Displaying 7 results from an estimated 7 matches for "vsxfmamutate".
2016 Mar 25
1
[VSXFMAMutate] OldFMAReg may be wrongly rewritten
On Tue, Mar 22, 2016 at 5:13 PM Eric Christopher wrote:
> I think we can probably go ahead and throw this up on Phabricator for
> review. I'd probably bring in Matthias for review as well.
>
> Thanks!
>
> -eric
>
(Follow up on the discuss in IRC)
I tried to bridge PPC backend and commuteInstruction, not sure if
correctly, but here I got some non-optimal results: in 12
2016 Feb 22
0
[VSXFMAMutate] OldFMAReg may be wrongly rewritten
On Fri, Feb 19, 2016 at 5:10 PM Tim Shen <timshen at google.com> wrote:
> I wonder if we can fix this by making the transformation simpler, that is,
> instead of doing:
>
I wrote a prototype (see attach) for this idea, it actually improves some
of the test cases (e.g. fma-assoc.ll: test_FMADD_ASSOC1), but pessimize
several other cases (e.g. test_FMADD_ASSOC_EXT1).
I'm not
2016 Feb 29
2
[VSXFMAMutate] OldFMAReg may be wrongly rewritten
Ping?
On Mon, Feb 22, 2016 at 1:06 PM Tim Shen <timshen at google.com> wrote:
> On Fri, Feb 19, 2016 at 5:10 PM Tim Shen <timshen at google.com> wrote:
>
>> I wonder if we can fix this by making the transformation simpler, that
>> is, instead of doing:
>>
>
> I wrote a prototype (see attach) for this idea, it actually improves some
> of the test cases
2016 Feb 20
2
[VSXFMAMutate] OldFMAReg may be wrongly rewritten
Example:
target triple = "powerpc64le-unknown-linux-gnu"
define void @TestFoo() {
entry_bb:
br label %loop_bb
loop_bb:
%tmp = load float, float* undef
%tmp1 = fmul float %tmp, 0x401F25E360000000
%tmp2 = fadd float %tmp1, 0x3FC1A7B960000000
%tmp3 = select i1 undef, float 0x401F25E360000000, float %tmp2
store float %tmp3, float* undef
br label %loop_bb
}
The code above is
2016 Mar 23
0
[VSXFMAMutate] OldFMAReg may be wrongly rewritten
...instructions with in-out registers.
>>>
>>> On Wed, Mar 2, 2016 at 3:40 PM Eric Christopher <echristo at gmail.com>
>>> wrote:
>>>
>>>> Relaying an IRC conversation here:
>>>>
>>>> Currently we're going to disable the VSXFMAMutate pass in ToT and talk
>>>> about ways to either rewrite or fix it to deal with more complicated cfgs.
>>>>
>>>> Thanks!
>>>>
>>>> -eric
>>>>
>>>> On Mon, Feb 29, 2016 at 2:21 PM Tim Shen <timshen at google.com>...
2016 Mar 05
2
[VSXFMAMutate] OldFMAReg may be wrongly rewritten
...rithm is unaware of any FMA related knowledge though, so it can be
used for any mixed instructions with in-out registers.
On Wed, Mar 2, 2016 at 3:40 PM Eric Christopher <echristo at gmail.com> wrote:
> Relaying an IRC conversation here:
>
> Currently we're going to disable the VSXFMAMutate pass in ToT and talk
> about ways to either rewrite or fix it to deal with more complicated cfgs.
>
> Thanks!
>
> -eric
>
> On Mon, Feb 29, 2016 at 2:21 PM Tim Shen <timshen at google.com> wrote:
>
>> Ping?
>>
>> On Mon, Feb 22, 2016 at 1:06 PM Tim Sh...
2016 Mar 16
2
[VSXFMAMutate] OldFMAReg may be wrongly rewritten
...;> be used for any mixed instructions with in-out registers.
>>
>> On Wed, Mar 2, 2016 at 3:40 PM Eric Christopher <echristo at gmail.com>
>> wrote:
>>
>>> Relaying an IRC conversation here:
>>>
>>> Currently we're going to disable the VSXFMAMutate pass in ToT and talk
>>> about ways to either rewrite or fix it to deal with more complicated cfgs.
>>>
>>> Thanks!
>>>
>>> -eric
>>>
>>> On Mon, Feb 29, 2016 at 2:21 PM Tim Shen <timshen at google.com> wrote:
>>>
>>...