llvm dev - Dec 2015 - LICM doesn't work for IntrReadMem intrinsic function

On 1 December 2015 at 12:33, Xiangyang Guo via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
> if I have the following IR, LICM doesn't work again, even if I use
> '-loop-rotate' firstly. In this IR, the return value of intrinsic
function
> "foo" is used by another function "func2" as parameter.
However, for the
> intrinsic function "foo", the parameters are still loop
invariant. Do you
> know why this happens? Thanks a lot.
Is there anything telling LLVM that func2 won't write to memory and
change the value returned by llvm.foo?

Cheers.

Tim.

Actually, no such information is attached to "func2".

For the following IR, if "func2" can only access the memory related to
the
first parameter "%b" and it will not change the value related to
second
parameter "%2", do you know how I can tell LLVM these information? And
if I
can do this, probably the LICM will work for llvm.foo? Thanks

 %2 = call i32 @llvm.foo(i32 %a, i32 1)
 %3 = call i32 @_Z5func2ii(i32 %b, i32 %2)

2015-12-01 15:51 GMT-05:00 Tim Northover <t.p.northover at gmail.com>:
> On 1 December 2015 at 12:33, Xiangyang Guo via llvm-dev
> <llvm-dev at lists.llvm.org> wrote:
> > if I have the following IR, LICM doesn't work again, even if I use
> > '-loop-rotate' firstly. In this IR, the return value of
intrinsic
> function
> > "foo" is used by another function "func2" as
parameter.  However, for the
> > intrinsic function "foo", the parameters are still loop
invariant. Do you
> > know why this happens? Thanks a lot.
>
> Is there anything telling LLVM that func2 won't write to memory and
> change the value returned by llvm.foo?
>
> Cheers.
>
> Tim.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20151201/ff1931ac/attachment.html>

Opps, the IR is

  %2 = call i32 @llvm.foo(i32 %a, i32 1)
  %3 = call i32 @_Z5func2ii(i32 %b, i32 %2)

2015-12-01 15:59 GMT-05:00 Xiangyang Guo <eceguo at gmail.com>:
> Actually, no such information is attached to "func2".
>
> For the following IR, if "func2" can only access the memory
related to the
> first parameter "%b" and it will not change the value related to
second
> parameter "%2", do you know how I can tell LLVM these
information? And if I
> can do this, probably the LICM will work for llvm.foo? Thanks
>
>  %2 = call i32 @llvm.foo(i32 %a, i32 1)
>  %3 = call i32 @_Z5func2ii(i32 %b, i32 %2)
>
> 2015-12-01 15:51 GMT-05:00 Tim Northover <t.p.northover at
gmail.com>:
>
>> On 1 December 2015 at 12:33, Xiangyang Guo via llvm-dev
>> <llvm-dev at lists.llvm.org> wrote:
>> > if I have the following IR, LICM doesn't work again, even if I
use
>> > '-loop-rotate' firstly. In this IR, the return value of
intrinsic
>> function
>> > "foo" is used by another function "func2" as
parameter.  However, for
>> the
>> > intrinsic function "foo", the parameters are still loop
invariant. Do
>> you
>> > know why this happens? Thanks a lot.
>>
>> Is there anything telling LLVM that func2 won't write to memory and
>> change the value returned by llvm.foo?
>>
>> Cheers.
>>
>> Tim.
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20151201/fa7a9884/attachment.html>

Hi, Tim,

I see your point. Only if I set the func2 to 'readonly', the llvm.foo
can
be loop invariant. Thanks.

Regards,

Xiangyang

2015-12-01 15:51 GMT-05:00 Tim Northover <t.p.northover at gmail.com>:
> On 1 December 2015 at 12:33, Xiangyang Guo via llvm-dev
> <llvm-dev at lists.llvm.org> wrote:
> > if I have the following IR, LICM doesn't work again, even if I use
> > '-loop-rotate' firstly. In this IR, the return value of
intrinsic
> function
> > "foo" is used by another function "func2" as
parameter.  However, for the
> > intrinsic function "foo", the parameters are still loop
invariant. Do you
> > know why this happens? Thanks a lot.
>
> Is there anything telling LLVM that func2 won't write to memory and
> change the value returned by llvm.foo?
>
> Cheers.
>
> Tim.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20151201/06d1876d/attachment.html>