llvm-project/clang/lib/CodeGen
Bill Wendling 985d1c5d15 Add the 'target-cpu' and 'target-features' attributes to functions.
The back-end will use these values to reconfigure code generation for different
features.

llvm-svn: 175308
2013-02-15 21:30:01 +00:00
..
ABIInfo.h Add support for AArch64 target. 2013-01-31 12:13:10 +00:00
BackendUtil.cpp [msan] Run more optimizations after MemorySanitizer pass. 2013-01-31 09:53:29 +00:00
CGBlocks.cpp Use the correct field to copy/dispose a __block variable. 2013-01-22 03:56:22 +00:00
CGBlocks.h Remove useless 'llvm::' qualifier from names like StringRef and others that are 2013-01-12 19:30:44 +00:00
CGBuilder.h Rewrite #includes for llvm/Foo.h to llvm/IR/Foo.h as appropriate to 2013-01-02 11:45:17 +00:00
CGBuiltin.cpp [ubsan] Add support for -fsanitize-blacklist 2013-01-18 11:30:38 +00:00
CGCUDANV.cpp Remove useless 'llvm::' qualifier from names like StringRef and others that are 2013-01-12 19:30:44 +00:00
CGCUDARuntime.cpp Sort all of Clang's files under 'lib', and fix up the broken headers 2012-12-04 09:13:33 +00:00
CGCUDARuntime.h CUDA: IR generation support for device stubs 2011-10-06 18:51:56 +00:00
CGCXX.cpp Sort all of Clang's files under 'lib', and fix up the broken headers 2012-12-04 09:13:33 +00:00
CGCXXABI.cpp Cleanup some clang code to use new type functions instead of using cast<>. 2012-10-25 15:39:14 +00:00
CGCXXABI.h Abstract out emitting the vdtor calls and do it properly when using -cxx-abi microsoft; also fix vdtor calls for the ARM ABI 2013-02-15 14:45:22 +00:00
CGCall.cpp Add the 'target-cpu' and 'target-features' attributes to functions. 2013-02-15 21:30:01 +00:00
CGCall.h Use the AttributeSet instead of AttributeWithIndex. 2013-01-27 02:46:53 +00:00
CGClass.cpp ubsan: Add checking for invalid downcasts. Per [expr.static.cast]p2 and p11, 2013-02-13 21:18:23 +00:00
CGCleanup.cpp Make sure we don't emit IR for unused EH cleanups. PR13359. 2012-08-02 00:10:24 +00:00
CGCleanup.h Documentation cleanup: 2012-06-15 22:10:14 +00:00
CGDebugInfo.cpp Add OpenCL samplers as Clang builtin types and check sampler related restrictions. 2013-02-07 10:55:47 +00:00
CGDebugInfo.h Revert r174246, accidentally committed. 2013-02-02 00:39:32 +00:00
CGDecl.cpp Revert r174246, accidentally committed. 2013-02-02 00:39:32 +00:00
CGDeclCXX.cpp [ubsan] Add support for -fsanitize-blacklist 2013-01-18 11:30:38 +00:00
CGException.cpp Call __cxa_begin_catch with the current exception before 2013-02-12 03:51:46 +00:00
CGExpr.cpp ubsan: Add checking for invalid downcasts. Per [expr.static.cast]p2 and p11, 2013-02-13 21:18:23 +00:00
CGExprAgg.cpp patch for PR9027 and // rdar://11861085 2013-01-25 23:57:05 +00:00
CGExprCXX.cpp Abstract out emitting the vdtor calls and do it properly when using -cxx-abi microsoft; also fix vdtor calls for the ARM ABI 2013-02-15 14:45:22 +00:00
CGExprComplex.cpp Implement OpenCL event_t as Clang builtin type, including event_t related OpenCL restrictions (OpenCL 1.2 spec 6.9) 2013-01-20 12:31:11 +00:00
CGExprConstant.cpp Add a new LangOpt NativeHalfType. This option allows for native half/fp16 2013-01-23 11:56:20 +00:00
CGExprScalar.cpp ubsan: Add checking for invalid downcasts. Per [expr.static.cast]p2 and p11, 2013-02-13 21:18:23 +00:00
CGObjC.cpp In ARC, emit non-peepholed +1s within the full-expression instead 2013-02-12 00:25:08 +00:00
CGObjCGNU.cpp Remove useless 'llvm::' qualifier from names like StringRef and others that are 2013-01-12 19:30:44 +00:00
CGObjCMac.cpp Properly assemble PHIs after a null-checked invoke of objc_msgSend. 2013-02-12 05:53:35 +00:00
CGObjCRuntime.cpp Rework the bitfield access IR generation to address PR13619 and 2012-12-06 11:14:44 +00:00
CGObjCRuntime.h objectiveC++: When throwing c++ exception of 2013-01-10 19:02:56 +00:00
CGOpenCLRuntime.cpp Add OpenCL samplers as Clang builtin types and check sampler related restrictions. 2013-02-07 10:55:47 +00:00
CGOpenCLRuntime.h Rewrite #includes for llvm/Foo.h to llvm/IR/Foo.h as appropriate to 2013-01-02 11:45:17 +00:00
CGRTTI.cpp Add OpenCL samplers as Clang builtin types and check sampler related restrictions. 2013-02-07 10:55:47 +00:00
CGRecordLayout.h Rewrite #includes for llvm/Foo.h to llvm/IR/Foo.h as appropriate to 2013-01-02 11:45:17 +00:00
CGRecordLayoutBuilder.cpp First pass at abstracting out a class for the target C++ ABI. 2013-01-25 22:30:49 +00:00
CGStmt.cpp PR14566: Debug Info: avoid top level lexical blocks in functions 2013-01-26 22:16:26 +00:00
CGVTT.cpp simplify a bunch of code to use the well-known LLVM IR types computed by CodeGenModule. 2012-02-07 00:39:47 +00:00
CGVTables.cpp The standard ARM C++ ABI dictates that inline functions are 2013-01-25 22:31:03 +00:00
CGVTables.h The standard ARM C++ ABI dictates that inline functions are 2013-01-25 22:31:03 +00:00
CGValue.h patch for PR9027 and // rdar://11861085 2013-01-25 23:57:05 +00:00
CMakeLists.txt clang/lib: [CMake] Update tblgen'd dependencies. 2012-07-27 06:18:33 +00:00
CodeGenAction.cpp Pass the target options through to code generation. 2013-02-14 08:09:20 +00:00
CodeGenFunction.cpp Emit virtual/deleting destructors properly with -cxx-abi microsoft, PR15058 2013-02-13 08:37:51 +00:00
CodeGenFunction.h ubsan: Add checking for invalid downcasts. Per [expr.static.cast]p2 and p11, 2013-02-13 21:18:23 +00:00
CodeGenModule.cpp Pass the target options through to code generation. 2013-02-14 08:09:20 +00:00
CodeGenModule.h Pass the target options through to code generation. 2013-02-14 08:09:20 +00:00
CodeGenTBAA.cpp Rewrite #includes for llvm/Foo.h to llvm/IR/Foo.h as appropriate to 2013-01-02 11:45:17 +00:00
CodeGenTBAA.h Rewrite #includes for llvm/Foo.h to llvm/IR/Foo.h as appropriate to 2013-01-02 11:45:17 +00:00
CodeGenTypes.cpp Add OpenCL samplers as Clang builtin types and check sampler related restrictions. 2013-02-07 10:55:47 +00:00
CodeGenTypes.h Rewrite #includes for llvm/Foo.h to llvm/IR/Foo.h as appropriate to 2013-01-02 11:45:17 +00:00
ItaniumCXXABI.cpp Abstract out emitting the vdtor calls and do it properly when using -cxx-abi microsoft; also fix vdtor calls for the ARM ABI 2013-02-15 14:45:22 +00:00
Makefile BUILD_ARCHIVE is the default for libraries, no need to set it. 2010-07-18 00:14:47 +00:00
MicrosoftCXXABI.cpp Abstract out emitting the vdtor calls and do it properly when using -cxx-abi microsoft; also fix vdtor calls for the ARM ABI 2013-02-15 14:45:22 +00:00
ModuleBuilder.cpp Pass the target options through to code generation. 2013-02-14 08:09:20 +00:00
README.txt
TargetInfo.cpp Add support for AArch64 target. 2013-01-31 12:13:10 +00:00
TargetInfo.h Fix the required args count for variadic blocks. 2012-12-07 07:03:17 +00:00

README.txt

IRgen optimization opportunities.

//===---------------------------------------------------------------------===//

The common pattern of
--
short x; // or char, etc
(x == 10)
--
generates an zext/sext of x which can easily be avoided.

//===---------------------------------------------------------------------===//

Bitfields accesses can be shifted to simplify masking and sign
extension. For example, if the bitfield width is 8 and it is
appropriately aligned then is is a lot shorter to just load the char
directly.

//===---------------------------------------------------------------------===//

It may be worth avoiding creation of alloca's for formal arguments
for the common situation where the argument is never written to or has
its address taken. The idea would be to begin generating code by using
the argument directly and if its address is taken or it is stored to
then generate the alloca and patch up the existing code.

In theory, the same optimization could be a win for block local
variables as long as the declaration dominates all statements in the
block.

NOTE: The main case we care about this for is for -O0 -g compile time
performance, and in that scenario we will need to emit the alloca
anyway currently to emit proper debug info. So this is blocked by
being able to emit debug information which refers to an LLVM
temporary, not an alloca.

//===---------------------------------------------------------------------===//

We should try and avoid generating basic blocks which only contain
jumps. At -O0, this penalizes us all the way from IRgen (malloc &
instruction overhead), all the way down through code generation and
assembly time.

On 176.gcc:expr.ll, it looks like over 12% of basic blocks are just
direct branches!

//===---------------------------------------------------------------------===//