[llvm-dev] Unnecessary spill/fill issue

[Lawrence, Peter via llvm-dev]

Jason,
          I've been trying to solve a similar problem by playing with the isReMaterializable
Machine instruction attribute.   Try setting that attribute on the "movabsq" instruction.
Try also isAsCheapAsAMove attribute.  No guarantees though, as I have not yet fully solved
my problem either.

I am hoping this attribute means what it says in the context of reg-alloc (it is as cheap or
cheaper to re-execute the instruction where the value is needed rather than spill/reload),
but I have also seen it used in LICM in a way that isn't so obvious. YMMV.

--Peter Lawrence.




Does anyone have any insight into this problem? Is there a way to minimize
excessive spill/fill for this kind of scenario?
Thanks,
Jason


On Fri, May 6, 2016 at 10:44 AM, Jason <thesurprises at gmail.com> wrote:

> Hi, I am using mcjit in llvm 3.6 to jit kernels to x86 avx2. I've noticed
> some inefficient use of the stack around constant vectors. In one example,
> I have code that computes a series of constant vectors at compile time.
> Each vector has a single use. In the final asm, I see a series of spills at
> the top of the function of all the constant vectors immediately to stack,
> then each use references the stack pointer directly:
>
> Lots of these at top of function:
>
> movabsq $.LCPI0_212, %rbx
> vmovaps (%rbx), %ymm0
> vmovaps %ymm0, 2816(%rsp)       # 32-byte Spill
>
> Later on, each use references the stack pointer:
>
> vpaddd 2816(%rsp), %ymm4, %ymm1 # 32-byte Folded Reload
>
> It seems the spill to stack is unnecessary. In one particularly bad
> kernel, I have 128 8-wide constant vectors, and so there is 4KB of stack
> use just for these constants. I think a better approach could be to load
> the constant vector pointers as needed:
>
> movabsq $.LCPI0_212, %rbx
> vpaddd (%rbx), %ymm4, %ymm1
>
>
> Thanks,
> Jason