[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

aeubanks abandoned this revision. aeubanks added a comment. reworked into https://github.com/llvm/llvm-project/pull/69030 Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D149800/new/ https://reviews.llvm.org/D149800 __

[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

davidxl added a comment. Sorry I missed the question. What I meant is that a lot of the profile guided size optimization (PGSO) done by Hiroshi can be adapted to use the attribute based method proposed in the patch as we don't need two different mechanisms. Repository: rG LLVM Github Monorep

[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

yamauchi added a comment. In D149800#4330831 , @davidxl wrote: > This patch uses a cleaner method than the previous effort. There is some > differences: > > 1. yamauchi's patch works for both iFDO and autoFDO, while this patch limits > to iFDO > 2. yama

[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

aeubanks added a comment. In D149800#4330831 , @davidxl wrote: > This patch uses a cleaner method than the previous effort. There is some > differences: > > 1. yamauchi's patch works for both iFDO and autoFDO, while this patch limits > to iFDO > 2. yama

[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

davidxl added a comment. This patch uses a cleaner method than the previous effort. There is some differences: 1. yamauchi's patch works for both iFDO and autoFDO, while this patch limits to iFDO 2. yamauchi's patch also handles size optimization (IIRC) at granularity smaller than function Lo

[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

aeubanks added a comment. In D149800#4317275 , @mtrofin wrote: > high level question, why not have it as a pass that runs after profiles (of > whatever kind - instrumented or sample-based) are ingested. The pass would > attribute functions as described.

[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

tejohnson added a subscriber: yamauchi. tejohnson added a comment. Previously @yamauchi did a bunch of work related to this called PGSO (Profile Guided Size Optimization). See the functions in https://github.com/llvm-mirror/llvm/blob/master/lib/Transforms/Utils/SizeOpts.cpp, which are used to g

[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

mtrofin added a comment. high level question, why not have it as a pass that runs after profiles (of whatever kind - instrumented or sample-based) are ingested. The pass would attribute functions as described. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.l

[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

aeubanks added a comment. looking for feedback if this makes sense at all Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D149800/new/ https://reviews.llvm.org/D149800 ___ cfe-commits mailing list cfe-commi

[PATCH] D149800: [WIP][PGO] Add ability to mark cold functions as optsize/minsize/optnone

aeubanks created this revision. Herald added subscribers: wlei, Enna1, ormris, wenlei, steven_wu, hiraditya. Herald added a project: All. aeubanks requested review of this revision. Herald added projects: clang, LLVM. Herald added subscribers: llvm-commits, cfe-commits. The performance of cold fun