Quentin Colombet a09cf4d3ee [MachineSink] Improve the compile time by preserving the dominance information ... as long as possible. ** Context ** Each time the dominance information is modified, the dominator tree analysis switches in a slow query mode. After a few queries without any modification on the dominator tree, it performs an expensive update of its internal structure to provide fast queries again. ** Problem ** Prior to this patch, the MachineSink pass was splitting the critical edges on demand while relying heavy on the dominator tree information. In some cases, this leads to pathological behavior where: - We end up in the slow query mode right after splitting an edge. - We update the dominance information. - We break the dominance information again, thus ending up in the slow query mode and so on. ** Proposed Solution ** To mitigate this effect, this patch postpones all the splitting of the edges at the end of each iteration of the main loop. The benefits are: - The dominance information is valid for the life time of an iteration. - This simplifies the code as we do not have to special treat instructions that are sunk on critical edges. Indeed, the related block will be available through the next iteration. The downside is that when edges splitting is required, this incurs an additional iteration of the main loop compared to the previous scheme. ** Performance ** Thanks to this patch, the motivating example compiles in 6+ minutes instead of 10+ minutes. No test case added as the motivating example as nothing special but being huge! I have measured only noise for both the compile time and the runtime on the llvm test-suite + SPECs with Os and O3. Note: The current implementation of MachineBasicBlock::SplitCriticalEdge also uses the dominance information and therefore, hits this problem. A subsequent patch will address that. <rdar://problem/17894619> llvm-svn: 215410		2014-08-11 23:52:01 +00:00
..
Analysis	In LVI(Lazy Value Info), originally value on a BB can only be caculated once,	2014-08-11 05:02:04 +00:00
AsmParser	Remove dead code in condition	2014-08-05 18:22:58 +00:00
Bitcode	BitcodeReader: Fix non-determinism in use-list order	2014-08-05 17:49:48 +00:00
CodeGen	[MachineSink] Improve the compile time by preserving the dominance information	2014-08-11 23:52:01 +00:00
DebugInfo	DWOHolder takes ownership of the argument constructor, use std::unique_ptr.	2014-07-31 20:26:42 +00:00
ExecutionEngine	[MCJIT] Simplify immediate decoding code in the RuntimeDyldMachO hierarchy.	2014-08-08 23:12:22 +00:00
IR	Fix typos:	2014-08-11 18:04:46 +00:00
IRReader	Update the MemoryBuffer API to use ErrorOr.	2014-07-06 17:43:13 +00:00
LineEditor
Linker	Fix use of uninitialized variable.	2014-08-11 17:07:34 +00:00
LTO	Don't internalize all but main by default.	2014-08-05 20:10:38 +00:00
MC	MC: Diagnose an unexpected token in COFF .section instead of asserting	2014-08-11 18:34:43 +00:00
Object	AArch64: add support for dynamic-loader relocations	2014-08-11 10:10:27 +00:00
Option	Generic: add range-adapter for option parsing.	2014-07-09 13:03:37 +00:00
ProfileData	Coverage: add HasCodeBefore flag to a mapping region.	2014-08-04 18:00:51 +00:00
Support	ADT: remove MinGW32 and Cygwin OSType enum	2014-08-09 23:12:20 +00:00
TableGen	[tablegen] - Eliminate memory leaks in TGParser.cpp	2014-08-08 00:29:54 +00:00
Target	R600/SIInstrInfo.cpp: Suppress an warning. [-Wunused-variable]	2014-08-11 23:03:38 +00:00
Transforms	InstCombine: Combine (add (and %a, %b) (or %a, %b)) to (add %a, %b)	2014-08-11 22:32:02 +00:00
CMakeLists.txt
LLVMBuild.txt
Makefile