Lines Matching full:sinking

11 // code into the preheader block, or by sinking code to the exit blocks if it is
13 // live in registers, thus hoisting and sinking "invariant" loads and stores.
160 // Experimentally, memory promotion carries less importance than sinking and
459 // us to sink instructions in one pass, without iteration. After sinking in runOnLoop()
479 // preheader for SSA updater, so also avoid sinking when no preheader in runOnLoop()
582 // we just delete it instead of sinking it. in sinkRegion()
1398 // We need to sink a callsite to a unique funclet. Avoid sinking if the in isNotUsedOrFoldableInLoop()
1433 // Sinking call-sites need to be handled differently from other in cloneInstructionInExitBlock()
1485 // OPT: If this shows up in a profile, we can instead finish sinking all in cloneInstructionInExitBlock()
1488 // sinking bottom-up. in cloneInstructionInExitBlock()
1636 LLVM_DEBUG(dbgs() << "LICM sinking instruction: " << I << "\n"); in sink()
1638 // Iterate over users to be ready for actual sinking. Replace users via in sink()
1690 << "sinking " << ore::NV("Inst", &I); in sink()
1711 // First check if I is worth sinking for all uses. Sink only when it is worth in sink()
1962 /// Try to promote memory values to scalars by sinking stores out of the
2133 // start sinking stores into unwind edges (see above), this will break. in promoteLoopAccessesToScalars()
2371 // For sinking, we'd need to check all Defs below this use. The getClobbering in pointerInvalidatedByLoop()
2381 // However sinking the load outside the loop, below the store is incorrect. in pointerInvalidatedByLoop()
2393 // When sinking, the source block may not be part of the loop so check it. in pointerInvalidatedByLoop()