2005-04-13 |
Nate Begeman | Implement the fold shift X, zext(Y) -> shift X, Y at... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Nate Begeman | Add CodeGen tests for the recent SelectionDAG transforms |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Nate Begeman | Disbale the broken fold of shift + sz[ext] for now |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | fix an infinite loop |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | fix some serious miscompiles on ia64, alpha, and ppc |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | avoid work when possible, perhaps fix the problem nate... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Andrew Lenharth | WOW, function calls still seem to work after this. |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Andrew Lenharth | prepare for func call optimization |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Andrew Lenharth | regression case for faster call sequence |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Andrew Lenharth | check that casts still use zap |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Duraid Madina | * add the shladd instruction |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Andrew Lenharth | add matches for SxADDL and company, as well as simplify... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Implement expansion of unsigned i64 -> FP. |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Duraid Madina | * if ANDing with a constant of the form: |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Andrew Lenharth | added s4addl matching test |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Andrew Lenharth | added all flavors of zap for anding |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Make expansion of uint->fp cast assert out instead... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Fix some mysteriously missing {}'s which cause the... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | add back the optimization that Nate added for shl X... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Oops, remove these too. |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | remove one more occurance of this that snuck in |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Remove support for ZERO_EXTEND_INREG. This pessimizes... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Remove special handling of ZERO_EXTEND_INREG. This... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Elimate handling of ZERO_EXTEND_INREG. This causes... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Z_E_I is gone |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Instead of making ZERO_EXTEND_INREG nodes, use the... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Remove all foldings of ZERO_EXTEND_INREG, moving them... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Add a new helper method which returns the and that... |
commit | commitdiff | tree | snapshot |
2005-04-13 |
Chris Lattner | Remove the ZERO_EXTEND_INREG node which is redundant... |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Nate Begeman | Fold shift x, [sz]ext(y) -> shift x, y |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Nate Begeman | Fold shift by size larger than type size to undef |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Nate Begeman | Implement setcc op, -1 sequences |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | promote extload i1 -> extload i8 |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | add an argument to allow avoiding deleting phi nodes. |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | add an argument. |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | Get rid of this for_each loop |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Duraid Madina | * OK, after changing to use liveIn/liveOut instead... |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Andrew Lenharth | Get rid of idefs for arguments (oops) |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Andrew Lenharth | Get rid of idefs for arguments |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | Put out* into the allocation order, allowing the regist... |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | Make sure to realize that calls use their argument... |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Duraid Madina | stop emitting IDEFs for args - change to using liveIn... |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Nate Begeman | Initial support for allocation condition registers |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | Fix a crash analyzing MultiSource/Benchmarks/MallocBench/gs |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | Remove some redundant checks, add a couple of new ones... |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | Emit comparisons against the sign bit better. Codegen... |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | Emit long comparison against -1 better. Instead of... |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Chris Lattner | canonicalize x <u 1 -> x == 0. On this testcase: |
commit | commitdiff | tree | snapshot |
2005-04-12 |
Nate Begeman | Implement bitfield clears |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Nate Begeman | Update PPC readme. Remove things that are done or... |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Chris Lattner | Teach the dag mechanism that this: |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Chris Lattner | Fix expansion of shifts by exactly NVT bits on arch... |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Chris Lattner | IA64 supports this operation. |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Chris Lattner | ORo sets CR0 |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Chris Lattner | Revert the previous patch, which I didn't mean to check in. |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Chris Lattner | Fix a minor bug (ORo didn't mark that it set CR0). |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Duraid Madina | hmm, should probably change addImm() to take 64-bit... |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Duraid Madina | rename addU64Imm() to addImm64() |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Nate Begeman | Add recording variants of ISD::AND and ISD::OR. This... |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Duraid Madina | assorted fixes: |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Reid Spencer | Ensure that the arguments passed to sys::Program::Execu... |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Jeff Cohen | Eliminate tabs |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Jeff Cohen | Eliminate major source of VC++ "possible loss of data... |
commit | commitdiff | tree | snapshot |
2005-04-11 |
Nate Begeman | Fix libcall code to not pass a NULL Chain to LowerCallTo |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | Don't bother sign/zext_inreg'ing the result of an and... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | Teach legalize to deal with targets that don't support... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | The first argument to ExecuteAndWait should be the... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | fix this testcase so the regex doesn't match the functi... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | don't zextload fp values! |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Duraid Madina | * store immediate values as int64_t, not int. come... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Nate Begeman | Fix another fixme: factor out the constant fp generatio... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Nate Begeman | Fix 64 bit argument loading that straddles the args... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | Until we have a dag combiner, promote using zextload... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | Fold zext_inreg(zextload), likewise for sext's |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | add a simple xform |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Nate Begeman | Remove unnecessary Implicit Defs. Since r0 is not... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | make this harder |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | oops add ~ |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | new testcase for previously unsupported unary complex... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Nate Begeman | Make sure that BRCOND branches can be converted into... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Nate Begeman | Don't hand ISD::CALL nodes off to SelectExprFP. This... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | Fix a thinko. If the operand is promoted, pass the... |
commit | commitdiff | tree | snapshot |
2005-04-10 |
Chris Lattner | rename getPPCOpcodeForSetCCNumber -> getPPCOpcodeForSet... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Nate Begeman | fix ISD::BRCONDTWOWAY codegen to not deference the... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | Fix CodeGen/Generic/2005-05-09-GlobalInPHI.ll, which... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | new testcase that used to crash the ppc fe. It could... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | add a little peephole optimization. This allows us... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | do not set the root to null if an argument is dead |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Nate Begeman | Add rlwnm instruction for variable rotate |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | Fix a crash on 173.applu by asking for a constant bigge... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | Switch this instruction selector over to using liveins... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | there is no need to remove this instruction, linscan... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | Adjust live intervals to support a livein set |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | Use live out sets for return values instead of imp_defs... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | Consider the livein/out set for a function, allowing... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Chris Lattner | add routines to track the livein/out set for a function |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Duraid Madina | ok, the "ia64 has a boatload of registers" joke stopped... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Nate Begeman | Optimize FSEL a bit for fneg arguments. This fixes... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Nate Begeman | Add a testcase to make sure that we don't emit two... |
commit | commitdiff | tree | snapshot |
2005-04-09 |
Nate Begeman | Add cases to cover the rest of the patterns we should... |
commit | commitdiff | tree | snapshot |
next |