Chandler Carruth 68ba25c47d Completely re-write the algorithm behind MachineBlockPlacement based on ... discussions with Andy. Fundamentally, the previous algorithm is both counter productive on several fronts and prioritizing things which aren't necessarily the most important: static branch prediction. The new algorithm uses the existing loop CFG structure information to walk through the CFG itself to layout blocks. It coalesces adjacent blocks within the loop where the CFG allows based on the most likely path taken. Finally, it topologically orders the block chains that have been formed. This allows it to choose a (mostly) topologically valid ordering which still priorizes fallthrough within the structural constraints. As a final twist in the algorithm, it does violate the CFG when it discovers a "hot" edge, that is an edge that is more than 4x hotter than the competing edges in the CFG. These are forcibly merged into a fallthrough chain. Future transformations that need te be added are rotation of loop exit conditions to be fallthrough, and better isolation of cold block chains. I'm also planning on adding statistics to model how well the algorithm does at laying out blocks based on the probabilities it receives. The old tests mostly still pass, and I have some new tests to add, but the nested loops are still behaving very strangely. This almost seems like working-as-intended as it rotated the exit branch to be fallthrough, but I'm not convinced this is actually the best layout. It is well supported by the probabilities for loops we currently get, but those are pretty broken for nested loops, so this may change later. llvm-svn: 142743		2011-10-23 09:18:45 +00:00
..
Analysis	Make SCEV's brute force analysis stronger in two ways. Firstly, we should be	2011-10-22 19:58:20 +00:00
Archive
Assembler	Fix a typo in the bitcode reader in the handling of atomic stores. Reported by David Meyer on llvmdev.	2011-09-19 19:41:28 +00:00
Bindings/Ocaml	OCaml bindings: add some missing functions and testcases.	2011-10-14 20:38:33 +00:00
Bitcode
BugPoint
CodeGen	Completely re-write the algorithm behind MachineBlockPlacement based on	2011-10-23 09:18:45 +00:00
DebugInfo	Add a new wrapper node for a DILexicalBlock that encapsulates it and a	2011-10-11 22:59:11 +00:00
ExecutionEngine	Excluding ARM JIT tests until someone can fix this compilation path.	2011-08-26 23:39:30 +00:00
Feature	Update tests to the newest EH syntax.	2011-08-25 01:30:18 +00:00
Integer
lib	s/tblgen/llvm-tblgen/g in a few missed places, including the tests	2011-10-06 13:39:59 +00:00
Linker	Let this test pass even if 'int' is somewhere in its directory path.	2011-10-15 18:07:16 +00:00
MC	Add X86 RORX instruction	2011-10-23 07:34:00 +00:00
Object	llvm-objdump: Add static symbol table dumping.	2011-10-18 19:32:17 +00:00
Other	Update to new EH scheme.	2011-08-31 21:44:24 +00:00
Scripts	Fix the bitwidth of the remaining fields.	2011-08-04 17:00:11 +00:00
TableGen	Add Paste Test	2011-10-19 13:04:50 +00:00
Transforms	The element insertion code in scalar replacement doesn't handle incorrect	2011-10-23 07:02:10 +00:00
Unit
Verifier
CMakeLists.txt	test/CMakeLists.txt: Depend on llvm-objdump. "make check" is expected to resolve test-dependent targets on CMake build.	2011-09-27 21:54:50 +00:00
lit.cfg	s/tblgen/llvm-tblgen/g in a few missed places, including the tests	2011-10-06 13:39:59 +00:00
lit.site.cfg.in
Makefile	test/Makefile: Inspect $(PROJ_OBJ_ROOT)/tools/clang/Makefile instead of $(PROJ_SRC_ROOT)/tools/clang for "check-all".	2011-10-16 02:54:14 +00:00
Makefile.tests
site.exp.in
TestRunner.sh