isel has its own particular features that it wants in the CFG, in order to
reduce the number of times a constant is computed, etc. Make sure that we
clean up the CFG before doing any other things for isel. Doing so can
dramatically reduce the number of split edges and reduce the number of
places that constants get computed. For example, this shrinks
CodeGen/Generic/phi-immediate-factoring.ll from 44 to 37 instructions on X86,
and from 21 to 17 MBB's in the output. This is primarily a code size win,
not a performance win.
This implements CodeGen/Generic/phi-immediate-factoring.ll and PR1296.
llvm-svn: 35575
Make sure that the CTPOP result is casted to i32 as the bit counting
intrinsics all return i32 now (this affects CTLZ and CTTZ as well).
llvm-svn: 35567
Document changes to bswap and bit counting intrinsics. bswap's name now
requires two types in the suffix per overloaded intrinsic naming rules.
The ctpop, cttz, and ctlz intrinsics were changed to always return i32.
llvm-svn: 35550
Update these test cases to use proper signatures for bswap which is now
and overloaded intrinsic. Its name must be of the form llvm.bswap.i32.i32
since both the parameter and the result or of type "iAny". Also, the
bit counting intrinsics changed to always return i32.
llvm-svn: 35548
Implement bswap, ctpop, ctlz, and cttz utilizing the new overloaded
intrinsic feature of tblgen. By defining llvm_int_ty to have an ValueType
of iAny, any intrinsic using that type becomes an overloaded intrinsic.
This signals that a) any integer type could be used for that argument and
b) the intrinsic needs a suffix to match each such argument for the
actual intrinsic.
llvm-svn: 35544
Intrinsic functions might need to be generated (if they are overloaded) so
returning a const char* is no longer possible. Change getOperationName to
return a std::string instead.
llvm-svn: 35543
Change the interface to the getName, getType, getDeclaration methods to
include the "actual" types of the result/arguments. This is necessary to
support overloaded intrinsics.
llvm-svn: 35542
Implement "actual" argument types for the Intrinsic member functions. This
involves changing the getName, getType, and getDeclaration methods to have
optional parameters for the actual types. These are necessary in order for
the type/name to be constructed properly for overloaded intrinsics. Only
the caller knows the actual argument types desired.
llvm-svn: 35541
1. Clear up confusion between "GotBits" and "ExpectBits". GotBits is the
type actually provided. ExpectedBits is the type expected for the
intrinsics. Before this patch, it was reversed!
2. Implement checks for overloaded intrinsics. This involves computing the
suffix expected and making sure the suffix matches the function name. It
also includes some intrinsic-specific checks such as ensuring that the
bswap parameter and result are the same width and an even number of bytes.
llvm-svn: 35540
Implement code generation for overloaded intrinsic functions. The basic
difference is that "actual" argument types must be provided when
constructing intrinsic names and types. Also, for recognition, only the
prefix is examined. If it matches, the suffix is assumed to match. The
suffix is checked by the Verifier, however.
llvm-svn: 35539
