Log - 16999da951677a94a2f30d98c8126ff175f457e1 - fp2-dev/platform/external/llvm

9b041c9 Anti-dependency breaking needs to be careful not to use reserved regs by Jim Grosbach · 15 years ago
9c2a034 Use std::vector instead of TargetRegisterInfo::FirstVirtualRegister. This time by Bill Wendling · 15 years ago
e4a4147 Reserve a goodly amount of room for the vectors. by Bill Wendling · 15 years ago
38306d5 Use std::vector instead of TargetRegisterInfo::FirstVirtualRegister. by Bill Wendling · 15 years ago
7fa889b revert bill's patches in an attempt to fix the buildbot. by Chris Lattner · 15 years ago
75a5b71 Fix headers. by Bill Wendling · 15 years ago
e010409 Use std::vector instead of a hard-coded array. The length of that array could by Bill Wendling · 15 years ago
7e1b566 Convert the last use of getPhysicalRegisterRegClass and remove it. by Rafael Espindola · 15 years ago
46df4eb Make post-ra scheduling, anti-dep breaking, and register scavenger (conservatively) aware of predicated instructions. This enables ARM to move if-conversion before post-ra scheduler. by Evan Cheng · 15 years ago
086723d Not all entries in the range will have an SUnit. Check for that when looking by Jim Grosbach · 15 years ago
533934e Update debug information when breaking anti-dependencies. rdar://7759363 by Jim Grosbach · 15 years ago
66db3a0 Make BreakAntiDependencies' SUnits argument const, and make the Begin by Dan Gohman · 15 years ago
347fa3f Tidy whitespace. by Bob Wilson · 15 years ago
518bb53 move target-independent opcodes out of TargetInstrInfo by Chris Lattner · 16 years ago
2973b57 80 column and whitespace cleanup by Jim Grosbach · 16 years ago
5393b25 by David Greene · 16 years ago
990d285 <rdar://problem/7453528>. Track only physical registers that are valid for the target. by David Goodwin · 16 years ago
2601329 <rdar://problem/6721894>. Allow multiple registers to be renamed together (super and sub) if necessary to break an anti-dependence. by David Goodwin · 16 years ago
00621ef Restructure code to allow renaming of multiple-register groups for anti-dep breaking. by David Goodwin · 16 years ago
557bbe6 Remove some old experimental code that is no longer needed. Remove additional, speculative scheduling pass as its cost did not translate into significant performance improvement. Minor tweaks. by David Goodwin · 16 years ago
3e72d30 Fix a couple of problems with maintaining liveness information for antidep breaking. by David Goodwin · 16 years ago
87d21b9 Allow target to specify regclass for which antideps will only be broken along the critical path. by David Goodwin · 16 years ago
12dd99d Rename registers to break output dependencies in addition to anti-dependencies. by David Goodwin · 16 years ago
0855dee Allow targets to specify register classes whose member registers should not be renamed to break anti-dependencies. by David Goodwin · 16 years ago
7040d6e Fix bug in aggressive antidep breaking; liveness was not updated correctly for regions that do not have antidep candidates. by David Goodwin · 16 years ago
98f2f1a Replace std::map.at() with std::map[]. by David Goodwin · 16 years ago
5409783 Break anti-dependencies using free registers in a round-robin manner to avoid introducing new anti-dependencies. by David Goodwin · 16 years ago
4de099d Do a scheduling pass ignoring anti-dependencies to identify candidate registers that should be renamed. by David Goodwin · 16 years ago
5b3c308 Between scheduling regions, correctly maintain anti-dep breaking state so that we don't incorrectly rename registers that span these regions. by David Goodwin · 16 years ago
67a8a7b Fix a couple of bugs in aggressive anti-dep breaking. by David Goodwin · 16 years ago
e10deca Allow the aggressive anti-dep breaker to process the same region multiple times. This is necessary because new anti-dependencies are exposed when "current" ones are broken. by David Goodwin · 16 years ago
3487771 Add aggressive anti-dependence breaker. Currently it is not the default for any target. Enable with -break-anti-dependencies=all. by David Goodwin · 16 years ago