- Aug 02, 2007
-
-
Dan Gohman authored
Generalize isPSHUFDMask and add a unary SHUFPD pattern so that SHUFPD's memory operand alignment can be tested as well, with a fix to avoid breaking MMX's use of isPSHUFDMask. llvm-svn: 40756
-
Dan Gohman authored
llvm-svn: 40754
-
Owen Anderson authored
llvm-svn: 40751
-
Owen Anderson authored
llvm-svn: 40750
-
Owen Anderson authored
llvm-svn: 40749
-
Chris Lattner authored
llvm-svn: 40748
-
Owen Anderson authored
llvm-svn: 40746
-
Christopher Lamb authored
llvm-svn: 40745
-
Chris Lattner authored
llvm-svn: 40743
-
Chris Lattner authored
casts in the input. llvm-svn: 40741
-
Chris Lattner authored
llvm-svn: 40740
-
Chris Lattner authored
llvm-svn: 40739
-
Chris Lattner authored
gvn, gvnpre, dse, and predsimplify. To see these, use: make check-line-length llvm-svn: 40738
-
Devang Patel authored
exit edge to preserve LCSSA. Fix dominance frontier update during loop unswitch. This fixes PR 1589, again llvm-svn: 40737
-
Dan Gohman authored
X86InstrInfo::isReallyTriviallyReMaterializable knows how to handle with the isReMaterializable flag so that it is given a chance to handle them. Without hoisting constant-pool loads from loops this isn't very visible, though it does keep CodeGen/X86/constant-pool-remat-0.ll from making a copy of the constant pool on the stack. llvm-svn: 40736
-
Chris Lattner authored
operations of casts. This implements InstCombine/zext-fold.ll llvm-svn: 40726
-
Anders Carlsson authored
llvm-svn: 40725
-
Anders Carlsson authored
Add extend and extOrTrunc methods that do sign or zero extension depending on whether the integer is signed or not llvm-svn: 40724
-
Evan Cheng authored
llvm-svn: 40723
-
Evan Cheng authored
llvm-svn: 40722
-
Evan Cheng authored
llvm-svn: 40721
-
Chris Lattner authored
llvm-svn: 40720
-
Reid Spencer authored
This fixes test/Feature/llvm2cpp.ll llvm-svn: 40714
-
Scott Michel authored
llvm-svn: 40712
-
Christopher Lamb authored
llvm-svn: 40711
-
Evan Cheng authored
simply specify them as results and let scheduledag handle them. That is, instead of SDOperand Flag = DAG.getTargetNode(Opc, MVT::i32, MVT::Flag, ...) SDOperand Result = DAG.getCopyFromReg(Chain, X86::EAX, MVT::i32, Flag) Just write: SDOperand Result = DAG.getTargetNode(Opc, MVT::i32, MVT::i32, ...) And let scheduledag emit the move from X86::EAX to a virtual register. llvm-svn: 40710
-
Evan Cheng authored
llvm-svn: 40709
-
Evan Cheng authored
llvm-svn: 40703
-
Evan Cheng authored
llvm-svn: 40702
-
Evan Cheng authored
llvm-svn: 40701
-
Devang Patel authored
llvm-svn: 40698
-
Evan Cheng authored
llvm-svn: 40697
-
Reid Spencer authored
failure to assemble). llvm-svn: 40696
-
Devang Patel authored
exit edge to preserve LCSSA. Fix dominance frontier update during loop unswitch. This fixes PR 1589. llvm-svn: 40695
-
Owen Anderson authored
llvm-svn: 40692
-
- Aug 01, 2007
-
-
Evan Cheng authored
llvm-svn: 40691
-
Evan Cheng authored
llvm-svn: 40690
-
Evan Cheng authored
llvm-svn: 40689
-
Evan Cheng authored
llvm-svn: 40688
-
Lauro Ramos Venancio authored
llvm-svn: 40687
-