summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Emit NOT instructions.Chris Lattner2005-01-111-1/+14
| | | | llvm-svn: 19455
* Implement the LOADABLE_MODULE option when building a shared library. ThisReid Spencer2005-01-111-4/+13
| | | | | | | | | | | | | passes the -module option on the libtool command line to ensure that the shared library being built can be dlopened and dlsym can work on that module. LOADABLE_MODULE should be sent only in conjunction with the SHARED_LIBRARY directive. It should generally be used for any module that is intended to be the target of an LLVM -load option. Note that loadable modules will not have the lib prefix but otherwise look like shared libraries. This is per the libtool recommendations and prevents these special shared libraries from being linked in via -l option to the linker. llvm-svn: 19454
* shift X, 0 -> XChris Lattner2005-01-111-0/+6
| | | | llvm-svn: 19453
* Fix a bug emitting branches that broke a lot of programs.Chris Lattner2005-01-111-12/+22
| | | | llvm-svn: 19452
* Be more careful where we set ContainsFPCode. We were missing a set in theChris Lattner2005-01-111-15/+10
| | | | | | | | int -> FP casting code. Note that we don't have to set it for FP operations that take FP values as operands: whatever produces the FP value will set the flag. llvm-svn: 19451
* Fix a major bug in setcc/cmov folding, where we accidentallyChris Lattner2005-01-111-6/+16
| | | | | | inverted the sense of the comparison. llvm-svn: 19450
* Take register pressure into account when we have to decide whether toChris Lattner2005-01-111-41/+232
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | evaluate the LHS or the RHS of an operation first. This causes good things to happen. For example, instead of compiling a loop to this: .LBBstrength_result7_1: # loopentry movl 16(%esp), %edi movl (%edi), %edi ;;; LOAD movl (%ecx), %ebx movl $2, (%eax,%ebx,4) movl (%edx), %ebx movl %esi, %ebp addl $21, %ebp addl $42, %esi cmpl $0, %edi ;;; USE cmovne %esi, %ebp cmpl %ebp, %ebx movl %ebp, %esi jg .LBBstrength_result7_1 We now compile it to this: .LBBstrength_result7_1: # loopentry movl %edi, %ebx addl $42, %ebx addl $21, %edi movl (%ecx), %ebp ;; LOAD cmpl $0, %ebp ;; USE cmovne %ebx, %edi movl (%edx), %ebx movl $2, (%eax,%ebx,4) movl (%esi), %ebx cmpl %edi, %ebx jg .LBBstrength_result7_1 Which reduces register pressure enough (in this case) to avoid spilling in the loop. As another example, consider the CodeGen/X86/regpressure.ll testcase. We used to generate this code for both cases: regpressure1: subl $32, %esp movl %esi, 12(%esp) movl %edi, 8(%esp) movl %ebx, 4(%esp) movl %ebp, (%esp) movl 36(%esp), %ecx movl (%ecx), %eax movl 4(%ecx), %edx movl %edx, 24(%esp) movl 8(%ecx), %edx movl %edx, 16(%esp) movl 12(%ecx), %edx movl 16(%ecx), %esi movl 20(%ecx), %edi movl 24(%ecx), %ebx movl %ebx, 28(%esp) movl 28(%ecx), %ebx movl 32(%ecx), %ebp movl %ebp, 20(%esp) movl 36(%ecx), %ecx imull 24(%esp), %eax imull 16(%esp), %eax imull %edx, %eax imull %esi, %eax imull %edi, %eax imull 28(%esp), %eax imull %ebx, %eax imull 20(%esp), %eax imull %ecx, %eax movl (%esp), %ebp movl 4(%esp), %ebx movl 8(%esp), %edi movl 12(%esp), %esi addl $32, %esp ret This code is basically trying to do all of the loads first, then execute all of the multiplies. Because we run out of registers, lots of spill code happens. We now generate this code for both cases: regpressure1: movl 4(%esp), %ecx movl (%ecx), %eax movl 4(%ecx), %edx imull %edx, %eax movl 8(%ecx), %edx imull %edx, %eax movl 12(%ecx), %edx imull %edx, %eax movl 16(%ecx), %edx imull %edx, %eax movl 20(%ecx), %edx imull %edx, %eax movl 24(%ecx), %edx imull %edx, %eax movl 28(%ecx), %edx imull %edx, %eax movl 32(%ecx), %edx imull %edx, %eax movl 36(%ecx), %ecx imull %ecx, %eax ret which is much nicer (when we fold loads into the muls it will be even better). The old instruction selector used to produce the good code for regpressure1 but not for regpressure2, as it depended on the order of operations in the LLVM code. llvm-svn: 19449
* The pattern isel is aggressively codegen'ing all of the loads in theseChris Lattner2005-01-111-0/+72
| | | | | | | | | | functions together at the start of the basic block, causing massive spillage. The old isel codegened the loads wherever they happened to land, so it generated good code for the first case, but bad code for the second. We really want the pattern isel to generate (the same) good code for both. llvm-svn: 19448
* Print SelectionDAGs bottom up, include extra info in the node labelsChris Lattner2005-01-111-3/+38
| | | | llvm-svn: 19447
* Add support for bottom-up graphs.Chris Lattner2005-01-112-3/+20
| | | | llvm-svn: 19446
* Add a marker for the graph root.Chris Lattner2005-01-101-0/+6
| | | | llvm-svn: 19445
* Put the operation name in each node, put the function name on the graph.Chris Lattner2005-01-101-0/+17
| | | | llvm-svn: 19444
* Split out SDNode::getOperationName into its own method.Chris Lattner2005-01-101-89/+88
| | | | llvm-svn: 19443
* Add a helper method.Chris Lattner2005-01-101-0/+3
| | | | llvm-svn: 19442
* Implement initial selectiondag printing support. This gets us a niceChris Lattner2005-01-101-0/+48
| | | | | | graph with no labels! :) llvm-svn: 19441
* Add support for graph operations, and add a viewGraph method to SelectionDAG.Chris Lattner2005-01-102-0/+77
| | | | llvm-svn: 19440
* Add a helper methodChris Lattner2005-01-101-1/+5
| | | | llvm-svn: 19439
* Fold setcc instructions into selects.Chris Lattner2005-01-101-18/+116
| | | | llvm-svn: 19438
* Add conditional moves for the parity flag.Chris Lattner2005-01-102-0/+36
| | | | llvm-svn: 19437
* Lower to the correct functions. This fixes FreeBench/fourinarowChris Lattner2005-01-101-2/+2
| | | | llvm-svn: 19436
* Implement 8-bit multiply for X86.Chris Lattner2005-01-101-1/+6
| | | | llvm-svn: 19435
* Rework constant pool handling so that function constant pools are no longerChris Lattner2005-01-101-21/+24
| | | | | | | leaked to the system. Now they are destroyed with the JITMemoryManager is destroyed. llvm-svn: 19434
* Apply feedback from Chris.Jeff Cohen2005-01-102-4/+5
| | | | llvm-svn: 19432
* Apply feed back from Chris:Jeff Cohen2005-01-104-6/+5
| | | | | | | 1. Rename createLoaderPass to CreateProfileLoaderPass 2. Opt shouldn't use the pass registered in CodeGen. llvm-svn: 19431
* Implement a couple of more simplifications. This lets us codegen:Chris Lattner2005-01-101-12/+43
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | int test2(int * P, int* Q, int A, int B) { return P+A == P; } into: test2: movl 4(%esp), %eax movl 12(%esp), %eax shll $2, %eax cmpl $0, %eax sete %al movzbl %al, %eax ret instead of: test2: movl 4(%esp), %eax movl 12(%esp), %ecx leal (%eax,%ecx,4), %ecx cmpl %eax, %ecx sete %al movzbl %al, %eax ret ICC is producing worse code: test2: movl 4(%esp), %eax #8.5 movl 12(%esp), %edx #8.5 lea (%edx,%edx), %ecx #9.9 addl %ecx, %ecx #9.9 addl %eax, %ecx #9.9 cmpl %eax, %ecx #9.16 movl $0, %eax #9.16 sete %al #9.16 ret #9.16 as is GCC (looks like our old code): test2: movl 4(%esp), %edx movl 12(%esp), %eax leal (%edx,%eax,4), %ecx cmpl %edx, %ecx sete %al movzbl %al, %eax ret llvm-svn: 19430
* Fix incorrect constant folds, fixing Stepanov after the SHR patch.Chris Lattner2005-01-101-4/+4
| | | | llvm-svn: 19429
* Update System project in Visual Studio to reflect renamed files.Jeff Cohen2005-01-101-120/+8
| | | | llvm-svn: 19428
* Constant fold shifts, turning this loop:Chris Lattner2005-01-101-0/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | .LBB_Z5test0PdS__3: # no_exit.1 fldl data(,%eax,8) fldl 24(%esp) faddp %st(1) fstl 24(%esp) incl %eax movl $16000, %ecx sarl $3, %ecx cmpl %eax, %ecx fstpl 16(%esp) #FP_REG_KILL jg .LBB_Z5test0PdS__3 # no_exit.1 into: .LBB_Z5test0PdS__3: # no_exit.1 fldl data(,%eax,8) fldl 24(%esp) faddp %st(1) fstl 24(%esp) incl %eax cmpl $2000, %eax fstpl 16(%esp) #FP_REG_KILL jl .LBB_Z5test0PdS__3 # no_exit.1 llvm-svn: 19427
* Rename Unix/*.cpp and Win32/*.cpp to have a *.inc suffix so that the sillyReid Spencer2005-01-0923-15/+15
| | | | | | | gdb debugger doesn't get confused on which file it is reading (the one in lib/System or the one in lib/System/{Win32,Unix}) llvm-svn: 19426
* Add some folds for == and != comparisons. This allows us toChris Lattner2005-01-091-41/+58
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | codegen this loop in stepanov: no_exit.i: ; preds = %entry, %no_exit.i, %then.i, %_Z5checkd.exit %i.0.0 = phi int [ 0, %entry ], [ %i.0.0, %no_exit.i ], [ %inc.0, %_Z5checkd.exit ], [ %inc.012, %then.i ] ; <int> [#uses=3] %indvar = phi uint [ %indvar.next, %no_exit.i ], [ 0, %entry ], [ 0, %then.i ], [ 0, %_Z5checkd.exit ] ; <uint> [#uses=3] %result_addr.i.0 = phi double [ %tmp.4.i.i, %no_exit.i ], [ 0.000000e+00, %entry ], [ 0.000000e+00, %then.i ], [ 0.000000e+00, %_Z5checkd.exit ] ; <double> [#uses=1] %first_addr.0.i.2.rec = cast uint %indvar to int ; <int> [#uses=1] %first_addr.0.i.2 = getelementptr [2000 x double]* %data, int 0, uint %indvar ; <double*> [#uses=1] %inc.i.rec = add int %first_addr.0.i.2.rec, 1 ; <int> [#uses=1] %inc.i = getelementptr [2000 x double]* %data, int 0, int %inc.i.rec ; <double*> [#uses=1] %tmp.3.i.i = load double* %first_addr.0.i.2 ; <double> [#uses=1] %tmp.4.i.i = add double %result_addr.i.0, %tmp.3.i.i ; <double> [#uses=2] %tmp.2.i = seteq double* %inc.i, getelementptr ([2000 x double]* %data, int 0, int 2000) ; <bool> [#uses=1] %indvar.next = add uint %indvar, 1 ; <uint> [#uses=1] br bool %tmp.2.i, label %_Z10accumulateIPddET0_T_S2_S1_.exit, label %no_exit.i To this: .LBB_Z4testIPddEvT_S1_T0__1: # no_exit.i fldl data(,%eax,8) fldl 16(%esp) faddp %st(1) fstpl 16(%esp) incl %eax movl %eax, %ecx shll $3, %ecx cmpl $16000, %ecx #FP_REG_KILL jne .LBB_Z4testIPddEvT_S1_T0__1 # no_exit.i instead of this: .LBB_Z4testIPddEvT_S1_T0__1: # no_exit.i fldl data(,%eax,8) fldl 16(%esp) faddp %st(1) fstpl 16(%esp) incl %eax leal data(,%eax,8), %ecx leal data+16000, %edx cmpl %edx, %ecx #FP_REG_KILL jne .LBB_Z4testIPddEvT_S1_T0__1 # no_exit.i llvm-svn: 19425
* Add last four createXxxPass functionsJeff Cohen2005-01-096-0/+44
| | | | llvm-svn: 19424
* Fix VC++ compilation errorJeff Cohen2005-01-091-0/+1
| | | | llvm-svn: 19423
* Print the DAG out more like a DAG in nested format.Chris Lattner2005-01-091-2/+18
| | | | llvm-svn: 19422
* Print out nodes sorted by their address to make it easier to find them in a ↵Chris Lattner2005-01-091-2/+5
| | | | | | list. llvm-svn: 19421
* Codegen (Reg|imm)+&GV as an LEA, because we cannot put it into the immediate ↵Chris Lattner2005-01-091-2/+2
| | | | | | | | | | | | | | | | field of an ADDri (due to current restrictions on MachineOperand :( ). This allows us to generate: leal Data+16000, %edx instead of: movl $Data, %edx addl $16000, %edx llvm-svn: 19420
* Add a simple transformation. This allows us to compile one of the innerChris Lattner2005-01-091-0/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | loops in stepanov to this: .LBB_Z5test0PdS__2: # no_exit.1 fldl data(,%eax,8) fldl 24(%esp) faddp %st(1) fstl 24(%esp) incl %eax cmpl $2000, %eax fstpl 16(%esp) #FP_REG_KILL jl .LBB_Z5test0PdS__2 instead of this: .LBB_Z5test0PdS__2: # no_exit.1 fldl data(,%eax,8) fldl 24(%esp) faddp %st(1) fstl 24(%esp) incl %eax movl $data, %ecx movl %ecx, %edx addl $16000, %edx subl %ecx, %edx movl %edx, %ecx sarl $2, %ecx shrl $29, %ecx addl %ecx, %edx sarl $3, %edx cmpl %edx, %eax fstpl 16(%esp) #FP_REG_KILL jl .LBB_Z5test0PdS__2 The old instruction selector produced: .LBB_Z5test0PdS__2: # no_exit.1 fldl 24(%esp) faddl data(,%eax,8) fstl 24(%esp) movl %eax, %ecx incl %ecx incl %eax leal data+16000, %edx movl $data, %edi subl %edi, %edx movl %edx, %edi sarl $2, %edi shrl $29, %edi addl %edi, %edx sarl $3, %edx cmpl %edx, %ecx fstpl 16(%esp) #FP_REG_KILL jl .LBB_Z5test0PdS__2 # no_exit.1 Which is even worse! llvm-svn: 19419
* Fix copy and pasto's for FP -> Int. This fixes fldryChris Lattner2005-01-091-2/+2
| | | | llvm-svn: 19418
* Fix a bug legalizing call instructions (make sure to remember all resultChris Lattner2005-01-091-16/+10
| | | | | | values), and eliminate some switch statements. llvm-svn: 19417
* Fix a minor bug legalizing dynamic_stackalloc. This allows us to compileChris Lattner2005-01-091-0/+2
| | | | | | | | std::__pad<wchar_t, std::char_traits<wchar_t> >::_S_pad(std::ios_base&, wchar_t, wchar_t*, wchar_t const*, int, int, bool) from libstdc++ llvm-svn: 19416
* Teach legalize to deal with DYNAMIC_STACKALLOC (aka a dynamic llvm alloca)Chris Lattner2005-01-091-1/+16
| | | | llvm-svn: 19415
* Initial implementation of FP->INT and INT->FP castsChris Lattner2005-01-091-3/+220
| | | | | | Also, fix zero_extend from bool to i8, which fixes Shootout/objinst. llvm-svn: 19414
* Get lib/Analysis/DataStructure to compile with VC++Jeff Cohen2005-01-096-2/+75
| | | | llvm-svn: 19412
* Workaround a VC++ bug. Microsoft puts the hash_map class in namespace ↵Jeff Cohen2005-01-091-1/+1
| | | | | | | | | | stdext, but this classes uses a utility function in namespace std. But Microsoft apparently assumes everyone will "using namespace std;". As LLVM doesn't.... Add a "use std::_Distance;" to get it working. llvm-svn: 19411
* Fix a subtle bug involving constant expr casts from int to fpChris Lattner2005-01-091-2/+2
| | | | llvm-svn: 19410
* Handle static alloca arguments to PHI nodes.Chris Lattner2005-01-091-3/+8
| | | | llvm-svn: 19409
* Implement varargs and returnaddress/frameaddress intrinsics. With thisChris Lattner2005-01-091-1/+66
| | | | | | patch, all of SingleSource/UnitTests passes. llvm-svn: 19408
* Use new interfaces to correctly lower varargs and return/frame address ↵Chris Lattner2005-01-091-25/+61
| | | | | | intrinsics. llvm-svn: 19407
* Add interfaces to lower varargs and return/frame address intrinsics.Chris Lattner2005-01-091-0/+31
| | | | llvm-svn: 19406
* Test fails on all platforms, not just linuxJeff Cohen2005-01-081-1/+1
| | | | llvm-svn: 19405
* Add support for llvm.setjmp and longjmp. Only 3 SingleSource/UnitTests fail ↵Chris Lattner2005-01-081-1/+8
| | | | | | now. llvm-svn: 19404
OpenPOWER on IntegriCloud