Update LLVM baseline to 992d5d0c

[zyw-bot]

LLVM baseline is updated from llvm/llvm-project@d1a52e4 to llvm/llvm-project@992d5d0.

Commits in this update:

32fe59e6a1ae TailDuplicator: Use getDefBlock (#220903)
73d2d31b9990 [AArch64] Correct FDIV/FSQRT throughput for Neoverse V3ae. (#219113)
191b4f522082 X86: Handle undef push operands in call frame optimization (#220368)
8038da43de43 [AMDGPU] MC support for indexed VBUFFER resources for gfx13 (#218649)
c6f9cf2bf5a3 [AArch64] Add Neoverse V3AE scheduling info for v264 frecpe (#219115)
ed6cedcd145c [AArch64][GlobalISel] Create uzp in shuffle(v, undefined) situations (#220535)
542e90edacd7 IR: Intersect and drop the addrspacecast nonnull flag (#217904)
ae0347b0275b [AArch64][PAC] Fix the list of registers clobbered by PAUTH_EPILOGUE (#220190)
1830dcfa42c0 IR: Add nonnull flag to addrspacecast (#217903)
9b4c98114d4a [IPSCCP] Do not delete non-replaceable constants during globals cleanup (#220854)
382165766eaa [ARM] Name caller and callee in unsupported hard-float calling convention error message (#217693)
d27381ed50f8 [TTI] Cache type legalization costs (NFC). (#220701)
35259509c54b [OpenMPOpt] Preserve debug locations when emitting barriers. (#220554)
d9b65e6e580f [LLVM][CodeGen][SVE] Improve code quality of predicate vector extracts. (#220254)
edd28877bc19 [VPlan] Simplify (x && y) | !x -> !x || y (#220801)
778919df9ee1 Remove unused MachineBlockFrequencyInfo::isIrrLoopHeader. NFC (#220841)
0d22d4cc858e [AMDGPU] Price the packed form of a vector of i1 (#217327)
833ee6a5c5be [AArch64][Vectorizer] Enable scalable factor-3 interleaving (#217370)
68d0ae1c0de8 [SCEV] Express SeqMinMax de-dup visitor tersely (NFC) (#220539)

Changes in statistics

No significant changes.

Diff report

Scope	Files	Added Lines	Removed Lines
Original	1	9	11
Kept	1	9	11

Download the full diff list: z_fulldiff.csv

Delta	Added Lines	Removed Lines	File
-2	9	11	llvm/PPCISelLowering

Artifacts

Selected raw llvm-relaxed-diff outputs (1 file pairs, up to 100 MB) can be downloaded from https://github.com/dtcxzyw/llvm-opt-benchmark-nightly/actions/runs/33759529330.

Grouped changes

report: add < sub