- 3fe6c1d Legalize is no longer limited to cleverness with just constant shift amounts. by Chris Lattner · 18 years ago
- cf9d0ac Fold the PPCISD shifts when presented with 0 inputs. This occurs for code by Chris Lattner · 18 years ago
- c356a57 Reflects MachineConstantPoolEntry changes. by Evan Cheng · 18 years ago
- 3a9ec24 For PR387: by Reid Spencer · 18 years ago
- f6e190f Fix a bug in a recent refactoring that broke a bunch of stuff. by Chris Lattner · 18 years ago
- e219945 Eliminate use of getNode that takes a vector. by Chris Lattner · 18 years ago
- 79e490a Convert vectors to fixed sized arrays and smallvectors. Eliminate use of getNode that takes a vector. by Chris Lattner · 18 years ago
- 325f0a1 Fix miscompilation of float vector returns. Compile code to this: by Chris Lattner · 18 years ago
- 0d72a20 Fix some ppc64 issues with vector code. by Chris Lattner · 18 years ago
- 35d86fe Rename RelocModel::PIC to PIC_, to avoid conflicts with -DPIC. by Chris Lattner · 18 years ago
- d998938 Implement Regression/CodeGen/PowerPC/bswap-load-store.ll by folding bswaps by Chris Lattner · 18 years ago
- f89437d Implement 64-bit select, bswap, etc. by Chris Lattner · 18 years ago
- 5f9faea PPC doesn't have bit converts to/from i64 by Chris Lattner · 18 years ago
- 563ecfb Implement 64-bit undef, sub, shl/shr, srem/urem by Chris Lattner · 18 years ago
- 7b0c58c Use i32 for shift amounts instead of i64. This gets bisort working. by Chris Lattner · 18 years ago
- c08f902 Implement a bunch of 64-bit cleanliness work. With this, treeadd builds (but by Chris Lattner · 18 years ago
- c91a475 Improve PPC64 calling convention support by Chris Lattner · 18 years ago
- ef95710 Correct returns of 64-bit values, though they seemed to work before... by Chris Lattner · 18 years ago
- 059ca0f fix some assumptions that pointers can only be 32-bits. With this, we can by Chris Lattner · 18 years ago
- a7a5854 Rename some subtarget features. A CPU now can *have* 64-bit instructions, by Chris Lattner · 18 years ago
- a7dc4a5 Type of extract_element index operand should be iPTR. by Evan Cheng · 18 years ago
- 4a45abf Fix a problem exposed by the local allocator. CALL instructions are not marked by Chris Lattner · 18 years ago
- 7b05350 Always reserve space for 8 spilled GPRs. GCC apparently assumes that this by Chris Lattner · 18 years ago
- 6848be1 Change RET node to include signness information of the return values. i.e. by Evan Cheng · 18 years ago
- 4360bdc CALL node change (arg / sign pairs instead of just arguments). by Evan Cheng · 18 years ago
- d74ea2b Patches to make the LLVM sources more -pedantic clean. Patch provided by Chris Lattner · 18 years ago
- 2ef5e89 Fix CodeGen/Generic/vector.ll:test_div with altivec. by Chris Lattner · 18 years ago
- 5734012 Handle SETO* like we handle SET*, restoring behavior after Evan's setcc by Chris Lattner · 18 years ago
- c703a8f Make PPC call lowering more aggressive, making the isel matching code simple by Chris Lattner · 18 years ago
- 9a2a497 Switch PPC over to a call-selection model where the lowering code creates by Chris Lattner · 18 years ago
- c8b682c 3 changes, 2 of which are cleanup one of which changes codegen: by Chris Lattner · 18 years ago
- c04ba7a implement passing/returning vector regs to calls, at least non-varargs calls. by Chris Lattner · 18 years ago
- abde460 Instead of implementing LowerCallTo directly, let the default impl produce an by Chris Lattner · 18 years ago
- af4ec0c Simplify the argument counting logic by only incrementing the index. by Chris Lattner · 18 years ago
- b375b5e Simplify the dead argument handling code. by Chris Lattner · 18 years ago
- be4849a Vector args passed in registers don't reserve stack space. by Chris Lattner · 18 years ago
- 8ab5fe5 Switch the PPC backend over to using FORMAL_ARGUMENTS for formal argument by Chris Lattner · 18 years ago
- 00402c7 Fit in 80 cols by Chris Lattner · 18 years ago
- b65e725 Remove dead var, fix bad override. by Chris Lattner · 18 years ago
- 25b8b8c Fix CodeGen/Generic/2006-04-28-Sign-extend-bool.ll by Chris Lattner · 18 years ago
- 37efe67 JumpTable support! What this represents is working asm and jit support for by Nate Begeman · 18 years ago
- 0090120 Fix a crash on: by Chris Lattner · 18 years ago
- f70f8d9 pretty print node name by Chris Lattner · 18 years ago
- 90564f2 Implement an important entry from README_ALTIVEC: by Chris Lattner · 18 years ago
- cea2aa7 Use vmladduhm to do v8i16 multiplies which is faster and simpler than doing by Chris Lattner · 18 years ago
- 19a8152 Implement v16i8 multiply with this code: by Chris Lattner · 18 years ago
- 72dd9bd Lower v8i16 multiply into this code: by Chris Lattner · 18 years ago
- e7c768e Custom lower v4i32 multiplies into a cute sequence, instead of having legalize by Chris Lattner · 18 years ago
- dbce85d Make sure to check splats of every constant we can, handle splat(31) by by Chris Lattner · 18 years ago
- bdd558c Teach the ppc backend to use rol and vsldoi to generate splatted constants. by Chris Lattner · 18 years ago
- 6876e66 Make some code more general, adding support for constant formation of several by Chris Lattner · 18 years ago
- c408382 Learn how to make odd splatted constants in range [17,29]. This implements by Chris Lattner · 18 years ago
- 4a998b9 Pull some code out into a helper function. by Chris Lattner · 18 years ago
- 5913810 Implement a TODO: for any shuffle that can be viewed as a v4[if]32 shuffle, by Chris Lattner · 18 years ago
- f3f69de Implement a TODO: have the legalizer canonicalize a bunch of operations to by Chris Lattner · 18 years ago
- b17f167 Make the BUILD_VECTOR lowering code much more aggressive w.r.t constant vectors. by Chris Lattner · 18 years ago
- 730b456 Fix a crash when faced with a shuffle vector that has an undef in its mask. by Chris Lattner · 18 years ago
- b097aa9 Allow undef in a shuffle mask by Chris Lattner · 19 years ago
- 1a635d6 Move the rest of the PPCTargetLowering::LowerOperation cases out into by Chris Lattner · 19 years ago
- f1b4708 Pull the VECTOR_SHUFFLE and BUILD_VECTOR lowering code out into separate by Chris Lattner · 19 years ago
- a39d798 Force non-darwin targets to use a static relo model. This fixes PR734, by Chris Lattner · 19 years ago
- ac225ca Add a new way to match vector constants, which make it easier to bang bits of by Chris Lattner · 19 years ago
- e87192a Rename get_VSPLI_elt -> get_VSPLTI_elt by Chris Lattner · 19 years ago
- 2b1c325 Ensure that zero vectors are always v4i32, which forces them to CSE with by Chris Lattner · 19 years ago
- bee9836 Vector function results go into V2 according to GCC. The darwin ABI doc by Chris Lattner · 19 years ago
- 06c2435 Move some return-handling code from lowerarguments to the ISD::RET handling stuff. by Chris Lattner · 19 years ago
- a1d95e1 properly mark vector selects as expanded to select_cc by Chris Lattner · 19 years ago
- 710ff32 Add VRRC select support by Chris Lattner · 19 years ago
- 79d9a88 Implement PowerPC/CodeGen/vec_splat.ll:spltish to use vsplish instead of a by Chris Lattner · 19 years ago
- 140a58f Change the interface to the predicate that determines if vsplti* can be used. by Chris Lattner · 19 years ago
- 9021799 Make sure to return the result in the right type. by Chris Lattner · 19 years ago
- f24380e Match vpku[hw]um(x,x). by Chris Lattner · 19 years ago
- caad163 Add support for matching vmrg(x,x) patterns by Chris Lattner · 19 years ago
- 116cc48 Pattern match vmrg* instructions, which are now lowered by the CFE into shuffles. by Chris Lattner · 19 years ago
- d0608e1 Support pattern matching vsldoi(x,y) and vsldoi(x,x), which allows the f.e. to by Chris Lattner · 19 years ago
- ddb739e Compile the vpkuhum/vpkuwum intrinsics into vpkuhum/vpkuwum instead of into by Chris Lattner · 19 years ago
- 9b42bdd Fix CodeGen/PowerPC/2006-04-05-splat-ish.ll by Chris Lattner · 19 years ago
- 278158b Fallthrough to expand if a VECTOR_SHUFFLE cannot be custom lowered. by Evan Cheng · 19 years ago
- 098e699 Fix some broken logic that would cause us to codegen {2147483647,2147483647,2147483647,2147483647} as 'vspltisb v0, -1'. by Chris Lattner · 19 years ago
- 7ff7e67 Ask legalize to promote all vector shuffles to be v16i8 instead of having to by Chris Lattner · 19 years ago
- b683144 Revert accidentally committed hunks. by Chris Lattner · 19 years ago
- 01cae07 Make sure to mark unsupported SCALAR_TO_VECTOR operations as expand. by Chris Lattner · 19 years ago
- bbe77de Inform the dag combiner that the predicate compares only return a low bit. by Chris Lattner · 19 years ago
- 541f91b Custom lower all BUILD_VECTOR's so that we can compile vec_splat_u8(8) into by Chris Lattner · 19 years ago
- e3fea5a Rearrange code a bit by Chris Lattner · 19 years ago
- 32a988a Add, sub and shuffle are legal for all vector types by Chris Lattner · 19 years ago
- 33497cc note to self: *save* file, then check it in by Chris Lattner · 19 years ago
- 4468c22 Implement an item from the readme, folding vcmp/vcmp. instructions with by Chris Lattner · 19 years ago
- a17b155 Lower vector compares to VCMP nodes, just like we lower vector comparison by Chris Lattner · 19 years ago
- 36f4b0d Mark INSERT_VECTOR_ELT as expand by Chris Lattner · 19 years ago
- 98e70cc Add a few more altivec intrinsics by Nate Begeman · 19 years ago
- 7f20b13 Use normal lvx for scalar_to_vector instead of lve*x. They do the exact by Chris Lattner · 19 years ago
- 48b61a7 Tblgen doesn't like multiple SDNode<> definitions that map to the sameenum value. Split them into separate enums. by Chris Lattner · 19 years ago
- f15485a SelectionDAGISel can now natively handle Switch instructions, in the same by Nate Begeman · 19 years ago
- 6d92cad Codegen vector predicate compares. by Chris Lattner · 19 years ago
- 5b6a01b Remove PPC:isZeroVector, use ISD::isBuildVectorAllZeros instead by Evan Cheng · 19 years ago
- 8d052bc Add some basic patterns for other datatypes by Chris Lattner · 19 years ago
- 9c61dcf Codegen things like: by Chris Lattner · 19 years ago
- 7fbcef7 Disable the i32->float G5 optimization. It is unsafe, as documented in the by Chris Lattner · 19 years ago
- 64b3a08 add support for using vxor to build zero vectors. This implements by Chris Lattner · 19 years ago