On 09/26/2013 05:48 PM, Alexander Graf wrote:
> +    for (i = 0; i < (8 / ebytes); i++) {
> +        simd_ld(tcg_tmp, freg_offs_n + (i * ebytes), size);
> +        tcg_gen_shli_i64(tcg_tmp, tcg_tmp, shift);
> +        simd_st(tcg_tmp, freg_offs_d + (i * ebytes * 2), size + 1);
> +    }

Are we really going to open-code all of the simd insns?

If so, then it makes me wonder if we ought to handle them as tcg registers
instead of continually loading and storing into the backing storage.


r~

Reply via email to