llvm-6502/lib/Target/Sparc
Dan Gohman 8be6bbe5bf Eliminate the ISel priority queue, which used the topological order for a
priority function. Instead, just iterate over the AllNodes list, which is
already in topological order. This eliminates a fair amount of bookkeeping,
and speeds up the isel phase by about 15% on many testcases.

The impact on most targets is that AddToISelQueue calls can be simply removed.

In the x86 target, there are two additional notable changes.

The rule-bending AND+SHIFT optimization in MatchAddress that creates new
pre-isel nodes during isel is now a little more verbose, but more robust.
Instead of either creating an invalid DAG or creating an invalid topological
sort, as it has historically done, it can now just insert the new nodes into
the node list at a position where they will be consistent with the topological
ordering.

Also, the address-matching code has logic that checked to see if a node was
"already selected". However, when a node is selected, it has all its uses
taken away via ReplaceAllUsesWith or equivalent, so it won't recieve any
further visits from MatchAddress. This code is now removed.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@58748 91177308-0d34-0410-b5e6-96231b3b80d8
2008-11-05 04:14:16 +00:00
..
CMakeLists.txt
DelaySlotFiller.cpp
FPMover.cpp
Makefile
README.txt
Sparc.h Avoid creating two TargetLowering objects for each target. 2008-10-03 16:55:19 +00:00
Sparc.td
SparcAsmPrinter.cpp Ignore extra 'r' modifier for now 2008-10-10 20:29:50 +00:00
SparcCallingConv.td Fix a thinko and unbreak sparc default CC 2008-10-10 21:47:37 +00:00
SparcInstrFormats.td
SparcInstrInfo.cpp Const-ify several TargetInstrInfo methods. 2008-10-16 01:49:15 +00:00
SparcInstrInfo.h Const-ify several TargetInstrInfo methods. 2008-10-16 01:49:15 +00:00
SparcInstrInfo.td Change CALLSEQ_BEGIN and CALLSEQ_END to take TargetConstant's as 2008-10-11 22:08:30 +00:00
SparcISelDAGToDAG.cpp Eliminate the ISel priority queue, which used the topological order for a 2008-11-05 04:14:16 +00:00
SparcISelLowering.cpp Teach DAGCombine to fold constant offsets into GlobalAddress nodes, 2008-10-18 02:06:02 +00:00
SparcISelLowering.h Teach DAGCombine to fold constant offsets into GlobalAddress nodes, 2008-10-18 02:06:02 +00:00
SparcRegisterInfo.cpp Switch the MachineOperand accessors back to the short names like 2008-10-03 15:45:36 +00:00
SparcRegisterInfo.h
SparcRegisterInfo.td
SparcSubtarget.cpp
SparcSubtarget.h
SparcTargetAsmInfo.cpp
SparcTargetAsmInfo.h
SparcTargetMachine.cpp Fix command-line option printing to print two spaces where needed, 2008-10-14 20:25:08 +00:00
SparcTargetMachine.h Avoid creating two TargetLowering objects for each target. 2008-10-03 16:55:19 +00:00

To-do
-----

* Keep the address of the constant pool in a register instead of forming its
  address all of the time.
* We can fold small constant offsets into the %hi/%lo references to constant
  pool addresses as well.
* When in V9 mode, register allocate %icc[0-3].
* Add support for isel'ing UMUL_LOHI instead of marking it as Expand.
* Emit the 'Branch on Integer Register with Prediction' instructions.  It's
  not clear how to write a pattern for this though:

float %t1(int %a, int* %p) {
        %C = seteq int %a, 0
        br bool %C, label %T, label %F
T:
        store int 123, int* %p
        br label %F
F:
        ret float undef
}

codegens to this:

t1:
        save -96, %o6, %o6
1)      subcc %i0, 0, %l0
1)      bne .LBBt1_2    ! F
        nop
.LBBt1_1:       ! T
        or %g0, 123, %l0
        st %l0, [%i1]
.LBBt1_2:       ! F
        restore %g0, %g0, %g0
        retl
        nop

1) should be replaced with a brz in V9 mode.

* Same as above, but emit conditional move on register zero (p192) in V9 
  mode.  Testcase:

int %t1(int %a, int %b) {
        %C = seteq int %a, 0
        %D = select bool %C, int %a, int %b
        ret int %D
}

* Emit MULX/[SU]DIVX instructions in V9 mode instead of fiddling 
  with the Y register, if they are faster.

* Codegen bswap(load)/store(bswap) -> load/store ASI

* Implement frame pointer elimination, e.g. eliminate save/restore for 
  leaf fns.
* Fill delay slots