[Mesa-dev] [PATCH] nir: add support for removing redundant stores to copy prop var

[Jason Ekstrand]

On Thu, Nov 8, 2018 at 1:39 AM Timothy Arceri <tarceri at itsqueeze.com> wrote:

> For example the following type of thing is seen in TCS from
> a number of Vulkan and DXVK games:
>
>         vec1 32 ssa_557 = deref_var &oPatch (shader_out float)
>         vec1 32 ssa_558 = intrinsic load_deref (ssa_557) ()
>         vec1 32 ssa_559 = deref_var &oPatch at 42 (shader_out float)
>         vec1 32 ssa_560 = intrinsic load_deref (ssa_559) ()
>         vec1 32 ssa_561 = deref_var &oPatch at 43 (shader_out float)
>         vec1 32 ssa_562 = intrinsic load_deref (ssa_561) ()
>         intrinsic store_deref (ssa_557, ssa_558) (1) /* wrmask=x */
>         intrinsic store_deref (ssa_559, ssa_560) (1) /* wrmask=x */
>         intrinsic store_deref (ssa_561, ssa_562) (1) /* wrmask=x */
>
> No shader-db changes on i965 (SKL).
>
> vkpipeline-db results RADV (VEGA):
>
> Totals from affected shaders:
> SGPRS: 8152 -> 8016 (-1.67 %)
> VGPRS: 6780 -> 7068 (4.25 %)
> Spilled SGPRs: 0 -> 0 (0.00 %)
> Spilled VGPRs: 0 -> 0 (0.00 %)
> Private memory VGPRs: 0 -> 0 (0.00 %)
> Scratch size: 0 -> 0 (0.00 %) dwords per thread
> Code Size: 480420 -> 468576 (-2.47 %) bytes
> LDS: 0 -> 0 (0.00 %) blocks
> Max Waves: 1027 -> 992 (-3.41 %)
> Wait states: 0 -> 0 (0.00 %)
>
> The Max Waves and VGPRS changes here are misleading. What is
> happening is a bunch of TCS outputs are being optimised away as
> they are now recognised as unused. This results in more varyings
> being compacted via nir_compact_varyings() which can results in
> more register pressure. This is an existing problem independent
> of this patch. I've run some benchmarks and haven't noticed any
> regrssions in affected games.
> ---
>  src/compiler/nir/nir_opt_copy_prop_vars.c | 30 +++++++++++++++--------
>  1 file changed, 20 insertions(+), 10 deletions(-)
>
> diff --git a/src/compiler/nir/nir_opt_copy_prop_vars.c
> b/src/compiler/nir/nir_opt_copy_prop_vars.c
> index 7a21ad56c79..fdabc1769a5 100644
> --- a/src/compiler/nir/nir_opt_copy_prop_vars.c
> +++ b/src/compiler/nir/nir_opt_copy_prop_vars.c
> @@ -676,18 +676,28 @@ copy_prop_vars_block(struct copy_prop_var_state
> *state,
>        }
>
>        case nir_intrinsic_store_deref: {
> -         struct value value = {
> -            .is_ssa = true
> -         };
> -
> -         for (unsigned i = 0; i < intrin->num_components; i++)
> -            value.ssa[i] = intrin->src[1].ssa;
> -
>           nir_deref_instr *dst = nir_src_as_deref(intrin->src[0]);
> -         unsigned wrmask = nir_intrinsic_write_mask(intrin);
>           struct copy_entry *entry =
> -            get_entry_and_kill_aliases(copies, dst, wrmask);
> -         store_to_entry(state, entry, &value, wrmask);
> +            lookup_entry_for_deref(copies, dst, nir_derefs_equal_bit);
> +         if (entry && entry->src.ssa[0] == intrin->src[1].ssa) {
>

I think you need to check all of the ssa values in the copy entry here.
Maybe a little value_equals_ssa helper?.  Otherwise, this looks good to me.


> +            /* If we are storing the value from a load of the same var the
> +             * store is redundant so remove it.
> +             */
> +            nir_instr_remove(instr);
> +         } else {
> +            struct value value = {
> +               .is_ssa = true
> +            };
> +
> +            for (unsigned i = 0; i < intrin->num_components; i++)
> +               value.ssa[i] = intrin->src[1].ssa;
> +
> +            unsigned wrmask = nir_intrinsic_write_mask(intrin);
> +            struct copy_entry *entry =
> +               get_entry_and_kill_aliases(copies, dst, wrmask);
> +            store_to_entry(state, entry, &value, wrmask);
> +         }
> +
>           break;
>        }
>
> --
> 2.19.1
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20181112/a897884d/attachment-0001.html>