MODULARIZE(1)

👑 MODULARIZE(1) 👑

okay bestie, welcome to the User Commands manual page for modularize — yeah, version 15. and it's giving iconic but also a LOT. like, there are SO many flags on this thing. but we're going through every single one together, no skips, no cap. 💅

🌸 NAME 🌸

modularize — manual page for modularize 15

✨ DESCRIPTION ✨

OVERVIEW: modularize.

USAGE:

modularize [options] <list of one or more header list files> <arguments to be passed to front end>...

so you hand it one or more header list files, plus whatever arguments you want forwarded along to the front end. that's the whole vibe. 🦋

🎀 OPTIONS 🎀

here comes the parade, bestie. every flag, every value, nothing skipped, periodt. 👑

🌈 Color Options

• --color — throws colors into your output (default=autodetect). yes, it autodetects by default. it's giving sparkle. 💖

🧚 General options

• -I <path> — an include path for the coverage check.
• --aarch64-neon-syntax=<value> — picks which flavor of NEON code the AArch64 backend spits out:
  • =generic — emit generic NEON assembly
  • =apple — emit Apple-style NEON assembly
• --aarch64-use-aa — turns on using AA during codegen.
• --abort-on-max-devirt-iterations-reached — bails out the moment the max iterations for the devirtualization CGSCC repeat pass get hit.
• --addrsig — emits an address-significance table.
• --align-loops=<uint> — the default alignment for loops.
• --allow-ginsert-as-artifact — lets G_INSERT count as an artifact. it's a hack to dodge AMDGPU test infinite loops.
• --amdgpu-bypass-slow-div — skips the 64-bit divide for dynamic 32-bit values.
• --amdgpu-disable-loop-alignment — don't align or prefetch loops, thanks.
• --amdgpu-dpp-combine — turns on the DPP combiner.
• --amdgpu-dump-hsa-metadata — dumps the AMDGPU HSA Metadata.
• --amdgpu-enable-merge-m0 — merges and hoists M0 initializations.
• --amdgpu-enable-power-sched — schedules stuff so mAI power bursts stay minimal.
• --amdgpu-igrouplp — turns on building Instruction Groups and ordering them for scheduling.
• --amdgpu-promote-alloca-to-vector-limit=<uint> — the max byte size to even consider promoting alloca to vector.
• --amdgpu-sdwa-peephole — turns on the SDWA peepholer.
• --amdgpu-use-aa-in-codegen — turns on using AA during codegen.
• --amdgpu-verify-hsa-metadata — verifies the AMDGPU HSA Metadata.
• --amdgpu-vgpr-index-mode — uses GPR indexing mode instead of movrel for vector indexing.
• --arm-add-build-attributes — no description in the original, it just exists. 💀
• --arm-implicit-it=<value> — lets conditional instructions live outside of an IT block:
  • =always — accept in both ISAs, and emit implicit ITs in Thumb
  • =never — warn in ARM, reject in Thumb
  • =arm — accept in ARM, reject in Thumb
  • =thumb — warn in ARM, emit implicit ITs in Thumb
• --asm-show-inst — emits the internal instruction representation into the assembly file.
• --atomic-counter-update-promoted — does the counter update with atomic fetch add, but only for promoted counters.
• --atomic-first-counter — uses atomic fetch add for the first counter in a function (which is usually the entry counter).
• --basic-block-sections=<all | <function list (file)> | labels | none> — emits basic blocks into separate sections.
• --block-check-header-list-only — only warns you if #include directives sit inside extern or namespace blocks and the included header is in the header list.
• --bounds-checking-single-trap — uses one trap block per function.
• --cfg-hide-cold-paths=<number> — hides blocks whose relative frequency is below the value you gave.
• --cfg-hide-deoptimize-paths — no description in the original.
• --cfg-hide-unreachable-paths — no description in the original.
• --code-model=<value> — picks the code model:
  • =tiny — tiny code model
  • =small — small code model
  • =kernel — kernel code model
  • =medium — medium code model
  • =large — large code model
• --cost-kind=<value> — the target cost kind:
  • =throughput — reciprocal throughput
  • =latency — instruction latency
  • =code-size — code size
  • =size-latency — code size and latency
• --coverage-check-only — only does the coverage check. nothing else.
• --data-sections — emits data into separate sections.
• --debug-entry-values — turns on debug info for the debug entry values.
• --debug-info-correlate — uses debug info to correlate profiles.
• --debugger-tune=<value> — tunes the debug info for a specific debugger:
  • =gdb — gdb
  • =lldb — lldb
  • =dbx — dbx
  • =sce — SCE targets (e.g. PS4)
• --debugify-func-limit=<ulong> — sets the max number of processed functions per pass.
• --debugify-level=<value> — the kind of debug info to add:
  • =locations — locations only
  • =location+variables — locations and variables
• --debugify-quiet — mutes the verbose debugify output.
• --denormal-fp-math=<value> — selects which denormal numbers the code is allowed to require:
  • =ieee — IEEE 754 denormal numbers
  • =preserve-sign — the sign of a flushed-to-zero number is preserved in the sign of 0
  • =positive-zero — denormals are flushed to positive zero
• --denormal-fp-math-f32=<value> — same deal, but for float. selects which denormal numbers the code is allowed to require:
  • =ieee — IEEE 754 denormal numbers
  • =preserve-sign — the sign of a flushed-to-zero number is preserved in the sign of 0
  • =positive-zero — denormals are flushed to positive zero
• --disable-i2p-p2i-opt — kills the inttoptr/ptrtoint roundtrip optimization.
• --disable-promote-alloca-to-lds — disables promoting alloca to LDS.
• --disable-promote-alloca-to-vector — disables promoting alloca to vector.
• --disable-tail-calls — never emits tail calls. ever.
• --display-file-lists — shows you lists of good files (no compile errors), problem files, and a combined list where problem files come with a '#' in front.
• --do-counter-promotion — does counter register promotion.
• --dot-cfg-mssa=<file name for generated dot file> — the file name for the generated dot file.
• --dwarf-version=<int> — the Dwarf version.
• --dwarf64 — generates debugging info in the 64-bit DWARF format.
• --emit-call-site-info — emits call site debug information, as long as debug info is enabled.
• --emit-dwarf-unwind=<value> — controls whether DWARF EH frame entries get emitted:
  • =always — always emit EH frame entries
  • =no-compact-unwind — only emit EH frame entries when compact unwind isn't available
  • =default — use the target platform default
• --emscripten-cxx-exceptions-allowed=<string> — the list of function names where Emscripten-style exception handling is turned on (peep the emscripten EMSCRIPTEN_CATCHING_ALLOWED options).
• --emulated-tls — uses the emulated TLS model.
• --enable-approx-func-fp-math — turns on FP math optimizations that assume approx func.
• --enable-cse-in-irtranslator — should turn on CSE in irtranslator.
• --enable-cse-in-legalizer — should turn on CSE in Legalizer.
• --enable-emscripten-cxx-exceptions — WebAssembly Emscripten-style exception handling.
• --enable-emscripten-sjlj — WebAssembly Emscripten-style setjmp/longjmp handling.
• --enable-gvn-hoist — turns on the GVN hoisting pass (default = off).
• --enable-gvn-memdep — no description in the original.
• --enable-gvn-sink — turns on the GVN sinking pass (default = off).
• --enable-jmc-instrument — instruments functions with a call to __CheckForDebuggerJustMyCode.
• --enable-load-in-loop-pre — no description in the original.
• --enable-load-pre — no description in the original.
• --enable-loop-simplifycfg-term-folding — no description in the original.
• --enable-name-compression — turns on name/filename string compression.
• --enable-no-infs-fp-math — turns on FP math optimizations that assume no +-Infs.
• --enable-no-nans-fp-math — turns on FP math optimizations that assume no NaNs.

---

and that's only part 1, bestie — the flags keep coming, so don't go anywhere. 🌈

• --enable-no-signed-zeros-fp-math — bestie, this turns on FP math optimizations that just assume the sign of 0 doesn't matter 💅 like okay sure
• --enable-no-trapping-fp-math — enables setting the FP exceptions build attribute to not use exceptions 🫶
• --enable-split-backedge-in-load-pre — no description given, it's just vibing there, mysterious ✨
• --enable-unsafe-fp-math — enables optimizations that may decrease FP precision 😭 so like, fast but slightly unhinged, know what you're signing up for
• --exception-model=<value> — the exception model, pick your fighter 🥊
  • =default — the default exception handling model 💖
  • =dwarf — DWARF-like CFI based exception handling 🧚
  • =sjlj — SjLj exception handling
  • =arm — ARM EHABI exceptions
  • =wineh — the Windows exception model 🪟
  • =wasm — WebAssembly exception handling 🌐
• --execute-concurrency=<uint> — how many threads get used to process all the files in parallel 🏃‍♀️💨 set it to 0 for hardware concurrency. Heads up though, this flag only applies to all-TUs (iykyk)
• --executor=<string> — the name of the executor to use, that's it, that's the whole thing 👑
• --experimental-debug-variable-locations — use the experimental new value-tracking variable locations 🧪 lowkey mad scientist energy
• --fatal-warnings — treat warnings as errors 😤 no chill, no mercy
• --filetype=<value> — choose a file type. not all types are supported by all targets, so don't come for me when yours isn't there 🙈
  • =asm — emit an assembly (.s) file
  • =obj — emit a native object (.o) file
  • =null — emit nothing, for performance testing 🕳️
• --filter=<string> — only process files that match this filter 🎯 again, this flag only applies to all-TUs
• --float-abi=<value> — choose your float ABI type 💐
  • =default — target default float ABI type
  • =soft — soft float ABI (implied by -soft-float)
  • =hard — hard float ABI (uses FP registers) 🔥
• --force-dwarf-frame-section — always emit a debug frame section, periodt 💅
• --fp-contract=<value> — enables aggressive formation of fused FP ops ✨
  • =fast — fuse FP ops whenever it's profitable 💰
  • =on — only fuse the 'blessed' FP ops 🙏
  • =off — only fuse FP ops when the result won't be affected 🧘
• --frame-pointer=<value> — specifies the frame pointer elimination optimization 🦋
  • =all — disable frame pointer elimination
  • =non-leaf — disable frame pointer elimination for non-leaf frame 🍃
  • =none — enable frame pointer elimination
• --fs-profile-debug-bw-threshold=<uint> — only show a debug message if the source branch weight is greater than this value 📊
• --fs-profile-debug-prob-diff-threshold=<uint> — only show a debug message if the branch probility is greater than this value (in percentage) 📈 (yes it says "probility", we don't edit the ancestors' typo, we preserve the lore)
• --function-sections — emit functions into separate sections ✂️ main character gets her own room
• --generate-merged-base-profiles — when generating nested context-sensitive profiles, always generate an extra base profile for a function with all its context profiles merged into it 🫂
• --gpsize=<uint> — Global Pointer Addressing Size. the default size is 8 🌈
• --hash-based-counter-split — rename the counter variable of a comdat function based on cfg hash #️⃣
• --hot-cold-split — enable the hot-cold splitting pass 🔥❄️
• --ignore-xcoff-visibility — don't emit the visibility attribute for asm in AIX OS, or give all symbols 'unspecified' visibility in an XCOFF object file 🙈
• --import-all-index — import all external functions in index 📥
• --incremental-linker-compatible — when used with filetype=obj, emit an object file that can be used with an incremental linker 🔗
• --instcombine-code-sinking — enable code sinking 🛁
• --instcombine-guard-widening-window=<uint> — how wide an instruction window to bypass looking for another guard 🪟
• --instcombine-max-iterations=<uint> — limit the max number of instruction combining iterations 🔁
• --instcombine-max-num-phis=<uint> — max number phis to handle in intptr/ptrint folding 🌀
• --instcombine-max-sink-users=<uint> — max number of undroppable users for instruction sinking ⚓
• --instcombine-maxarray-size=<uint> — max array size considered when doing a combine 📏
• --instcombine-negator-enabled — should we attempt to sink negations? 🤔 big question honestly
• --instcombine-negator-max-depth=<uint> — what's the maximal lookup depth when trying to check for viability of negation sinking 🕳️
• --instrprof-atomic-counter-update-all — make all profile counter updates atomic (for testing only) 🧪
• --internalize-public-api-file=<filename> — a file containing a list of symbol names to preserve 📄
• --internalize-public-api-list=<list> — a list of symbol names to preserve 📝
• --iterative-counter-promotion — allow counter promotion across the whole loop nest 🔄
• --lower-global-dtors-via-cxa-atexit — lower llvm.global_dtors (global destructors) via __cxa_atexit 💀
• --lto-embed-bitcode=<value> — embed LLVM bitcode in object files produced by LTO 📦
  • =none — do not embed
  • =optimized — embed after all optimization passes ✨
  • =post-merge-pre-opt — embed post merge, but before optimizations
• --lto-pass-remarks-filter=<regex> — only record optimization remarks from passes whose names match the given regular expression 🔍
• --lto-pass-remarks-format=<format> — the format used for serializing remarks (default: YAML) 📃
• --lto-pass-remarks-output=<filename> — output filename for pass remarks 💾
• --march=<string> — architecture to generate code for (see --version) 🏛️
• --matrix-default-layout=<value> — sets the default matrix layout 🧮
  • =column-major — use column-major layout
  • =row-major — use row-major layout
• --mattr=<a1,+a2,-a3,...> — target specific attributes (-mattr=help for details) 🎛️
• --max-counter-promotions=<int> — max number of allowed counter promotions
• --max-counter-promotions-per-loop=<uint> — max number counter promotions per loop, to avoid increasing register pressure too much 😮‍💨
• --mc-relax-all — when used with filetype=obj, relax all fixups in the emitted object file 🛋️
• --mcabac — tbd 🫥 (the devs really said "we'll figure it out later" and honestly? iconic)
• --mcpu=<cpu-name> — target a specific cpu type (-mcpu=help for details) 🖥️
• --meabi=<value> — set EABI type (default depends on triple) 🧬
  • =default — triple default EABI version
  • =4 — EABI version 4
  • =5 — EABI version 5
  • =gnu — EABI GNU
• --merror-missing-parenthesis — error for missing parenthesis around predicate registers 🚨
• --merror-noncontigious-register — error for register names that aren't contigious ⚠️ (again, keeping the ancestral spelling, it's part of the family history)
• --mhvx — enable Hexagon Vector eXtensions 🐝
• --mhvx=<value> — enable Hexagon Vector eXtensions 🐝
  • =v60 — build for HVX v60
  • =v62 — build for HVX v62
  • =v65 — build for HVX v65
  • =v66 — build for HVX v66
  • =v67 — build for HVX v67
  • =v68 — build for HVX v68
  • =v69 — build for HVX v69 ⭐
• --mips-compact-branches=<value> — MIPS Specific: Compact branch policy 🦖
  • =never — do not use compact branches if possible
  • =optimal — use compact branches where appropriate (default) 👑
  • =always — always use compact branches if possible
• --mips16-constant-islands — enable mips16 constant islands 🏝️
• --mips16-hard-float — enable mips16 hard float 🔥
• --mir-strip-debugify-only — should mir-strip-debug only strip debug info from debugified modules by default 🤔
• --misexpect-tolerance=<uint> — prevents emiting diagnostics when profile counts are within N% of the threshold.. 📉 (two periods, we respect the punctuation)
• --mno-compound — disable looking for compound instructions for Hexagon 🚫
• --mno-fixup — disable fixing up resolved relocations for Hexagon 🔧
• --mno-ldc1-sdc1 — expand double precision loads and stores to their single precision counterparts 🔬
• --mno-pairing — disable looking for duplex instructions for Hexagon 👯
• --module-map-path=<string> — turns module map output ON and tells it where to write it (path or filename, bestie, your call). If you don't give a path and you did pass --prefix, then the prefix gets used as the file path. Slay.
• --mwarn-missing-parenthesis — warns when predicate registers are missing their parenthesis 🫶
• --mwarn-noncontigious-register — warns about register names that arent contigious (yes, that's the spelling in the flag, do NOT come for me 😭)
• --mwarn-sign-mismatch — warns when a signed value and an unsigned value are mismatching 💀
• --no-coverage-check — literally just don't do the coverage check. periodt.
• --no-deprecated-warn — suppress all deprecated warnings, every single one, silence ✨
• --no-discriminators — disables generation of discriminator information.
• --no-type-check — suppresses type errors (Wasm) 🙈
• --no-warn — suppress all warnings. all of them. nothing gets through. iconic.
• --no-xray-index — don't emit the xray_fn_idx section
• --nozero-initialized-in-bss — don't place zero-initialized symbols into the bss section
• --nvptx-sched4reg — NVPTX Specific: schedule for register pressue (again, that's the original wording, typo and all 😌)
• --opaque-pointers — use opaque pointers
• --poison-checking-function-local — check that returns are non-poison (for testing)
• --prefix=<string> — prepend header file paths with this prefix. If you don't specify it, the files are considered to be relative to the header list file. ngl this one matters
• --print-pipeline-passes — prints a -passes compatible string describing the pipeline (best-effort only, so don't be shocked if it's not perfect 🫠)
• --problem-files-list=<string> — list of files with compilation or modularization problems for assistant mode. This will be excluded.
• --r600-ir-structurize — use the StructurizeCFG IR pass
• --rdf-dump — no description given, it's just out here existing, mysterious main character energy 🦋
• --rdf-limit=<uint> — also no description. iconic silence. 💅
• --relax-elf-relocations — emit GOTPCRELX/REX_GOTPCRELX instead of GOTPCREL on x86-64 ELF
• --relocation-model=<value> — choose relocation model 👑
  • =static — non-relocatable code
  • =pic — fully relocatable, position independent code
  • =dynamic-no-pic — relocatable external references, non-relocatable code
  • =ropi — code and read-only data relocatable, accessed PC-relative
  • =rwpi — read-write data relocatable, accessed relative to static base
  • =ropi-rwpi — combination of ropi and rwpi
• --root-module=<string> — specify the name of the root module.
• --runtime-counter-relocation — enable relocating counters at runtime.
• --safepoint-ir-verifier-print-only — no description, just vibes ✨
• --sample-profile-check-record-coverage=<N> — emit a warning if less than N% of records in the input profile are matched to the IR.
• --sample-profile-check-sample-coverage=<N> — emit a warning if less than N% of samples in the input profile are matched to the IR.
• --sample-profile-max-propagate-iterations=<uint> — maximum number of iterations to go through when propagating sample block/edge weights through the CFG.
• --skip-ret-exit-block — suppress counter promotion if exit blocks contain ret.
• --speculative-counter-promotion-max-exiting=<uint> — the max number of exiting blocks of a loop to allow speculative counter promotion
• --speculative-counter-promotion-to-loop — when the option is false, if the target block is in a loop, the promotion will be disallowed unless the promoted counter update can be further/iteratively promoted into an acyclic region. (read it twice, it's fine, I believe in you 🫶)
• --split-machine-functions — split out cold basic blocks from machine functions based on profile information
• --stack-size-section — emit a section containing stack size metadata
• --stack-symbol-ordering — order local stack symbols.
• --stackrealign — force align the stack to the minimum alignment
• --strict-dwarf — use strict dwarf
• --summary-file=<string> — the summary file to use for function importing.
• --sve-tail-folding=<string> — control the use of vectorisation using tail-folding for SVE: 🌀
  • disabled — no loop types will vectorize using tail-folding
  • default — uses the default tail-folding settings for the target CPU
  • all — all legal loop types will vectorize using tail-folding
  • simple — use tail-folding for simple loops (not reductions or recurrences)
  • reductions — use tail-folding for loops containing reductions
  • recurrences — use tail-folding for loops containing first order recurrences
• --swift-async-fp=<value> — determine when the Swift async frame pointer should be set
  • =auto — determine based on deployment target
  • =always — always set the bit
  • =never — never set the bit
• --tail-predication=<value> — MVE tail-predication pass options
  • =disabled — don't tail-predicate loops
  • =enabled-no-reductions — enable tail-predication, but not for reduction loops
  • =enabled — enable tail-predication, including reduction loops
  • =force-enabled-no-reductions — enable tail-predication, but not for reduction loops, and force this which might be unsafe 😭
  • =force-enabled — enable tail-predication, including reduction loops, and force this which might be unsafe 😭

bestie. the two force-enabled variants are literally described as "might be unsafe" in the source. that's not me being dramatic, that's a direct quote. proceed with caution fr fr 💀

• --tailcallopt — turn fastcc calls into tail calls by (potentially) changing ABI.
• --thinlto-assume-merged — assume the input has already undergone ThinLTO function importing and the other pre-optimization pipeline changes.
• --thread-model=<value> — choose threading model
  • =posix — POSIX thread model
  • =single — single thread model
• --threads=<int> — no description provided, she's just there 💅
• --tls-size=<uint> — bit size of immediate TLS offsets
• --type-based-intrinsic-cost — calculate intrinsics cost based only on argument types
• --unique-basic-block-section-names — give unique names to every basic block section
• --unique-section-names — give unique names to every section
• --use-ctors — use .ctors instead of .init_array.
• --vec-extabi — enable the AIX Extended Altivec ABI.
• --verify-region-info — verify region info (time consuming ⏳ so like, maybe not on a deadline)
• --vp-counters-per-site=<number> — the average number of profile counters allocated per value profiling site.
• --vp-static-alloc — do static counter allocation for value profiler
• --wasm-enable-eh — WebAssembly exception handling
• --wasm-enable-sjlj — WebAssembly setjmp/longjmp handling
• --x86-align-branch=<string> — specify types of branches to align (plus separated list of types):
  • jcc — indicates conditional jumps
  • fused — indicates fused conditional jumps
  • jmp — indicates direct unconditional jumps
  • call — indicates direct and indirect calls
  • ret — indicates rets
  • indirect — indicates indirect unconditional jumps
• --x86-align-branch-boundary=<uint> — control how the assembler should align branches with NOP. If the boundary's size is not 0, it should be a power of 2 and no less than 32. Branches will be aligned to prevent from being across or against the boundary of specified size. The default value 0 does not align branches. (so 0 = no alignment, got it 🎀)
• --x86-branches-within-32B-boundaries — align selected instructions to mitigate negative performance impact of Intel's micro code update for errata skx102.

heads up bestie: this one may break assumptions about labels corresponding to particular instructions, and should be used with caution. not a vibe to yolo 🙈

• --x86-pad-max-prefix-size=<uint> — maximum number of prefixes to use for padding
• --xcoff-traceback-table — emit the XCOFF traceback table

---

✨ Generic Options ✨

these are the OGs, the universal besties, they work everywhere 🫶

• --help — display available options (--help-hidden for more)
• --help-list — display list of available options (--help-list-hidden for more)
• --version — display the version of this program

---

✨ Polly Options ✨ — Configure the polly loop optimizer

and the Polly lore continues in the next part, don't go anywhere 🦋

• --polly — turns the polly optimizer ON, bestie ✨ (and yes it only does its thing alongside -O1, -O2 or -O3, so don't forget those)
• --polly-2nd-level-tiling — enables a 2nd level loop of loop tiling 🧱 like tiling but make it double-stacked
• --polly-ast-print-accesses — prints the memory access functions, so you can see exactly what's being touched 👀
• --polly-context=<isl parameter set> — hands you extra constraints to slap onto the context parameters
• --polly-dce-precise-steps=<int> — the number of precise steps between two approximating iterations. A value of -1 schedules another approximation stage before the actual dead code elimination 💀
• --polly-delicm-max-ops=<int> — max number of isl operations to invest in lifetime analysis; 0 means no limit, go feral 🌀
• --polly-detect-full-functions — allows the detection of full functions 🫶
• --polly-dump-after — dumps the module after Polly transformations into a file suffixed with -after
• --polly-dump-after-file=<string> — same energy, but dumps the module after Polly transformations to the given file 📁
• --polly-dump-before — dumps the module before Polly transformations into a file suffixed with -before
• --polly-dump-before-file=<string> — dumps the module before Polly transformations to the given file 📁
• --polly-enable-simplify — simplifies the SCoP after optimizations 🧼
• --polly-ignore-func=<string> — ignores functions that match a regex. Multiple regexes can be comma separated. Scop detection will ignore all functions that match ANY of the regexes provided. so like, if it matches even one, it's getting ghosted 👻
• --polly-isl-arg=<argument> — an option passed to ISL
• --polly-on-isl-error-abort — aborts if an isl error is encountered 🚨 like no recovering, we're out
• --polly-only-func=<string> — only runs on functions that match a regex. Multiple regexes can be comma separated. Scop detection will run on all functions that match ANY of the regexes provided — so match any one and you're in the club 💅
• --polly-only-region=<identifier> — only runs on certain regions (the provided identifier must appear in the name of the region's entry block
• --polly-only-scop-detection — only runs scop detection, but no other optimizations. detection only, no notes 📝
• --polly-optimized-scops — Polly — dumps the polyhedral description of Scops optimized with the isl scheduling optimizer, and the set of post-scheduling transformations is applied on the schedule tree 🌳
• --polly-parallel — generates thread parallel code (isl codegen only) 🧵
• --polly-parallel-force — forces generation of thread parallel code, ignoring any cost model. no thoughts, just threads 😤
• --polly-pattern-matching-based-opts — performs optimizations based on pattern matching 🔍
• --polly-postopts — applies post-rescheduling optimizations such as tiling (requires -polly-reschedule)
• --polly-pragma-based-opts — applies user-directed transformation from metadata
• --polly-pragma-ignore-depcheck — skips the dependency check for pragma-based transformations 🙈 (the safety net? we removed it)
• --polly-process-unprofitable — processes scops that are unlikely to benefit from Polly optimizations. giving main character energy to the underdogs 🥺
• --polly-register-tiling — enables register tiling 🎀
• --polly-report — prints information about the activities of Polly 📊 very "what did you do today" of it
• --polly-reschedule — optimizes SCoPs using ISL
• --polly-show — highlights the code regions that will be optimized in a (CFG BBs and LLVM-IR instructions) 🌈
• --polly-show-only — highlights the code regions that will be optimized in a (CFG only BBs)
• --polly-stmt-granularity=<value> — algorithm to use for splitting basic blocks into multiple statements
  • =bb — one statement per basic block
  • =scalar-indep — scalar independence heuristic
  • =store — store-level granularity
• --polly-target=<value> — the hardware to target 🎯
  • =cpu — generate CPU code
• --polly-tiling — enables loop tiling 🧱
• --polly-vectorizer=<value> — selects the vectorization strategy
  • =none — no Vectorization
  • =polly — Polly internal vectorizer
  • =stripmine — strip-mine outer loops for the loop-vectorizer to trigger

---

and that's a wrap on the whole thing, bestie!! 💖 this page is from modularize 15, dated July 2023, and it's literally the MODULARIZE(1) manual page — so if you ever forget what any of these flags do, it's all right here waiting for you. slay, go optimize something gorgeous 🌸🦋👑