search for: reg1051

Displaying 5 results from an estimated 5 matches for "reg1051".

Did you mean: reg1031
2008 Jan 16
4
[LLVMdev] LiveInterval Questions
I had been assuming that give a LiveRange a, a.valno->def, if valid, would be the same as a.start. But this is apparently not always the case. For example: Predecessors according to CFG: 0x839d130 (#3) 0x8462780 (#35) 308 %reg1051 = MOV64rr %reg1227<kill> 312 %reg1052 = MOV64rr %reg1228<kill> 316 %reg1053 = MOV64rr %reg1229<kill> 320 %reg1054 = MOV64rr %reg1230<kill> 324 %reg1055<dead> = LEA64r %reg1047, 1, %reg1053, 0 328 %reg1135 = MOVSX64rr32 %reg1025 332 %reg1136 = MOV64rr %reg1135<kill&g...
2008 Jan 17
0
[LLVMdev] LiveInterval Questions
...08, at 11:49 AM, David Greene wrote: > I had been assuming that give a LiveRange a, a.valno->def, if > valid, would be the same as a.start. But this is apparently not > always the case. For example: > > Predecessors according to CFG: 0x839d130 (#3) 0x8462780 (#35) > 308 %reg1051 = MOV64rr %reg1227<kill> > 312 %reg1052 = MOV64rr %reg1228<kill> > 316 %reg1053 = MOV64rr %reg1229<kill> > 320 %reg1054 = MOV64rr %reg1230<kill> > 324 %reg1055<dead> = LEA64r %reg1047, 1, %reg1053, 0 > 328 %reg1135 = MOVSX64rr32 %reg1025 > 332 %reg113...
2007 Jul 12
1
[LLVMdev] backend problem with LiveInterval::removeRange
...a0:1 SU(6): 0x88c9c28: ch = BEQ 0x88c9ba0, 0x88c8a08, 0x88ca710, 0x88ca778 Selected machine code: bb20: 0x88c8140, LLVM BB @0x88bec48, ID#8: %reg1047 = LUi <ga:flags.2176> %reg1048 = LW 0, <fi#6> %reg1049 = ADDiu %reg1047, <ga:flags.2176> %reg1050 = ADDu %reg1049, %reg1048 %reg1051 = LBu 0, %reg1050 %reg1052 = ADDiu %ZERO, 0 BEQ %reg1051, %reg1052, mbb<cond_next46,0x88c8430> Successors according to CFG: 0x88c8430 (#13) 0x88c8218 (#9) Total amount of phi nodes to update: 0 Replacing.1 0x88c8d00: ch = TokenFactor 0x88c9540:1, 0x88c9540:1 With: 0x88c9540: i32,ch = l...
2008 Jan 17
0
[LLVMdev] LiveInterval Questions
...008, at 11:49 AM, David Greene wrote: > I had been assuming that give a LiveRange a, a.valno->def, if > valid, would be the same as a.start. But this is apparently not > always the case. For example: > > Predecessors according to CFG: 0x839d130 (#3) 0x8462780 (#35) > 308 %reg1051 = MOV64rr %reg1227<kill> > 312 %reg1052 = MOV64rr %reg1228<kill> > 316 %reg1053 = MOV64rr %reg1229<kill> > 320 %reg1054 = MOV64rr %reg1230<kill> > 324 %reg1055<dead> = LEA64r %reg1047, 1, %reg1053, 0 > 328 %reg1135 = MOVSX64rr32 %reg1025 > 332 %reg113...
2004 Jun 22
3
[LLVMdev] Linearscan allocator bug?
...1041 = move %reg1042 %ar7 = + %ar7, 8 %reg1043 = - %ar7, 1 store %reg1043, %reg1041 %reg1044 = - %ar7, 2 store %reg1044, %reg1029 %reg1045 = - %ar7, 3 store %reg1045, %reg1046 %reg1047 = - %ar7, 4 %reg1048 = move 0 store %reg1047, %reg1048 %reg1049 = - %ar7, 5 store %reg1049, %reg1050 %reg1051 = - %ar7, 6 %reg1052 = move 0 store %reg1051, %reg1052 %reg1053 = - %ar7, 7 store %reg1053, %reg1039 call <ga:printf> %ar7 = - %ar7, 8 %reg1054 = move %gr7 return shortcirc_done.1 (0x8065d90, LLVM BB @0x8060320): %reg1060 = phi %reg1032, mbb<shortcirc_next.0.selectcont.selectcont...