David Green 2017b8f59b [LV] Allow tail folded reduction selects to remain in the loop ... The normal scheme for tail folding reductions is to use: loop: p = phi(0, a) mask = ... x = masked_load(..., mask) a = add(x, p) s = select(mask, a, p) This means we need to keep the register p and a alive out of the loop, plus the mask. On a target with predicated operations we can instead generate the phi as p = phi(0, s). This ensures the select in the loop and we can fold select(m, add(a, b), c) to something like a vaddt c, a, b using the m predicate. This in turn allows us to tail predicate the entire loop. Differential Revision: https://reviews.llvm.org/D84741		2020-08-20 14:31:14 +01:00
..
AggressiveInstCombine	[SVE] Remove calls to VectorType::getNumElements from AggressiveInstCombine	2020-08-14 16:40:34 -07:00
CFGuard
Coroutines	[Coroutines] Use to collect lifetime marker of in CoroFrame Differential Revision: https://reviews.llvm.org/D85279	2020-08-06 14:21:55 +08:00
Hello
InstCombine	[ValueTracking] define/use max recursion depth in header	2020-08-19 16:56:59 -04:00
Instrumentation	[MSAN] Avoid dangling ActualFnStart when replacing instruction	2020-08-14 23:50:38 +00:00
IPO	[Attributor] Handle CallBase case in AAValueConstantRange::initialize	2020-08-20 20:15:19 +09:00
ObjCARC	[NFC] Remove unused GetUnderlyingObject paramenter	2020-07-31 02:10:03 -07:00
Scalar	Reland "[LoopUnswitch] Fix incorrect Modified status"	2020-08-20 11:52:09 +02:00
Utils	Reland "[SLC] sprintf(dst, "%s", str) -> strcpy(dst, str)"	2020-08-15 12:14:57 +02:00
Vectorize	[LV] Allow tail folded reduction selects to remain in the loop	2020-08-20 14:31:14 +01:00
CMakeLists.txt
LLVMBuild.txt