| Commit message (Collapse) | Author | Age | Files | Lines | ||
|---|---|---|---|---|---|---|
| ... | ||||||
| * | Initial documentation for first-class aggregates changes. | Dan Gohman | 2008-05-12 | 1 | -0/+116 | |
| | | | | | llvm-svn: 51013 | |||||
| * | Fix a malformed %gt; | Dan Gohman | 2008-05-12 | 1 | -1/+1 | |
| | | | | | llvm-svn: 51011 | |||||
| * | Teach Legalize how to scalarize VSETCC | Nate Begeman | 2008-05-12 | 2 | -0/+37 | |
| | | | | | | | Teach X86 a few more vsetcc patterns. Custom lowering for unsupported ones is next. llvm-svn: 51009 | |||||
| * | Xform bitconvert(build_pair(load a, load b)) to a single load if the load ↵ | Evan Cheng | 2008-05-12 | 6 | -75/+63 | |
| | | | | | | | locations are at the right offset from each other. llvm-svn: 51008 | |||||
| * | New test for tail merging | Dale Johannesen | 2008-05-12 | 1 | -0/+145 | |
| | | | | | llvm-svn: 51007 | |||||
| * | Be more aggressive about tail-merging small blocks | Dale Johannesen | 2008-05-12 | 1 | -1/+12 | |
| | | | | | | | | if those blocks consist entirely of common instructions; merging will not add an extra branch in this case. llvm-svn: 51006 | |||||
| * | Constify isSourceDefinedByImplicitDef function. Otherwise, just formatting | Bill Wendling | 2008-05-12 | 1 | -45/+38 | |
| | | | | | | | changes that don't change functionality. llvm-svn: 51004 | |||||
| * | Constify the machine instruction passed into the | Bill Wendling | 2008-05-12 | 3 | -4/+5 | |
| | | | | | | | "is{Trivially,Really}ReMaterializable" methods. llvm-svn: 51001 | |||||
| * | Initial X86 codegen support for VSETCC. | Nate Begeman | 2008-05-12 | 3 | -11/+44 | |
| | | | | | llvm-svn: 51000 | |||||
| * | Further rework of tail merge algorithm. Not quite | Dale Johannesen | 2008-05-12 | 1 | -111/+77 | |
| | | | | | | | | | semantically identical, but little difference in either results or execution speed; but it's much easier to read, at least IMO. llvm-svn: 50999 | |||||
| * | Simplify some checks | Nate Begeman | 2008-05-12 | 1 | -4/+4 | |
| | | | | | llvm-svn: 50998 | |||||
| * | Fix a copy+paste bug; pseudo-instructions shouldn't have | Dan Gohman | 2008-05-12 | 1 | -21/+14 | |
| | | | | | | | encoding information. llvm-svn: 50997 | |||||
| * | Pointer comparisons should use icmp, not vicmp | Nate Begeman | 2008-05-12 | 1 | -1/+1 | |
| | | | | | llvm-svn: 50996 | |||||
| * | Go back to passing the analyses around as parameters. | Owen Anderson | 2008-05-12 | 1 | -21/+33 | |
| | | | | | llvm-svn: 50995 | |||||
| * | Pointer comparisons should be handled by icmp, not vicmp :) | Nate Begeman | 2008-05-12 | 2 | -3/+3 | |
| | | | | | llvm-svn: 50994 | |||||
| * | Forgot this. | Evan Cheng | 2008-05-12 | 1 | -7/+19 | |
| | | | | | llvm-svn: 50993 | |||||
| * | Hard code CmpInst back to i1 for now while I go track down what in the ↵ | Nate Begeman | 2008-05-12 | 1 | -2/+2 | |
| | | | | | | | | | bitcode reader/writer is assuming i1 This was breaking a bunch of tests llvm-svn: 50992 | |||||
| * | Refactor isConsecutiveLoad from X86 to TargetLowering so DAG combiner can ↵ | Evan Cheng | 2008-05-12 | 3 | -67/+95 | |
| | | | | | | | make use of it. llvm-svn: 50991 | |||||
| * | Revert the previous commit. Go ahead and hoist rematerializable instructions. | Bill Wendling | 2008-05-12 | 1 | -4/+0 | |
| | | | | | llvm-svn: 50990 | |||||
| * | Add support for vicmp/vfcmp codegen, more legalize support coming. | Nate Begeman | 2008-05-12 | 6 | -0/+115 | |
| | | | | | | | This is necessary to unbreak the build. llvm-svn: 50988 | |||||
| * | One real change - don't hoist something that's trivially rematerializable. It's | Bill Wendling | 2008-05-12 | 1 | -18/+19 | |
| | | | | | | | | | possible for it to produce worse code than before. The rest of this patch is code cleanup. llvm-svn: 50987 | |||||
| * | Fix build breakage | Nate Begeman | 2008-05-12 | 1 | -1/+1 | |
| | | | | | llvm-svn: 50986 | |||||
| * | Add two new instructions to the llvm IR, vicmp and vfcmp. see updated LangRef | Nate Begeman | 2008-05-12 | 18 | -1523/+2017 | |
| | | | | | | | for details. CodeGen support coming in a follow up patch llvm-svn: 50985 | |||||
| * | Make firstEightPowers const. | Dan Gohman | 2008-05-12 | 1 | -2/+2 | |
| | | | | | llvm-svn: 50975 | |||||
| * | Update comments. | Dan Gohman | 2008-05-12 | 2 | -5/+4 | |
| | | | | | llvm-svn: 50974 | |||||
| * | Filter option names to escape symbols not allowed as C++ identifiers. | Mikhail Glushenkov | 2008-05-12 | 3 | -8/+37 | |
| | | | | | | | | Makes it possible to use options with names like "Wa,". Also fixes the -Wall option handling as a side-effect. llvm-svn: 50973 | |||||
| * | Make it possible to choose between different compilation graph definitions ↵ | Mikhail Glushenkov | 2008-05-12 | 1 | -1/+2 | |
| | | | | | | | at compile-time. llvm-svn: 50972 | |||||
| * | Fix some error messages; Make LLVMC pass through the exit code of a failed tool. | Mikhail Glushenkov | 2008-05-12 | 3 | -9/+47 | |
| | | | | | llvm-svn: 50971 | |||||
| * | Fix a compile error on compilers that still want a return value | Dan Gohman | 2008-05-12 | 1 | -0/+1 | |
| | | | | | | | in a non-void function that calls abort. llvm-svn: 50969 | |||||
| * | Fix a missing break in the ISD::FLT_ROUNDS_ handling. Patch by giuma! | Dan Gohman | 2008-05-12 | 1 | -0/+1 | |
| | | | | | llvm-svn: 50967 | |||||
| * | Testcase for PR2264. | Duncan Sands | 2008-05-12 | 1 | -0/+10 | |
| | | | | | llvm-svn: 50965 | |||||
| * | Move the various analyses used by GVN into static variables so we don't have ↵ | Owen Anderson | 2008-05-12 | 1 | -30/+21 | |
| | | | | | | | to keep passing them around or refetching them. llvm-svn: 50963 | |||||
| * | prune #includes. | Chris Lattner | 2008-05-12 | 1 | -3/+0 | |
| | | | | | llvm-svn: 50962 | |||||
| * | Add a new SparsePropagation analysis utility, which allows you to do | Chris Lattner | 2008-05-12 | 2 | -0/+498 | |
| | | | | | | | | | SCCP like sparse lattice analysis with relative ease. Just pick your lattice function and implement the transfer function and you're good. Just make sure you don't break monotonicity ;-) llvm-svn: 50961 | |||||
| * | Make constructors target-specific. This fixes problems where the path would | Nick Lewycky | 2008-05-11 | 3 | -4/+19 | |
| | | | | | | | include backslashes on Windows. This should fix llvm-ld problems on win32. llvm-svn: 50960 | |||||
| * | Add note | Anton Korobeynikov | 2008-05-11 | 1 | -0/+38 | |
| | | | | | llvm-svn: 50959 | |||||
| * | Fix various DOUTs to not call the extremely expensive Value::getName() | Chris Lattner | 2008-05-11 | 1 | -5/+5 | |
| | | | | | | | | | method. DOUT statements are disabled when assertions are off, but the side effects of getName() are still evaluated. Just call getNameSTart, which is close enough and doesn't cause heap traffic. llvm-svn: 50958 | |||||
| * | Simplify code by using SwitchInst::findCaseValue instead of reimplementing it. | Chris Lattner | 2008-05-10 | 1 | -14/+2 | |
| | | | | | llvm-svn: 50957 | |||||
| * | Remove warnings when using -Wshorten-64-to-32. | Bill Wendling | 2008-05-10 | 1 | -3/+3 | |
| | | | | | llvm-svn: 50952 | |||||
| * | Testcase for PR2303. | Duncan Sands | 2008-05-10 | 1 | -0/+33 | |
| | | | | | llvm-svn: 50951 | |||||
| * | Fix linking of internal aliases | Anton Korobeynikov | 2008-05-10 | 1 | -28/+55 | |
| | | | | | llvm-svn: 50950 | |||||
| * | Prevent -W64-to-32-shortened warnings. | Bill Wendling | 2008-05-10 | 1 | -2/+2 | |
| | | | | | llvm-svn: 50947 | |||||
| * | Prevent warnings from the -Wshorten-64-to-32 flag. | Bill Wendling | 2008-05-10 | 1 | -5/+2 | |
| | | | | | llvm-svn: 50946 | |||||
| * | remove commented-out code, it is subsumed by ↵ | Gabor Greif | 2008-05-10 | 1 | -9/+0 | |
| | | | | | | | DECLARE_TRANSPARENT_OPERAND_ACCESSORS llvm-svn: 50944 | |||||
| * | merge of use-diet branch to trunk | Gabor Greif | 2008-05-10 | 18 | -584/+1528 | |
| | | | | | llvm-svn: 50943 | |||||
| * | Re-enable loop deletion by default. | Owen Anderson | 2008-05-10 | 1 | -0/+1 | |
| | | | | | llvm-svn: 50941 | |||||
| * | When transforming a vector_shuffle to a load, the base address must not be ↵ | Evan Cheng | 2008-05-10 | 2 | -0/+12 | |
| | | | | | | | an undef. llvm-svn: 50940 | |||||
| * | Set to 2.4 and regenerate configure. | Tanya Lattner | 2008-05-10 | 2 | -35/+35 | |
| | | | | | llvm-svn: 50935 | |||||
| * | Add nounwind. | Evan Cheng | 2008-05-10 | 1 | -2/+2 | |
| | | | | | llvm-svn: 50931 | |||||
| * | For now, abort when an ISD::VAARG is encountered on x86-64, rather | Dan Gohman | 2008-05-10 | 2 | -3/+18 | |
| | | | | | | | | | | than silently generate invalid code. llvm-gcc does not currently use VAArgInst; it lowers va_arg in the front-end. llvm-svn: 50930 | |||||

