| Commit message (Expand) | Author | Age | Files | Lines |
* | so hey, it turns out that the histogram was completely wrong, because | Chris Lattner | 2010-03-04 | 1 | -10/+32 |
* | change the new isel matcher to emit ComplexPattern matches | Chris Lattner | 2010-03-04 | 1 | -6/+8 |
* | enhance comment output to specify what recorded slot | Chris Lattner | 2010-03-04 | 1 | -2/+5 |
* | introduce a new SwitchTypeMatcher node (which is analogous to | Chris Lattner | 2010-03-03 | 1 | -10/+35 |
* | Rewrite chain handling validation and input TokenFactor handling | Chris Lattner | 2010-03-02 | 1 | -5/+0 |
* | add some missing \n's | Chris Lattner | 2010-03-02 | 1 | -11/+19 |
* | fixme resolved. | Chris Lattner | 2010-03-01 | 1 | -3/+0 |
* | remove a little hack I did for the old isel, not needed | Chris Lattner | 2010-03-01 | 1 | -4/+0 |
* | Missed a \n in previous commit. | Torok Edwin | 2010-03-01 | 1 | -0/+1 |
* | Add command-line flag to tblgen to turn off generating comments for the new | Torok Edwin | 2010-03-01 | 1 | -57/+117 |
* | eliminate the CheckMultiOpcodeMatcher code and have each | Chris Lattner | 2010-03-01 | 1 | -11/+1 |
* | add a new OPC_SwitchOpcode which is semantically equivalent | Chris Lattner | 2010-03-01 | 1 | -2/+50 |
* | enhance RecordNode and RecordChild comments to indicate what | Chris Lattner | 2010-03-01 | 1 | -2/+4 |
* | inline the node transforms and node predicates into the generated | Chris Lattner | 2010-03-01 | 1 | -13/+56 |
* | simplify some code now that chain/flag results are not stored in | Chris Lattner | 2010-02-28 | 1 | -1/+1 |
* | don't emit useless functions. These were producing | Chris Lattner | 2010-02-28 | 1 | -47/+56 |
* | change a few opcodes to use VBRs instead of embedding | Chris Lattner | 2010-02-28 | 1 | -60/+15 |
* | enhance the EmitNode/MorphNodeTo operands to take a bit that | Chris Lattner | 2010-02-28 | 1 | -1/+2 |
* | use MorphNodeTo instead of SelectNodeTo. SelectNodeTo | Chris Lattner | 2010-02-28 | 1 | -4/+4 |
* | enhance the new isel to use SelectNodeTo for most patterns, | Chris Lattner | 2010-02-28 | 1 | -6/+11 |
* | enhance EmitNodeMatcher to keep track of the recorded slot numbers | Chris Lattner | 2010-02-28 | 1 | -1/+15 |
* | add infrastructure to support forming selectnodeto. Not used yet | Chris Lattner | 2010-02-28 | 1 | -3/+6 |
* | change CheckOpcodeMatcher to hold the SDNodeInfo instead of | Chris Lattner | 2010-02-27 | 1 | -5/+5 |
* | add some helpful comments to the emitter | Chris Lattner | 2010-02-26 | 1 | -0/+6 |
* | change the scope node to include a list of children to be checked | Chris Lattner | 2010-02-25 | 1 | -48/+62 |
* | formatting. | Chris Lattner | 2010-02-25 | 1 | -6/+3 |
* | rename fooMatcherNode to fooMatcher. | Chris Lattner | 2010-02-25 | 1 | -109/+108 |
* | rename PushMatcherNode -> ScopeMatcherNode to more accurately | Chris Lattner | 2010-02-25 | 1 | -9/+9 |
* | contract movechild+checktype into a new checkchild node, shrinking the | Chris Lattner | 2010-02-24 | 1 | -1/+7 |
* | emit a histogram of the opcodes in comments. | Chris Lattner | 2010-02-24 | 1 | -2/+59 |
* | Since the new instruction selector now works, I don't need to keep | Chris Lattner | 2010-02-24 | 1 | -1/+1 |
* | implement a simple proof-of-concept optimization for | Chris Lattner | 2010-02-24 | 1 | -0/+7 |
* | The new isel was not properly handling patterns that covered | Chris Lattner | 2010-02-24 | 1 | -0/+9 |
* | really fix an off-by-one error | Chris Lattner | 2010-02-23 | 1 | -1/+1 |
* | switch the value# in OPC_CompleteMatch and OPC_EmitNode to use a | Chris Lattner | 2010-02-23 | 1 | -5/+29 |
* | add a new Push2 opcode for targets (like cellspu) which have | Chris Lattner | 2010-02-22 | 1 | -5/+18 |
* | add a new CheckMultiOpcode opcode for checking that a node | Chris Lattner | 2010-02-22 | 1 | -0/+9 |
* | emit table indexes before each row so that it is debuggable. | Chris Lattner | 2010-02-21 | 1 | -7/+11 |
* | fix a table size miscomputation, target opcodes are 2 bytes. | Chris Lattner | 2010-02-21 | 1 | -1/+1 |
* | emit to the right streams, to avoid emitting the push | Chris Lattner | 2010-02-21 | 1 | -15/+16 |
* | implement the last known missing feature: updating uses of results | Chris Lattner | 2010-02-21 | 1 | -5/+11 |
* | Lots of improvements to the new dagisel emitter. This gets it to | Chris Lattner | 2010-02-21 | 1 | -18/+126 |
* | add emitter support for integer constants and simple physreg references. | Chris Lattner | 2010-02-19 | 1 | -3/+14 |
* | add support for referencing registers and immediates, | Chris Lattner | 2010-02-18 | 1 | -2/+7 |
* | start sketching out the structure of code for result emission generation. | Chris Lattner | 2010-02-18 | 1 | -0/+3 |
* | rename the child field to 'next'. This is not a parent/child | Chris Lattner | 2010-02-18 | 1 | -12/+12 |
* | eliminate the MatcherNodeWithChild class, give the 'child' | Chris Lattner | 2010-02-18 | 1 | -5/+3 |
* | improve comments, the matcher is now feature complete, on to codegen. | Chris Lattner | 2010-02-17 | 1 | -1/+5 |
* | Emulate the current isel's "IsChainCompatible" logic for now. | Chris Lattner | 2010-02-17 | 1 | -0/+4 |
* | improve comments on OPC_Record to say what we're recording a node. | Chris Lattner | 2010-02-17 | 1 | -1/+3 |