatively execute instructions
745 ; GCN-O2-NEXT: Scalar Evolution Analysis
746 ; GCN-O2-NEXT: Straight line strength reduction
747 ; GCN-O2-NEXT: Early CSE
748 ; GCN-O2-NEXT: Scalar Evolution Analysis
749 ; GCN-O2-NEXT: Nary reassociation
750 ; GCN-O2-NEXT: Early CSE
751 ; GCN-O2-NEXT: Post-Dominator Tree Construction
752 ; GCN-O2-NEXT: Legacy Divergence Analysis
753 ; GCN-O2-NEXT: AMDGPU IR optimizations
754 ; GCN-O2-NEXT: Canonicalize natural loops
755 ; GCN-O2-NEXT: Scalar Evolution Analysis
756 ; GCN-O2-NEXT: Loop Pass Manager
757 ; GCN-O2-NEXT: Canonicalize Freeze Instructions in Loops
758 ; GCN-O2-NEXT: Induction Variable Users
759 ; GCN-O2-NEXT: Loop Strength Reduction
760 ; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
761 ; GCN-O2-NEXT: Function Alias Analysis Results
762 ; GCN-O2-NEXT: Merge contiguous icmps into a memcmp
763 ; GCN-O2-NEXT: Natural Loop Information
764 ; GCN-O2-NEXT: Lazy Branch Probability Analysis
765 ; GCN-O2-NEXT: Lazy Block Frequency Analysis
766 ; GCN-O2-NEXT: Expand memcmp() to load/stores
767 ; GCN-O2-NEXT: Lower constant intrinsics
768 ; GCN-O2-NEXT: Remove unreachable blocks from the CFG
769 ; GCN-O2-NEXT: Natural Loop Information
770 ; GCN-O2-NEXT: Post-Dominator Tree Construction
771 ; GCN-O2-NEXT: Branch Probability Analysis
772 ; GCN-O2-NEXT: Block Frequency Analysis
773 ; GCN-O2-NEXT: Constant Hoisting
774 ; GCN-O2-NEXT: Replace intrinsics with calls to vector library
775 ; GCN-O2-NEXT: Partially inline calls to library functions
776 ; GCN-O2-NEXT: Expand vector predication intrinsics
777 ; GCN-O2-NEXT: Scalarize Masked Memory Intrinsics
778 ; GCN-O2-NEXT: Expand reduction intrinsics
779 ; GCN-O2-NEXT: Natural Loop Information
780 ; GCN-O2-NEXT: TLS Variable Hoist
781 ; GCN-O2-NEXT: Early CSE
782 ; GCN-O2-NEXT: AMDGPU Attributor
783 ; GCN-O2-NEXT: CallGraph Construction
784 ; GCN-O2-NEXT: Call Graph SCC Pass Manager
785 ; GCN-O2-NEXT: AMDGPU Annotate Kernel Features
786 ; GCN-O2-NEXT: FunctionPass Manager
787 ; GCN-O2-NEXT: AMDGPU Lower Kernel Arguments
788 ; GCN-O2-NEXT: Dominator Tree Construction
789 ; GCN-O2-NEXT: Natural Loop Information
790 ; GCN-O2-NEXT: CodeGen Prepare
791 ; GCN-O2-NEXT: Dominator Tree Construction
792 ; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
793 ; GCN-O2-NEXT: Function Alias Analysis Results
794 ; GCN-O2-NEXT: Natural Loop Information
795 ; GCN-O2-NEXT: Scalar Evolution Analysis
796 ; GCN-O2-NEXT: GPU Load and Store Vectorizer
797 ; GCN-O2-NEXT: Lazy Value Information Analysis
798 ; GCN-O2-NEXT: Lower SwitchInst's to branches
799 ; GCN-O2-NEXT: Lower invoke and unwind, for unwindless code generators
800 ; GCN-O2-NEXT: Remove unreachable blocks from the CFG
801 ; GCN-O2-NEXT: Dominator Tree Construction
802 ; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
803 ; GCN-O2-NEXT: Function Alias Analysis Results
804 ; GCN-O2-NEXT: Flatten the CFG
805 ; GCN-O2-NEXT: Dominator Tree Construction
806 ; GCN-O2-NEXT: Post-Dominator Tree Construction
807 ; GCN-O2-NEXT: Natural Loop Information
808 ; GCN-O2-NEXT: Legacy Divergence Analysis
809 ; GCN-O2-NEXT: AMDGPU IR late optimizations
810 ; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
811 ; GCN-O2-NEXT: Function Alias Analysis Results
812 ; GCN-O2-NEXT: Code sinking
813 ; GCN-O2-NEXT: Legacy Divergence Analysis
814 ; GCN-O2-NEXT: Unify divergent function exit nodes
815 ; GCN-O2-NEXT: Lazy Value Information Analysis
816 ; GCN-O2-NEXT: Lower SwitchInst's to branches
817 ; GCN-O2-NEXT: Dominator Tree Construction
818 ; GCN-O2-NEXT: Natural Loop Information
819 ; GCN-O2-NEXT: Convert irreducible control-flow into natural loops
820 ; GCN-O2-NEXT: Fixup each natural loop to have a single exit block
821 ; GCN-O2-NEXT: Post-Dominator Tree Construction
822 ; GCN-O2-NEXT: Dominance Frontier Construction
823 ; GCN-O2-NEXT: Detect single entry single exit regions
824 ; GCN-O2-NEXT: Region Pass Manager
825 ; GCN-O2-NEXT: Structurize control flow
826 ; GCN-O2-NEXT: Post-Dominator Tree Construction
827 ; GCN-O2-NEXT: Natural Loop Information
828 ; GCN-O2-NEXT: Legacy Divergence Analysis
829 ; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
830 ; GCN-O2-NEXT: Function Alias Analysis Results
831 ; GCN-O2-NEXT: Memory SSA
832 ; GCN-O2-NEXT: AMDGPU Annotate Uniform Values
833 ; GCN-O2-NEXT: SI annotate control flow
834 ; GCN-O2-NEXT: LCSSA Verifier
835 ; GCN-O2-NEXT: Loop-Closed SSA Form Pass
836 ; GCN-O2-NEXT: Analysis if a function is memory bound
837 ; GCN-O2-NEXT: DummyCGSCCPass
838 ; GCN-O2-NEXT: FunctionPass Manager
839 ; GCN-O2-NEXT: Safe Stack instrumentation pass
840 ; GCN-O2-NEXT: Insert stack protectors
841 ; GCN-O2-NEXT: Dominator Tree Construction
842 ; GCN-O2-NEXT: Post-Dominator Tree Construction
843 ; GCN-O2-NEXT: Natural Loop Information
844 ; GCN-O2-NEXT: Legacy Divergence Analysis
845 ; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
846 ; GCN-O2-NEXT: Function Alias Analysis Results
847 ; GCN-O2-NEXT: Branch Probability Analysis
848 ; GCN-O2-NEXT: Lazy Branch Probability Analysis
849 ; GCN-O2-NEXT: Lazy Block Frequency Analysis
850 ; GCN-O2-NEXT: AMDGPU DAG->DAG Pattern Instruction Selection
851 ; GCN-O2-NEXT: MachineDominator Tree Construction
852 ; GCN-O2-NEXT: SI Fix SGPR copies
853 ; GCN-O2-NEXT: MachinePostDominator Tree Construction
854 ; GCN-O2-NEXT: SI Lower i1 Copies
855 ; GCN-O2-NEXT: Finalize ISel and expand pseudo-instructions
856 ; GCN-O2-NEXT: Lazy Machine Block Frequency Analysis
857 ; GCN-O2-NEXT: Early Tail Duplication
858 ; GCN-O2-NEXT: Optimize machine instruction PHIs
859 ; GCN-O2-NEXT: Slot index numbering
860 ; GCN-O2-NEXT: Merge disjoint stack slots
861 ; GCN-O2-NEXT: Local Stack Slot Allocation
862 ; GCN-O2-NEXT: Remove dead machine instructions
863 ; GCN-O2-NEXT: MachineDominator Tree Construction
864 ; GCN-O2-NEXT: Machine Natural Loop Construction
865 ; GCN-O2-NEXT: Machine Block Frequency Analysis
866 ; GCN-O2-NEXT: Early Machine Loop Invariant Code Motion
867 ; GCN-O2-NEXT: MachineDominator Tree Construction
868 ; GCN-O2-NEXT: Machine Block Frequency Analysis
869 ; GCN-O2-NEXT: Machine Common Subexpression Elimination
870 ; GCN-O2-NEXT: MachinePostDominator Tree Construction
871 ; GCN-O2-NEXT: Machine Cycle Info Analysis
872 ; GCN-O2-NEXT: Machine code sinking
873 ; GCN-O2-NEXT: Peephole Optimizations
874 ; GCN-O2-NEXT: Remove dead machine instructions
875 ; GCN-O2-NEXT: SI Fold Operands
876 ; GCN-O2-NEXT: GCN DPP Combine
877 ; GCN-O2-NEXT: SI Load Store Optimizer
878 ; GCN-O2-NEXT: SI Peephole SDWA
879 ; GCN-O2-NEXT: Machine Block Frequency Analysis
880 ; GCN-O2-NEXT: MachineDominator Tree Construction
881 ; GCN-O2-NEXT: Early Machine Loop Invariant Code Motion
882 ; GCN-O2-NEXT: MachineDominator Tree Construction
883 ; GCN-O2-NEXT: Machine Block Frequency Analysis
884 ; GCN-O2-NEXT: Machine Common Subexpression Elimination
885 ; GCN-O2-NEXT: SI Fold Operands
886 ; GCN-O2-NEXT: Remove dead machine instructions
887 ; GCN-O2-NEXT: SI Shrink Instructions
888 ; GCN-O2-NEXT: Register Usage Information Propagation
889 ; GCN-O2-NEXT: Detect Dead Lanes
890 ; GCN-O2-NEXT: Remove dead machine instructions
891 ; GCN-O2-NEXT: Process Implicit Definitions
892 ; GCN-O2-NEXT: Remove unreachable machine basic blocks
893 ; GCN-O2-NEXT: Live Variable Analysis
894 ; GCN-O2-NEXT: SI Optimize VGPR LiveRange
895 ; GCN-O2-NEXT: Eliminate PHI nodes for register allocation
896 ; GCN-O2-NEXT: SI Lower control flow pseudo instructions
897 ; GCN-O2-NEXT: Two-Address instruction pass
898 ; GCN-O2-NEXT: Slot index numbering
899 ; GCN-O2-NEXT: Live Interval Analysis
900 ; GCN-O2-NEXT: Machine Natural Loop Construction
901 ; GCN-O2-NEXT: Simple Register Coalescing
902 ; GCN-O2-NEXT: Rename Disconnected Subregister Components
903 ; GCN-O2-NEXT: AMDGPU Pre-RA optimizations
904 ; GCN-O2-NEXT: Machine Instruction Scheduler
905 ; GCN-O2-NEXT: MachinePostDominator Tree Construction
906 ; GCN-O2-NEXT: SI Whole Quad Mode
907 ; GCN-O2-NEXT: Virtual Register Map
908 ; GCN-O2-NEXT: Live Register Matrix
909 ; GCN-O2-NEXT: SI Pre-allocate WWM Registers
910 ; GCN-O2-NEXT: SI optimize exec mask operations pre-RA
911 ; GCN-O2-NEXT: SI Form memory clauses
912 ; GCN-O2-NEXT: Machine Natural Loop Construction
913 ; GCN-O2-NEXT: Machine Block Frequency Analysis
914 ; GCN-O2-NEXT: Debug Variable Analysis
915 ; GCN-O2-NEXT: Live Stack Slot Analysis
916 ; GCN-O2-NEXT: Virtual Register Map
917 ; GCN-O2-NEXT: Live Register Matrix
918 ; GCN-O2-NEXT: Bundle Machine CFG Edges
919 ; GCN-O2-NEXT: Spill Code Placement Analysis
920 ; GCN-O2-NEXT: Lazy Machine Block Frequency Analysis
921 ; GCN-O2-NEXT: Machine Optimization Remark Emitter
922 ; GCN-O2-NEXT: Greedy Register Allocator
923 ; GCN-O2-NEXT: Virtual Register Rewriter
924 ; GCN-O2-NEXT: SI lower SGPR spill instructions
925 ; GCN-O2-NEXT: Virtual Register Map
926 ; GCN-O2-NEXT: Live Register Matrix
927 ; GCN-O2-NEXT: Greedy Register Allocator
928 ; GCN-O2-NEXT: GCN NSA Reassign
929 ; GCN-O2-NEXT: Virtual Register Rewriter
930 ; GCN-O2-NEXT: Stack Slot Coloring
931 ; GCN-O2-NEXT: Machine Copy Propagation Pass
932 ; GCN-O2-NEXT: Machine Loop Invariant Code Motion
933 ; GCN-O2-NEXT: SI Fix VGPR copies
934 ; GCN-O2-NEXT: SI optimize exec mask operations
935 ; GCN-O2-NEXT: Remove Redundant DEBUG_VALUE analysis
936 ; GCN-O2-NEXT: Fixup Statepoint Caller Saved
937 ; GCN-O2-NEXT: PostRA Machine Sink
938 ; GCN-O2-NEXT: MachineDominator Tree Construction
939 ; GCN-O2-NEXT: Machine Natural Loop Construction
940 ; GCN-O2-NEXT: Machine Block Frequency Analysis
941 ; GCN-O2-NEXT: MachinePostDominator Tree Construction
942 ; GCN-O2-NEXT: Lazy Machine Block Frequency Analysis
943 ; GCN-O2-NEXT: Machine Optimization Remark Emitter
944 ; GCN-O2-NEXT: Shrink Wrapping analysis
945 ; GCN-O2-NEXT: Prologue/Epilogue Insertion & Frame Finalization
946 ; GCN-O2-NEXT: Control Flow Optimizer
947 ; GCN-O2-NEXT: Lazy Machine Block Frequency Analysis
948 ; GCN-O2-NEXT: Tail Duplication
949 ; GCN-O2-NEXT: Machine Copy Propagation Pass
950 ; GCN-O2-NEXT: Post-RA pseudo instruction expansion pass
951 ; GCN-O2-NEXT: SI Shrink Instructions
952 ; GCN-O2-NEXT: SI post-RA bundler
953 ; GCN-O2-NEXT: MachineDominator Tree Construction
954 ; GCN-O2-NEXT: Machine Natural Loop Construction
955 ; GCN-O2-NEXT: PostRA Machine Instruction Scheduler
956 ; GCN-O2-NEXT: Machine Block Frequency Analysis
957 ; GCN-O2-NEXT: MachinePostDominator Tree Construction
958 ; GCN-O2-NEXT: Branch Probability Basic Block Placement
959 ; GCN-O2-NEXT: Insert fentry calls
960 ; GCN-O2-NEXT: Insert XRay ops
961 ; GCN-O2-NEXT: GCN Create VOPD Instructions
962 ; GCN-O2-NEXT: SI Memory Legalizer
963 ; GCN-O2-NEXT: MachineDominator Tree Construction
964 ; GCN-O2-NEXT: Machine Natural Loop Construction
965 ; GCN-O2-NEXT: MachinePostDominator Tree Construction
966 ; GCN-O2-NEXT: SI insert wait instructions
967 ; GCN-O2-NEXT: Insert required mode register values
968 ; GCN-O2-NEXT: SI Insert Hard Clauses
969 ; GCN-O2-NEXT: SI Final Branch Preparation
970 ; GCN-O2-NEXT: SI peephole optimizations
971 ; GCN-O2-NEXT: Post RA hazard recognizer
972 ; GCN-O2-NEXT: Release VGPRs
973 ; GCN-O2-NEXT: AMDGPU Insert Delay ALU
974 ; GCN-O2-NEXT: Branch relaxation pass
975 ; GCN-O2-NEXT: Register Usage Information Collector Pass
976 ; GCN-O2-NEXT: Live DEBUG_VALUE analysis
977 ; GCN-O2-NEXT: Function register usage analysis
978 ; GCN-O2-NEXT: FunctionPass Manager
979 ; GCN-O2-NEXT: Lazy Machine Block Frequency Analysis
980 ; GCN-O2-NEXT: Machine Optimization Remark Emitter
981 ; GCN-O2-NEXT: AMDGPU Assembly Printer
982 ; GCN-O2-NEXT: Free MachineFunction
983 ; GCN-O2-NEXT:Pass Arguments: -domtree
984 ; GCN-O2-NEXT: FunctionPass Manager
985 ; GCN-O2-NEXT: Dominator Tree Construction
987 ; GCN-O3:Target Library Information
988 ; GCN-O3-NEXT:Target Pass Configuration
989 ; GCN-O3-NEXT:Machine Module Information
990 ; GCN-O3-NEXT:Target Transform Information
991 ; GCN-O3-NEXT:Assumption Cache Tracker
992 ; GCN-O3-NEXT:Profile summary info
993 ; GCN-O3-NEXT:AMDGPU Address space based Alias Analysis
994 ; GCN-O3-NEXT:External Alias Analysis
995 ; GCN-O3-NEXT:Type-Based Alias Analysis
996 ; GCN-O3-NEXT:Scoped NoAlias Alias Analysis
997 ; GCN-O3-NEXT:Argument Register Usage Information Storage
998 ; GCN-O3-NEXT:Create Garbage Collector Module Metadata
999 ; GCN-O3-NEXT:Machine Branch Probability Analysis
1000 ; GCN-O3-NEXT:Register Usage Information Storage
1001 ; GCN-O3-NEXT:Default Regalloc Eviction Advisor
1002 ; GCN-O3-NEXT: ModulePass Manager
1003 ; GCN-O3-NEXT: Pre-ISel Intrinsic Lowering
1004 ; GCN-O3-NEXT: AMDGPU Printf lowering
1005 ; GCN-O3-NEXT: FunctionPass Manager
1006 ; GCN-O3-NEXT: Dominator Tree Construction
1007 ; GCN-O3-NEXT: Lower ctors and dtors for AMDGPU
1008 ; GCN-O3-NEXT: FunctionPass Manager
1009 ; GCN-O3-NEXT: Early propagate attributes from kernels to functions
1010 ; GCN-O3-NEXT: AMDGPU Lower Intrinsics
1011 ; GCN-O3-NEXT: AMDGPU Inline All Functions
1012 ; GCN-O3-NEXT: CallGraph Construction
1013 ; GCN-O3-NEXT: Call Graph SCC Pass Manager
1014 ; GCN-O3-NEXT: Inliner for always_inline functions
1015 ; GCN-O3-NEXT: A No-Op Barrier Pass
1016 ; GCN-O3-NEXT: Lower OpenCL enqueued blocks
1017 ; GCN-O3-NEXT: Lower uses of LDS variables from non-kernel functions
1018 ; GCN-O3-NEXT: FunctionPass Manager
1019 ; GCN-O3-NEXT: Infer address spaces
1020 ; GCN-O3-NEXT: Expand Atomic instructions
1021 ; GCN-O3-NEXT: AMDGPU Promote Alloca
1022 ; GCN-O3-NEXT: Dominator Tree Construction
1023 ; GCN-O3-NEXT: SROA
1024 ; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
1025 ; GCN-O3-NEXT: Function Alias Analysis Results
1026 ; GCN-O3-NEXT: Memory SSA
1027 ; GCN-O3-NEXT: Natural Loop Information
1028 ; GCN-O3-NEXT: Canonicalize natural loops
1029 ; GCN-O3-NEXT: LCSSA Verifier
1030 ; GCN-O3-NEXT: Loop-Closed SSA Form Pass
1031 ; GCN-O3-NEXT: Scalar Evolution Analysis
1032 ; GCN-O3-NEXT: Lazy Branch Probability Analysis
1033 ; GCN-O3-NEXT: Lazy Block Frequency Analysis
1034 ; GCN-O3-NEXT: Loop Pass Manager
1035 ; GCN-O3-NEXT: Loop Invariant Code Motion
1036 ; GCN-O3-NEXT: Split GEPs to a variadic base and a constant offset for better CSE
1037 ; GCN-O3-NEXT: Speculatively execute instructions
1038 ; GCN-O3-NEXT: Scalar Evolution Analysis
1039 ; GCN-O3-NEXT: Straight line strength reduction
1040 ; GCN-O3-NEXT: Phi Values Analysis
1041 ; GCN-O3-NEXT: Function Alias Analysis Results
1042 ; GCN-O3-NEXT: Memory Dependence Analysis
1043 ; GCN-O3-NEXT: Optimization Remark Emitter
1044 ; GCN-O3-NEXT: Global Value Numbering
1045 ; GCN-O3-NEXT: Scalar Evolution Analysis
1046 ; GCN-O3-NEXT: Nary reassociation
1047 ; GCN-O3-NEXT: Early CSE
1048 ; GCN-O3-NEXT: Post-Dominator Tree Construction
1049 ; GCN-O3-NEXT: Legacy Divergence Analysis
1050 ; GCN-O3-NEXT: AMDGPU IR optimizations
1051 ; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
1052 ; GCN-O3-NEXT: Canonicalize natural loops
1053 ; GCN-O3-NEXT: Scalar Evolution Analysis
1054 ; GCN-O3-NEXT: Loop Pass Manager
1055 ; GCN-O3-NEXT: Canonicalize Freeze Instructions in Loops
1056 ; GCN-O3-NEXT: Induction Variable Users
1057 ; GCN-O3-NEXT: Loop Strength Reduction
1058 ; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
1059 ; GCN-O3-NEXT: Function Alias Analysis Results
1060 ; GCN-O3-NEXT: Merge contiguous icmps into a memcmp
1061 ; GCN-O3-NEXT: Natural Loop Information
1062 ; GCN-O3-NEXT: Lazy Branch Probability Analysis
1063 ; GCN-O3-NEXT: Lazy Block Frequency Analysis
1064 ; GCN-O3-NEXT: Expand memcmp() to load/stores
1065 ; GCN-O3-NEXT: Lower constant intrinsics
1066 ; GCN-O3-NEXT: Remove unreachable blocks from the CFG
1067 ; GCN-O3-NEXT: Natural Loop Information
1068 ; GCN-O3-NEXT: Post-Dominator Tree Construction
1069 ; GCN-O3-NEXT: Branch Probability Analysis
1070 ; GCN-O3-NEXT: Block Frequency Analysis
1071 ; GCN-O3-NEXT: Constant Hoisting
1072 ; GCN-O3-NEXT: Replace intrinsics with calls to vector library
1073 ; GCN-O3-NEXT: Partially inline calls to library functions
1074 ; GCN-O3-NEXT: Expand vector predication intrinsics
1075 ; GCN-O3-NEXT: Scalarize Masked Memory Intrinsics
1076 ; GCN-O3-NEXT: Expand reduction intrinsics
1077 ; GCN-O3-NEXT: Natural Loop Information
1078 ; GCN-O3-NEXT: TLS Variable Hoist
1079 ; GCN-O3-NEXT: Phi Values Analysis
1080 ; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
1081 ; GCN-O3-NEXT: Function Alias Analysis Results
1082 ; GCN-O3-NEXT: Memory Dependence Analysis
1083 ; GCN-O3-NEXT: Lazy Branch Probability Analysis
1084 ; GCN-O3-NEXT: Lazy Block Frequency Analysis
1085 ; GCN-O3-NEXT: Optimization Remark Emitter
1086 ; GCN-O3-NEXT: Global Value Numbering
1087 ; GCN-O3-NEXT: AMDGPU Attributor
1088 ; GCN-O3-NEXT: CallGraph Construction
1089 ; GCN-O3-NEXT: Call Graph SCC Pass Manager
1090 ; GCN-O3-NEXT: AMDGPU Annotate Kernel Features
1091 ; GCN-O3-NEXT: FunctionPass Manager
1092 ; GCN-O3-NEXT: AMDGPU Lower Kernel Arguments
1093 ; GCN-O3-NEXT: Dominator Tree Construction
1094 ; GCN-O3-NEXT: Natural Loop Information
1095 ; GCN-O3-NEXT: CodeGen Prepare
1096 ; GCN-O3-NEXT: Dominator Tree Construction
1097 ; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
1098 ; GCN-O3-NEXT: Function Alias Analysis Results
1099 ; GCN-O3-NEXT: Natural Loop Information
1100 ; GCN-O3-NEXT: Scalar Evolution Analysis
1101 ; GCN-O3-NEXT: GPU Load and Store Vectorizer
1102 ; GCN-O3-NEXT: Lazy Value Information Analysis
1103 ; GCN-O3-NEXT: Lower SwitchInst's to branches
1104 ; GCN-O3-NEXT: Lower invoke and unwind, for unwindless code generators
1105 ; GCN-O3-NEXT: Remove unreachable blocks from the CFG
1106 ; GCN-O3-NEXT: Dominator Tree Construction
1107 ; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
1108 ; GCN-O3-NEXT: Function Alias Analysis Results
1109 ; GCN-O3-NEXT: Flatten the CFG
1110 ; GCN-O3-NEXT: Dominator Tree Construction
1111 ; GCN-O3-NEXT: Post-Dominator Tree Construction
1112 ; GCN-O3-NEXT: Natural Loop Information
1113 ; GCN-O3-NEXT: Legacy Divergence Analysis
1114 ; GCN-O3-NEXT: AMDGPU IR late optimizations
1115 ; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
1116 ; GCN-O3-NEXT: Function Alias Analysis Results
1117 ; GCN-O3-NEXT: Code sinking
1118 ; GCN-O3-NEXT: Legacy Divergence Analysis
1119 ; GCN-O3-NEXT: Unify divergent function exit nodes
1120 ; GCN-O3-NEXT: Lazy Value Information Analysis
1121 ; GCN-O3-NEXT: Lower SwitchInst's to branches
1122 ; GCN-O3-NEXT: Dominator Tree Construction
1123 ; GCN-O3-NEXT: Natural Loop Information
1124 ; GCN-O3-NEXT: Convert irreducible control-flow into natural loops
1125 ; GCN-O3-NEXT: Fixup each natural loop to have a single exit block
1126 ; GCN-O3-NEXT: Post-Dominator Tree Construction
1127 ; GCN-O3-NEXT: Dominance Frontier Construction
1128 ; GCN-O3-NEXT: Detect single entry single exit regions
1129 ; GCN-O3-NEXT: Region Pass Manager
1130 ; GCN-O3-NEXT: Structurize control flow
1131 ; GCN-O3-NEXT: Post-Dominator Tree Construction
1132 ; GCN-O3-NEXT: Natural Loop Information
1133 ; GCN-O3-NEXT: Legacy Divergence Analysis
1134 ; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
1135 ; GCN-O3-NEXT: Function Alias Analysis Results
1136 ; GCN-O3-NEXT: Memory SSA
1137 ; GCN-O3-NEXT: AMDGPU Annotate Uniform Values
1138 ; GCN-O3-NEXT: SI annotate control flow
1139 ; GCN-O3-NEXT: LCSSA Verifier
1140 ; GCN-O3-NEXT: Loop-Closed SSA Form Pass
1141 ; GCN-O3-NEXT: Analysis if a function is memory bound
1142 ; GCN-O3-NEXT: DummyCGSCCPass
1143 ; GCN-O3-NEXT: FunctionPass Manager
1144 ; GCN-O3-NEXT: Safe Stack instrumentation pass
1145 ; GCN-O3-NEXT: Insert stack protectors
1146 ; GCN-O3-NEXT: Dominator Tree Construction
1147 ; GCN-O3-NEXT: Post-Dominator Tree Construction
1148 ; GCN-O3-NEXT: Natural Loop Information
1149 ; GCN-O3-NEXT: Legacy Divergence Analysis
1150 ; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
1151 ; GCN-O3-NEXT: Function Alias Analysis Results
1152 ; GCN-O3-NEXT: Branch Probability Analysis
1153 ; GCN-O3-NEXT: Lazy Branch Probability Analysis
1154 ; GCN-O3-NEXT: Lazy Block Frequency Analysis
1155 ; GCN-O3-NEXT: AMDGPU DAG->DAG Pattern Instruction Selection
1156 ; GCN-O3-NEXT: MachineDominator Tree Construction
1157 ; GCN-O3-NEXT: SI Fix SGPR copies
1158 ; GCN-O3-NEXT: MachinePostDominator Tree Construction
1159 ; GCN-O3-NEXT: SI Lower i1 Copies
1160 ; GCN-O3-NEXT: Finalize ISel and expand pseudo-instructions
1161 ; GCN-O3-NEXT: Lazy Machine Block Frequency Analysis
1162 ; GCN-O3-NEXT: Early Tail Duplication
1163 ; GCN-O3-NEXT: Optimize machine instruction PHIs
1164 ; GCN-O3-NEXT: Slot index numbering
1165 ; GCN-O3-NEXT: Merge disjoint stack slots
1166 ; GCN-O3-NEXT: Local Stack Slot Allocation
1167 ; GCN-O3-NEXT: Remove dead machine instructions
1168 ; GCN-O3-NEXT: MachineDominator Tree Construction
1169 ; GCN-O3-NEXT: Machine Natural Loop Construction
1170 ; GCN-O3-NEXT: Machine Block Frequency Analysis
1171 ; GCN-O3-NEXT: Early Machine Loop Invariant Code Motion
1172 ; GCN-O3-NEXT: MachineDominator Tree Construction
1173 ; GCN-O3-NEXT: Machine Block Frequency Analysis
1174 ; GCN-O3-NEXT: Machine Common Subexpression Elimination
1175 ; GCN-O3-NEXT: MachinePostDominator Tree Construction
1176 ; GCN-O3-NEXT: Machine Cycle Info Analysis
1177 ; GCN-O3-NEXT: Machine code sinking
1178 ; GCN-O3-NEXT: Peephole Optimizations
1179 ; GCN-O3-NEXT: Remove dead machine instructions
1180 ; GCN-O3-NEXT: SI Fold Operands
1181 ; GCN-O3-NEXT: GCN DPP Combine
1182 ; GCN-O3-NEXT: SI Load Store Optimizer
1183 ; GCN-O3-NEXT: SI Peephole SDWA
1184 ; GCN-O3-NEXT: Machine Block Frequency Analysis
1185 ; GCN-O3-NEXT: MachineDominator Tree Construction
1186 ; GCN-O3-NEXT: Early Machine Loop Invariant Code Motion
1187 ; GCN-O3-NEXT: MachineDominator Tree Construction
1188 ; GCN-O3-NEXT: Machine Block Frequency Analysis
1189 ; GCN-O3-NEXT: Machine Common Subexpression Elimination
1190 ; GCN-O3-NEXT: SI Fold Operands
1191 ; GCN-O3-NEXT: Remove dead machine instructions
1192 ; GCN-O3-NEXT: SI Shrink Instructions
1193 ; GCN-O3-NEXT: Register Usage Information Propagation
1194 ; GCN-O3-NEXT: Detect Dead Lanes
1195 ; GCN-O3-NEXT: Remove dead machine instructions
1196 ; GCN-O3-NEXT: Process Implicit Definitions
1197 ; GCN-O3-NEXT: Remove unreachable machine basic blocks
1198 ; GCN-O3-NEXT: Live Variable Analysis
1199 ; GCN-O3-NEXT: SI Optimize VGPR LiveRange
1200 ; GCN-O3-NEXT: Eliminate PHI nodes for register allocation
1201 ; GCN-O3-NEXT: SI Lower control flow pseudo instructions
1202 ; GCN-O3-NEXT: Two-Address instruction pass
1203 ; GCN-O3-NEXT: Slot index numbering
1204 ; GCN-O3-NEXT: Live Interval Analysis
1205 ; GCN-O3-NEXT: Machine Natural Loop Construction
1206 ; GCN-O3-NEXT: Simple Register Coalescing
1207 ; GCN-O3-NEXT: Rename Disconnected Subregister Components
1208 ; GCN-O3-NEXT: AMDGPU Pre-RA optimizations
1209 ; GCN-O3-NEXT: Machine Instruction Scheduler
1210 ; GCN-O3-NEXT: MachinePostDominator Tree Construction
1211 ; GCN-O3-NEXT: SI Whole Quad Mode
1212 ; GCN-O3-NEXT: Virtual Register Map
1213 ; GCN-O3-NEXT: Live Register Matrix
1214 ; GCN-O3-NEXT: SI Pre-allocate WWM Registers
1215 ; GCN-O3-NEXT: SI optimize exec mask operations pre-RA
1216 ; GCN-O3-NEXT: SI Form memory clauses
1217 ; GCN-O3-NEXT: Machine Natural Loop Construction
1218 ; GCN-O3-NEXT: Machine Block Frequency Analysis
1219 ; GCN-O3-NEXT: Debug Variable Analysis
1220 ; GCN-O3-NEXT: Live Stack Slot Analysis
1221 ; GCN-O3-NEXT: Virtual Register Map
1222 ; GCN-O3-NEXT: Live Register Matrix
1223 ; GCN-O3-NEXT: Bundle Machine CFG Edges
1224 ; GCN-O3-NEXT: Spill Code Placement Analysis
1225 ; GCN-O3-NEXT: Lazy Machine Block Frequency Analysis
1226 ; GCN-O3-NEXT: Machine Optimization Remark Emitter
1227 ; GCN-O3-NEXT: Greedy Register Allocator
1228 ; GCN-O3-NEXT: Virtual Register Rewriter
1229 ; GCN-O3-NEXT: SI lower SGPR spill instructions
1230 ; GCN-O3-NEXT: Virtual Register Map
1231 ; GCN-O3-NEXT: Live Register Matrix
1232 ; GCN-O3-NEXT: Greedy Register Allocator
1233 ; GCN-O3-NEXT: GCN NSA Reassign
1234 ; GCN-O3-NEXT: Virtual Register Rewriter
1235 ; GCN-O3-NEXT: Stack Slot Coloring
1236 ; GCN-O3-NEXT: Machine Copy Propagation Pass
1237 ; GCN-O3-NEXT: Machine Loop Invariant Code Motion
1238 ; GCN-O3-NEXT: SI Fix VGPR copies
1239 ; GCN-O3-NEXT: SI optimize exec mask operations
1240 ; GCN-O3-NEXT: Remove Redundant DEBUG_VALUE analysis
1241 ; GCN-O3-NEXT: Fixup Statepoint Caller Saved
1242 ; GCN-O3-NEXT: PostRA Machine Sink
1243 ; GCN-O3-NEXT: MachineDominator Tree Construction
1244 ; GCN-O3-NEXT: Machine Natural Loop Construction
1245 ; GCN-O3-NEXT: Machine Block Frequency Analysis
1246 ; GCN-O3-NEXT: MachinePostDominator Tree Construction
1247 ; GCN-O3-NEXT: Lazy Machine Block Frequency Analysis
1248 ; GCN-O3-NEXT: Machine Optimization Remark Emitter
1249 ; GCN-O3-NEXT: Shrink Wrapping analysis
1250 ; GCN-O3-NEXT: Prologue/Epilogue Insertion & Frame Finalization
1251 ; GCN-O3-NEXT: Control Flow Optimizer
1252 ; GCN-O3-NEXT: Lazy Machine Block Frequency Analysis
1253 ; GCN-O3-NEXT: Tail Duplication
1254 ; GCN-O3-NEXT: Machine Copy Propagation Pass
1255 ; GCN-O3-NEXT: Post-RA pseudo instruction expansion pass
1256 ; GCN-O3-NEXT: SI Shrink Instructions
1257 ; GCN-O3-NEXT: SI post-RA bundler
1258 ; GCN-O3-NEXT: MachineDominator Tree Construction
1259 ; GCN-O3-NEXT: Machine Natural Loop Construction
1260 ; GCN-O3-NEXT: PostRA Machine Instruction Scheduler
1261 ; GCN-O3-NEXT: Machine Block Frequency Analysis
1262 ; GCN-O3-NEXT: MachinePostDominator Tree Construction
1263 ; GCN-O3-NEXT: Branch Probability Basic Block Placement
1264 ; GCN-O3-NEXT: Insert fentry calls
1265 ; GCN-O3-NEXT: Insert XRay ops
1266 ; GCN-O3-NEXT: GCN Create VOPD Instructions
1267 ; GCN-O3-NEXT: SI Memory Legalizer
1268 ; GCN-O3-NEXT: MachineDominator Tree Construction
1269 ; GCN-O3-NEXT: Machine Natural Loop Construction
1270 ; GCN-O3-NEXT: MachinePostDominator Tree Construction
1271 ; GCN-O3-NEXT: SI insert wait instructions
1272 ; GCN-O3-NEXT: Insert required mode register values
1273 ; GCN-O3-NEXT: SI Insert Hard Clauses
1274 ; GCN-O3-NEXT: SI Final Branch Preparation
1275 ; GCN-O3-NEXT: SI peephole optimizations
1276 ; GCN-O3-NEXT: Post RA hazard recognizer
1277 ; GCN-O3-NEXT: Release VGPRs
1278 ; GCN-O3-NEXT: AMDGPU Insert Delay ALU
1279 ; GCN-O3-NEXT: Branch relaxation pass
1280 ; GCN-O3-NEXT: Register Usage Information Collector Pass
1281 ; GCN-O3-NEXT: Live DEBUG_VALUE analysis
1282 ; GCN-O3-NEXT: Function register usage analysis
1283 ; GCN-O3-NEXT: FunctionPass Manager
1284 ; GCN-O3-NEXT: Lazy Machine Block Frequency Analysis
1285 ; GCN-O3-NEXT: Machine Optimization Remark Emitter
1286 ; GCN-O3-NEXT: AMDGPU Assembly Printer
1287 ; GCN-O3-NEXT: Free MachineFunction
1288 ; GCN-O3-NEXT:Pass Arguments: -domtree
1289 ; GCN-O3-NEXT: FunctionPass Manager
1290 ; GCN-O3-NEXT: Dominator Tree Construction