[Bug c/112327] New: RVV: Redundant vmv1r for widen reduction

[Bug c/112327] New: RVV: Redundant vmv1r for widen reduction

[juzhe.zhong at rivai dot ai]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112327

            Bug ID: 112327
           Summary: RVV: Redundant vmv1r for widen reduction
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: juzhe.zhong at rivai dot ai
  Target Milestone: ---

#include "riscv_vector.h"

void rvv_dot_prod(int16_t *pSrcA, int16_t *pSrcB, uint32_t n, int64_t *result)
{
    size_t vl;
    vint16m4_t vSrcA, vSrcB;
    vint64m1_t vSum = __riscv_vmv_s_x_i64m1(0, 1);
    while (n > 0) {
        vl = __riscv_vsetvl_e16m4(n);
        vSrcA = __riscv_vle16_v_i16m4(pSrcA, vl);
        vSrcB = __riscv_vle16_v_i16m4(pSrcB, vl);
        vSum = __riscv_vwredsum_vs_i32m8_i64m1(__riscv_vwmul_vv_i32m8(vSrcA,
vSrcB, vl), vSum, vl);
        pSrcA += vl;
        pSrcB += vl;
        n -= vl;
    }
    *result = __riscv_vmv_x_s_i64m1_i64(vSum);
}

https://godbolt.org/z/sb8G7ExKP

GCC:

...
vmv1r.v v2,v1
...
vwredsum.vs     v1,v8,v2

Clang:

vwredsum.vs     v8, v24, v8

The root cause is that we don't allow vwredsum.vs vd,vs2,vs1, vs1 overlaps vd.

[Bug target/112327] RVV: Redundant vmv1r for widen reduction

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112327

--- Comment #1 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Pan Li <panli@gcc.gnu.org>:

https://gcc.gnu.org/g:1a0af6e5a99cd895a663f0221c25321ae802413f

commit r14-5067-g1a0af6e5a99cd895a663f0221c25321ae802413f
Author: Juzhe-Zhong <juzhe.zhong@rivai.ai>
Date:   Wed Nov 1 14:56:39 2023 +0800

    RISC-V: Allow dest operand and accumulator operand overlap of widen
reduction instruction[PR112327]

    Consider this following intrinsic code:

    void rvv_dot_prod(int16_t *pSrcA, int16_t *pSrcB, uint32_t n, int64_t
*result)
    {
        size_t vl;
        vint16m4_t vSrcA, vSrcB;
        vint64m1_t vSum = __riscv_vmv_s_x_i64m1(0, 1);
        while (n > 0) {
            vl = __riscv_vsetvl_e16m4(n);
            vSrcA = __riscv_vle16_v_i16m4(pSrcA, vl);
            vSrcB = __riscv_vle16_v_i16m4(pSrcB, vl);
            vSum =
__riscv_vwredsum_vs_i32m8_i64m1(__riscv_vwmul_vv_i32m8(vSrcA, vSrcB, vl), vSum,
vl);
            pSrcA += vl;
            pSrcB += vl;
            n -= vl;
        }
        *result = __riscv_vmv_x_s_i64m1_i64(vSum);
    }

    https://godbolt.org/z/vWd35W7G6

    Before this patch:

    ...
    Loop:
    ...
    vmv1r.v v2,v1
    ...
    vwredsum.vs     v1,v8,v2
    ...

    After this patch:

    ...
    Loop:
    ...
    vwredsum.vs     v1,v8,v1
    ...

            PR target/112327

    gcc/ChangeLog:

            * config/riscv/vector.md: Add '0'.

    gcc/testsuite/ChangeLog:

            * gcc.target/riscv/rvv/base/pr112327-1.c: New test.
            * gcc.target/riscv/rvv/base/pr112327-2.c: New test.

[Bug target/112327] RVV: Redundant vmv1r for widen reduction

[juzhe.zhong at rivai dot ai]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112327

JuzheZhong <juzhe.zhong at rivai dot ai> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #2 from JuzheZhong <juzhe.zhong at rivai dot ai> ---
Fixed