qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Richard Henderson <richard.henderson@linaro.org>
To: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>, qemu-devel@nongnu.org
Cc: qemu-riscv@nongnu.org, palmer@dabbelt.com,
	alistair.francis@wdc.com, dbarboza@ventanamicro.com,
	liwei1518@gmail.com, bmeng.cn@gmail.com,
	TANG Tiancheng <tangtiancheng.ttc@alibaba-inc.com>
Subject: Re: [PATCH v2 13/14] tcg/riscv: Implement vector roti/v/x shi ops
Date: Tue, 3 Sep 2024 08:15:49 -0700	[thread overview]
Message-ID: <97873524-9e47-44c4-b34a-a27e9833b0e1@linaro.org> (raw)
In-Reply-To: <20240830061607.1940-14-zhiwei_liu@linux.alibaba.com>

On 8/29/24 23:16, LIU Zhiwei wrote:
> @@ -2589,6 +2605,69 @@ void tcg_expand_vec_op(TCGOpcode opc, TCGType type, unsigned vece,
>               }
>           }
>           break;
> +    case INDEX_op_shli_vec:
> +        if (a2 > 31) {
> +            tcg_gen_shls_vec(vece, v0, v1, tcg_constant_i32(a2));
> +        } else {
> +            vec_gen_3(INDEX_op_rvv_shli_vec, type, vece, tcgv_vec_arg(v0),
> +                      tcgv_vec_arg(v1), a2);
> +        }
> +        break;
> +    case INDEX_op_shri_vec:
> +        if (a2 > 31) {
> +            tcg_gen_shrs_vec(vece, v0, v1, tcg_constant_i32(a2));
> +        } else {
> +            vec_gen_3(INDEX_op_rvv_shri_vec, type, vece, tcgv_vec_arg(v0),
> +                      tcgv_vec_arg(v1), a2);
> +        }
> +        break;
> +    case INDEX_op_sari_vec:
> +        if (a2 > 31) {
> +            tcg_gen_sars_vec(vece, v0, v1, tcg_constant_i32(a2));
> +        } else {
> +            vec_gen_3(INDEX_op_rvv_sari_vec, type, vece, tcgv_vec_arg(v0),
> +                      tcgv_vec_arg(v1), a2);
> +        }
> +        break;
> +    case INDEX_op_rotli_vec:
> +        t1 = tcg_temp_new_vec(type);
> +        tcg_gen_shli_vec(vece, t1, v1, a2);
> +        tcg_gen_shri_vec(vece, v0, v1, (8 << vece) - a2);
> +        tcg_gen_or_vec(vece, v0, v0, t1);
> +        tcg_temp_free_vec(t1);
> +        break;
> +    case INDEX_op_rotls_vec:
> +        t1 = tcg_temp_new_vec(type);
> +        t2 = tcg_temp_new_i32();
> +        tcg_gen_neg_i32(t2, temp_tcgv_i32(arg_temp(a2)));
> +        tcg_gen_shrs_vec(vece, v0, v1, t2);
> +        tcg_gen_shls_vec(vece, t1, v1, temp_tcgv_i32(arg_temp(a2)));
> +        tcg_gen_or_vec(vece, v0, v0, t1);
> +        tcg_temp_free_vec(t1);
> +        tcg_temp_free_i32(t2);
> +        break;

I'm trying to work out how much benefit there is here of expanding these early, as opposed 
to simply using TCG_REG_TMP0 when the immediate doesn't fit, or for rotls_vec negation.

> +    case INDEX_op_rotlv_vec:
> +        v2 = temp_tcgv_vec(arg_temp(a2));
> +        t1 = tcg_temp_new_vec(type);
> +        tcg_gen_neg_vec(vece, t1, v2);
> +        vec_gen_3(INDEX_op_shrv_vec, type, vece, tcgv_vec_arg(t1),
> +                  tcgv_vec_arg(v1), tcgv_vec_arg(t1));
> +        vec_gen_3(INDEX_op_shlv_vec, type, vece, tcgv_vec_arg(v0),
> +                  tcgv_vec_arg(v1), tcgv_vec_arg(v2));
> +        tcg_gen_or_vec(vece, v0, v0, t1);
> +        tcg_temp_free_vec(t1);
> +        break;
> +    case INDEX_op_rotrv_vec:
> +        v2 = temp_tcgv_vec(arg_temp(a2));
> +        t1 = tcg_temp_new_vec(type);
> +        tcg_gen_neg_vec(vece, t1, v2);
> +        vec_gen_3(INDEX_op_shlv_vec, type, vece, tcgv_vec_arg(t1),
> +                  tcgv_vec_arg(v1), tcgv_vec_arg(t1));
> +        vec_gen_3(INDEX_op_shrv_vec, type, vece, tcgv_vec_arg(v0),
> +                  tcgv_vec_arg(v1), tcgv_vec_arg(v2));
> +        tcg_gen_or_vec(vece, v0, v0, t1);
> +        tcg_temp_free_vec(t1);
> +        break;

And here we can use TCG_REG_V0 as the temporary, both for negation and shift intermediate.

     vrsub_vi  V0, a2, 0
     vshlv_vv  V0, a1, V0
     vshrv_vv  a0, a1, a2
     vor_vv    a0, a0, V0


r~


  reply	other threads:[~2024-09-03 15:17 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-30  6:15 [PATCH v2 00/14] tcg/riscv: Add support for vector LIU Zhiwei
2024-08-30  6:15 ` [PATCH v2 01/14] tcg/op-gvec: Fix iteration step in 32-bit operation LIU Zhiwei
2024-08-31 23:59   ` Richard Henderson
2024-08-30  6:15 ` [PATCH v2 02/14] util: Add RISC-V vector extension probe in cpuinfo LIU Zhiwei
2024-09-02  0:12   ` Richard Henderson
2024-08-30  6:15 ` [PATCH v2 03/14] tcg/riscv: Add basic support for vector LIU Zhiwei
2024-09-02  0:28   ` Richard Henderson
2024-08-30  6:15 ` [PATCH v2 04/14] tcg/riscv: Add riscv vset{i}vli support LIU Zhiwei
2024-09-02  1:06   ` Richard Henderson
2024-08-30  6:15 ` [PATCH v2 05/14] tcg/riscv: Implement vector load/store LIU Zhiwei
2024-09-02  1:31   ` Richard Henderson
2024-08-30  6:15 ` [PATCH v2 06/14] tcg/riscv: Implement vector mov/dup{m/i} LIU Zhiwei
2024-09-02  1:36   ` Richard Henderson
2024-08-30  6:16 ` [PATCH v2 07/14] tcg/riscv: Add support for basic vector opcodes LIU Zhiwei
2024-09-02  1:39   ` Richard Henderson
2024-08-30  6:16 ` [PATCH v2 08/14] tcg/riscv: Implement vector cmp ops LIU Zhiwei
2024-09-03  6:45   ` Richard Henderson
2024-09-03 14:51     ` Richard Henderson
2024-08-30  6:16 ` [PATCH v2 09/14] tcg/riscv: Implement vector neg ops LIU Zhiwei
2024-09-03 14:52   ` Richard Henderson
2024-08-30  6:16 ` [PATCH v2 10/14] tcg/riscv: Implement vector sat/mul ops LIU Zhiwei
2024-09-03 14:52   ` Richard Henderson
2024-08-30  6:16 ` [PATCH v2 11/14] tcg/riscv: Implement vector min/max ops LIU Zhiwei
2024-09-03 14:53   ` Richard Henderson
2024-08-30  6:16 ` [PATCH v2 12/14] tcg/riscv: Implement vector shs/v ops LIU Zhiwei
2024-09-03 14:54   ` Richard Henderson
2024-08-30  6:16 ` [PATCH v2 13/14] tcg/riscv: Implement vector roti/v/x shi ops LIU Zhiwei
2024-09-03 15:15   ` Richard Henderson [this message]
2024-09-04 15:25     ` LIU Zhiwei
2024-09-04 19:05       ` Richard Henderson
2024-09-05  1:40         ` LIU Zhiwei
2024-08-30  6:16 ` [PATCH v2 14/14] tcg/riscv: Enable native vector support for TCG host LIU Zhiwei
2024-09-03 15:02   ` Richard Henderson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=97873524-9e47-44c4-b34a-a27e9833b0e1@linaro.org \
    --to=richard.henderson@linaro.org \
    --cc=alistair.francis@wdc.com \
    --cc=bmeng.cn@gmail.com \
    --cc=dbarboza@ventanamicro.com \
    --cc=liwei1518@gmail.com \
    --cc=palmer@dabbelt.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-riscv@nongnu.org \
    --cc=tangtiancheng.ttc@alibaba-inc.com \
    --cc=zhiwei_liu@linux.alibaba.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).