Lines Matching refs:users
20 // "post-inc" users. This is not talking about post-increment addressing modes;
33 // case of post-increment users is users outside the loop.
42 // TODO: When truncation is free, truncate ICmp users' operands to make it a
128 /// bail out. This threshold is far beyond the number of users that LSR can
944 for (User *UR : UVal->users()) {
1004 /// still return true. Therefore, users still need to account for
1131 /// users of the same expression in common?
1853 /// chain collection, before LSR replaces IV users. During chain generation,
1925 /// dominate all the in-loop post-increment users.
1950 /// IV users can form a chain of IV increments.
1953 /// IV users that belong to profitable IVChains.
2420 // exits if there may be pre-inc users in intervening blocks.
2474 // possible for it to have multiple users. If it is not immediately before
2824 /// be less than the number required for the individual IV users. First prohibit
2825 /// any IV users that keep the IV live across increments (the Users set should
2843 LLVM_DEBUG(dbgs() << "Chain: " << *Chain.Incs[0].UserInst << " users:\n";
2987 // transitively follow its user and only add leaf IV users to the set.
2988 for (User *U : IVOper->users()) {
3057 // Ignore users that are part of a SCEV expression. This way we only
3247 // Skip IV users that are part of profitable IV Chains.
3388 /// TODO: Should this give more weight to users inside the loop?
3977 // As a special-case, handle special out-of-loop Basic users specially.
5318 // users.
5478 // Expand the new value definitions and update the users.
5537 // IVUsers analysis should only create users that are dominated by simple loop
5538 // headers. Since this loop should dominate all of its users, its user list
5558 // If loop preparation eliminates all interesting IV users, bail.