Displaying 3 results from an estimated 3 matches for "bb31".
Did you mean:
bb1
[LLVMdev] A question about GetElementPtr common subexpression elimination/loop invariant code motion
2007 Jan 29
2
[LLVMdev] A question about GetElementPtr common subexpression elimination/loop invariant code motion
....7, but I have
the feeling this optimization is already there somewhere)
I get the following llvm code :
int %test() {
entry:
%mat = alloca [7 x [7 x [7 x int]]], align 16 ; <[7 x
[7 x [7 x int]]]*> [#uses=2]
br label %cond_true
cond_true: ; preds = %bb31, %bb22, %cond_true, %entry
%j.1.2.ph = phi int [ 0, %entry ], [ %j.1.2.ph, %cond_true ], [
%tmp24, %bb22 ], [ 0, %bb31 ] ; <int> [#uses=4]
%i.0.0.ph = phi int [ 0, %entry ], [ %i.0.0.ph, %cond_true ], [
%i.0.0.ph, %bb22 ], [ %tmp33, %bb31 ] ; <int>...
2009 Feb 05
2
[LLVMdev] LLVM misses some cross-MBB and loop optimizations compared to GCC
...the live-out set of B2. This could help
to e.g. eliminate some useless reloads from spill slots, if the value
is available on the required physical register already. For example,
in the example below, the marked "movl 12(%esp), %ecx" instruction
could be eliminated.
.LBB2_2: # bb31
movl 12(%esp), %ecx
movl 8(%esp), %eax
cmpl $0, up+28(%eax,%ecx,4)
je .LBB2_9 # bb569
.LBB2_3: # bb41 ; <--- bb31 is the only predecessor of bb41
movl 12(%esp), %ecx ; <--- This could be eliminated!!!
movl 4(%e...
2009 Feb 06
0
[LLVMdev] LLVM misses some cross-MBB and loop optimizations compared to GCC
...could help
> to e.g. eliminate some useless reloads from spill slots, if the value
> is available on the required physical register already. For example,
> in the example below, the marked "movl 12(%esp), %ecx" instruction
> could be eliminated.
>
> .LBB2_2: # bb31
> movl 12(%esp), %ecx
> movl 8(%esp), %eax
> cmpl $0, up+28(%eax,%ecx,4)
> je .LBB2_9 # bb569
> .LBB2_3: # bb41 ; <--- bb31 is the only predecessor
> of bb41
> movl 12(%esp), %ecx ; <--- This could be el...